{
  "text": "哈喽大家\n就等了同传字幕1.0正式版\n今天终于和大家见面了\n从最初的小工具到现在\nAndroid、Mac还有大家呼声超高的Windows\n三个平台这次全都齐了\n不管你是睡前抱着手机躺看直播\n还是在电脑前边干活边挂着生肉台\n伸手\n让字母自己跟上\n他现在能做什么\n平时看日语或者英语的视频和直播\n他能实时听声音\n出字幕\n翻成中文或者你想看的任何语言\n列表里自带了中\n英\n日\n韩\n其他任何语言\n只要你自己填个名字\n它就能翻\n字幕会像这样浮在画面最上面\n随便拖动\n改大小\n换配色\n房屋处\n你可以选边听边出\n草稿随说随出\n一句说完立刻变正式字幕\n也可以切成稳稳地说完再出\n在手机上\n你可以单独选某一个会出声的APP来听\n到了电脑上\n它直接听整台电脑播放的声音\n哪怕关掉主窗口字幕\n也照样在\nMac的菜单栏和Windows的右下角拖盘点一下\n就能随时开关\n隐私方面也完全不用担心\n声音指摘你自己的设备本地识别\n绝对不会上传\n而且语音识别模型\n已经直接内置在安装包里\n装好就能用\n不用另外下载\n翻译的时候\n也只是把识别出来的纯文字\n借口\n抵挡\n你可以按内容类型\n写好背景\n数语表\n甚至直接点一下让大模型帮你自动生成术语\n翻译时它还会自动结合上文语境\n不会翻得前言不搭后语\n翻译接入的是你自己的大模型API\n像Deepstick\n建容\n计备\n啊\n会自动释出\n关闭模型思考过程的参数来降低延迟\n资金一请求头\n不需要Key的服务把Key留空就行\n另外设置里还提供了一个可选的Javi\n在线辅助功能\n一文检查\n所以平时用起来其实就三步\n填好你的大模型API并测试连接允许权限\n然后点一下开始字幕\n就可以安心看直播了\n老朋友可能已经发现了\n1.0版本除了加新功能\n我还做了几个减法\n首先是移除了之前内置的index免费翻译\n它是B\n占官方提供的接口\n用的人一多\n不稳定\n翻译效果其实也不如其他主流大模型\n我后来想明白了一件事\n其他提供免费翻译的软件\n它们本身是做付费服务的商业软件\n对免费功能的可用性\n也是有保障义务的\n当免费翻译用不了的时候\n用户去反馈\n指责甚至骂两句\n那是商业软件应该受着的\n但同传字幕是个开源项目\nIndex是逼不赚提供的服务\n它不稳定的问题\n本来不应该反馈到我这里\n可只要我把它内置在软件里\n大家自然会觉得这是我提供的服务\n我就对它有义务\n所以我决定把它移除掉\n当然\n你依然可以通过现有的自定义配置\n自主合规地建入从前的index翻译接口\n但那是你亲手填上去用的\n你会清楚\n这是你自己的选择\n第二个减法是移除了本机大模型部署\n因为这个项目的重心是做好工具本身\n也就是模型调用和翻译相关功能的开发优化\n而在安卓上面对五花八门的系统\nSoC CPU GPU和NPU\n让手机连续跑本地模型做实时翻译\n本身就是个性能功耗和负载炸弹\nows和Mac上也是一样\n不仅要适配各种显卡和统一内存\n而且上了桌面端\n大家肯定会想要更强的模型\n两臂\n继续支持这些\n会给项目引入大量非必要的开发成本\n如果你想在本地部署模型\n其实有很多更优秀的专门项目\n用它们在电脑上部署好\n再通过本地API接入同传字幕\n专业的事交给专业的工具\n才更符合这个项目的定位\n这还带来了一个附带的好处\n以前为了照顾国内网络环境\n模型下载失配\n现在只剩语音识别模型\n干脆直接内置进安装包\n把下载部分彻底剥离掉\n也方便大家通过Titop以外的渠道分享传播\n只保留BYOK\n你选你信任的服务\n你来掌控成本和稳定性\n摇完了功能\n最后想跟大家说说\n做这个项目的初衷\n其实特别简单\n火雨直播\n千万同传\n因为看的人太多\n突然炸掉了\n那天半夜直播结束\n越想越上头的我\n决定干脆自己动手搓一个\n一开始\n真的只是打算留着自己用的\n直播镜里\n看到不少人被迫去买付费翻译软件\n收起抱怨\n我就想既然自己用着看直播挺好的\n不如分享出来给大家一起用吧\n两个切片和一个卢播\n后来又做了个特别简陋\n垃圾的小视频\n哈哈即使当时只是想有个能放下载地址\n接受大家反馈的地方\n没想到\n一下子收到了各种各样的反馈和功能诉求\n这几天我几乎把所有能想到的功能\n开发和三个平台的适配全都干出来了\n他有了今天正式和大家见面的1.0\n接下来\n我还会持续优化他的性能表现\n和翻译效果\n比如现在\n电脑端用的还是和手机端\n一样的语音识别模型\n手机上是实在没得选\n但在桌面端\n未来其实还有很大的升级和选择空间\n真心感谢这段时间\n每一位提出反馈\n帮我测试捉虫的朋友\n希望这个深夜上头\n搓出来的小工具\n也能陪你度过\n看台夜晚\n项目代码和三个平台的安装包都在GitHub上\n搜索CMU Subtitle就能找到\n如果它也帮到了你\n别忘了点赞投币收藏支持一下呀\n我们下个版本见\n拜拜",
  "language": "zh",
  "duration": 450.517313,
  "segments": [
    {
      "start": 3.2200000000000006,
      "end": 3.9800000000000004,
      "text": "哈喽大家",
      "raw_asr_text": "哈喽大家",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 0,
      "words": [
        {
          "word": "哈",
          "start": 3.2200000000000006,
          "end": 3.58,
          "probability": 0.50634765625
        },
        {
          "word": "喽",
          "start": 3.58,
          "end": 3.68,
          "probability": 0.9775390625
        },
        {
          "word": "大家",
          "start": 3.68,
          "end": 3.9800000000000004,
          "probability": 0.8955078125
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.13841238822645813,
        "compression_ratio": 1.1423611111111112,
        "no_speech_prob": 2.5074888346243718e-11
      },
      "lid_status": "UNCERTAIN",
      "alternatives": [
        {
          "language": "zh",
          "text": "哈喽大家就等了同传字幕1.0正式版今天终于和大家见面了从最初的小工具到现在Android、Mac还有大家呼声超高的Windows三个平台这次全都齐了不管你是睡前抱着手机躺看直播还是在电脑前边干活边挂着生肉台现在都能用同一套伸手的体验让字母自己跟上"
        }
      ],
      "id": 0,
      "review_annotation": "",
      "row_id": "0",
      "bounded_row_id": 0,
      "parent_row_ids": [
        0
      ],
      "parent_row_id": 0,
      "packet_id": 0,
      "LID_evidence": [
        {
          "region_id": 0,
          "probabilities": {
            "zh": 0
          },
          "status": "UNCERTAIN"
        }
      ],
      "VAD_region_ids": [
        0
      ],
      "language_confidence": 0,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 1.9000000000000001,
      "packet_end": 28.900000000000002,
      "raw_text": "哈喽大家",
      "selected_text": "哈喽大家",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            0
          ],
          "packet_id": 0,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        0
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 0
    },
    {
      "start": 3.9800000000000004,
      "end": 6.66,
      "text": "就等了同传字幕1.0正式版",
      "raw_asr_text": "就等了同传字幕1.0正式版",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 0,
      "words": [
        {
          "word": "就",
          "start": 3.9800000000000004,
          "end": 4.44,
          "probability": 0.7802734375
        },
        {
          "word": "等",
          "start": 4.44,
          "end": 4.62,
          "probability": 0.99951171875
        },
        {
          "word": "了",
          "start": 4.62,
          "end": 4.78,
          "probability": 0.4794921875
        },
        {
          "word": "同",
          "start": 4.78,
          "end": 5.0200000000000005,
          "probability": 0.95751953125
        },
        {
          "word": "传",
          "start": 5.0200000000000005,
          "end": 5.22,
          "probability": 0.9765625
        },
        {
          "word": "字",
          "start": 5.22,
          "end": 5.38,
          "probability": 0.98681640625
        },
        {
          "word": "幕",
          "start": 5.38,
          "end": 5.5600000000000005,
          "probability": 0.98974609375
        },
        {
          "word": "1.",
          "start": 5.5600000000000005,
          "end": 5.78,
          "probability": 0.95654296875
        },
        {
          "word": "0",
          "start": 5.92,
          "end": 6.140000000000001,
          "probability": 0.99951171875
        },
        {
          "word": "正",
          "start": 6.140000000000001,
          "end": 6.28,
          "probability": 0.9951171875
        },
        {
          "word": "式",
          "start": 6.28,
          "end": 6.5,
          "probability": 0.99609375
        },
        {
          "word": "版",
          "start": 6.5,
          "end": 6.66,
          "probability": 0.9248046875
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.13841238822645813,
        "compression_ratio": 1.1423611111111112,
        "no_speech_prob": 2.5074888346243718e-11
      },
      "lid_status": "UNCERTAIN",
      "alternatives": [
        {
          "language": "zh",
          "text": "哈喽大家就等了同传字幕1.0正式版今天终于和大家见面了从最初的小工具到现在Android、Mac还有大家呼声超高的Windows三个平台这次全都齐了不管你是睡前抱着手机躺看直播还是在电脑前边干活边挂着生肉台现在都能用同一套伸手的体验让字母自己跟上"
        }
      ],
      "id": 1,
      "review_annotation": "",
      "row_id": "1",
      "bounded_row_id": 1,
      "parent_row_ids": [
        1
      ],
      "parent_row_id": 1,
      "packet_id": 0,
      "LID_evidence": [
        {
          "region_id": 1,
          "probabilities": {
            "zh": 0.9977821012338003
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        1
      ],
      "language_confidence": 0.9977821012338003,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 1.9000000000000001,
      "packet_end": 28.900000000000002,
      "raw_text": "就等了同传字幕1.0正式版",
      "selected_text": "就等了同传字幕1.0正式版",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            1
          ],
          "packet_id": 0,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        0
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 1
    },
    {
      "start": 7.380000000000001,
      "end": 8.94,
      "text": "今天终于和大家见面了",
      "raw_asr_text": "今天终于和大家见面了",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 0,
      "words": [
        {
          "word": "今天",
          "start": 7.380000000000001,
          "end": 7.74,
          "probability": 0.98681640625
        },
        {
          "word": "终",
          "start": 7.74,
          "end": 8.0,
          "probability": 0.99755859375
        },
        {
          "word": "于",
          "start": 8.0,
          "end": 8.12,
          "probability": 0.9990234375
        },
        {
          "word": "和",
          "start": 8.12,
          "end": 8.24,
          "probability": 0.9990234375
        },
        {
          "word": "大家",
          "start": 8.24,
          "end": 8.459999999999999,
          "probability": 1.0
        },
        {
          "word": "见",
          "start": 8.459999999999999,
          "end": 8.6,
          "probability": 0.9990234375
        },
        {
          "word": "面",
          "start": 8.6,
          "end": 8.78,
          "probability": 0.99951171875
        },
        {
          "word": "了",
          "start": 8.78,
          "end": 8.94,
          "probability": 0.99951171875
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.13841238822645813,
        "compression_ratio": 1.1423611111111112,
        "no_speech_prob": 2.5074888346243718e-11
      },
      "lid_status": "UNCERTAIN",
      "alternatives": [
        {
          "language": "zh",
          "text": "哈喽大家就等了同传字幕1.0正式版今天终于和大家见面了从最初的小工具到现在Android、Mac还有大家呼声超高的Windows三个平台这次全都齐了不管你是睡前抱着手机躺看直播还是在电脑前边干活边挂着生肉台现在都能用同一套伸手的体验让字母自己跟上"
        }
      ],
      "id": 2,
      "review_annotation": "",
      "row_id": "2",
      "bounded_row_id": 2,
      "parent_row_ids": [
        2
      ],
      "parent_row_id": 2,
      "packet_id": 0,
      "LID_evidence": [
        {
          "region_id": 2,
          "probabilities": {
            "zh": 0.9975494657244001
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        2
      ],
      "language_confidence": 0.9975494657244001,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 1.9000000000000001,
      "packet_end": 28.900000000000002,
      "raw_text": "今天终于和大家见面了",
      "selected_text": "今天终于和大家见面了",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            2
          ],
          "packet_id": 0,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        0
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 2
    },
    {
      "start": 8.94,
      "end": 11.26,
      "text": "从最初的小工具到现在",
      "raw_asr_text": "从最初的小工具到现在",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 0,
      "words": [
        {
          "word": "从",
          "start": 8.94,
          "end": 9.58,
          "probability": 0.9970703125
        },
        {
          "word": "最",
          "start": 9.58,
          "end": 9.78,
          "probability": 1.0
        },
        {
          "word": "初",
          "start": 9.78,
          "end": 9.92,
          "probability": 0.99951171875
        },
        {
          "word": "的",
          "start": 9.92,
          "end": 10.040000000000001,
          "probability": 1.0
        },
        {
          "word": "小",
          "start": 10.040000000000001,
          "end": 10.18,
          "probability": 1.0
        },
        {
          "word": "工",
          "start": 10.18,
          "end": 10.38,
          "probability": 0.9990234375
        },
        {
          "word": "具",
          "start": 10.38,
          "end": 10.56,
          "probability": 0.99951171875
        },
        {
          "word": "到",
          "start": 10.56,
          "end": 10.9,
          "probability": 0.998046875
        },
        {
          "word": "现在",
          "start": 10.9,
          "end": 11.26,
          "probability": 0.99951171875
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.13841238822645813,
        "compression_ratio": 1.1423611111111112,
        "no_speech_prob": 2.5074888346243718e-11
      },
      "lid_status": "UNCERTAIN",
      "alternatives": [
        {
          "language": "zh",
          "text": "哈喽大家就等了同传字幕1.0正式版今天终于和大家见面了从最初的小工具到现在Android、Mac还有大家呼声超高的Windows三个平台这次全都齐了不管你是睡前抱着手机躺看直播还是在电脑前边干活边挂着生肉台现在都能用同一套伸手的体验让字母自己跟上"
        }
      ],
      "id": 3,
      "review_annotation": "",
      "row_id": "3",
      "bounded_row_id": 3,
      "parent_row_ids": [
        3
      ],
      "parent_row_id": 3,
      "packet_id": 0,
      "LID_evidence": [
        {
          "region_id": 2,
          "probabilities": {
            "zh": 0.9975494657244001
          },
          "status": "STABLE"
        },
        {
          "region_id": 3,
          "probabilities": {
            "zh": 0.9963101983070374
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        2,
        3
      ],
      "language_confidence": 0.9975494657244001,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 1.9000000000000001,
      "packet_end": 28.900000000000002,
      "raw_text": "从最初的小工具到现在",
      "selected_text": "从最初的小工具到现在",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            3
          ],
          "packet_id": 0,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        0
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 3
    },
    {
      "start": 11.26,
      "end": 14.46,
      "text": "Android、Mac还有大家呼声超高的Windows",
      "raw_asr_text": "Android、Mac还有大家呼声超高的Windows",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 0,
      "words": [
        {
          "word": "And",
          "start": 11.26,
          "end": 11.84,
          "probability": 0.96142578125
        },
        {
          "word": "roid、",
          "start": 11.84,
          "end": 12.040000000000001,
          "probability": 0.99951171875
        },
        {
          "word": "Mac",
          "start": 12.34,
          "end": 12.58,
          "probability": 0.9892578125
        },
        {
          "word": "还有",
          "start": 12.58,
          "end": 12.94,
          "probability": 0.97802734375
        },
        {
          "word": "大家",
          "start": 12.94,
          "end": 13.22,
          "probability": 0.9931640625
        },
        {
          "word": "呼",
          "start": 13.22,
          "end": 13.38,
          "probability": 0.9638671875
        },
        {
          "word": "声",
          "start": 13.38,
          "end": 13.56,
          "probability": 0.99658203125
        },
        {
          "word": "超",
          "start": 13.56,
          "end": 13.76,
          "probability": 0.99462890625
        },
        {
          "word": "高",
          "start": 13.76,
          "end": 13.96,
          "probability": 0.9990234375
        },
        {
          "word": "的",
          "start": 13.96,
          "end": 14.06,
          "probability": 1.0
        },
        {
          "word": "Wind",
          "start": 14.06,
          "end": 14.200000000000001,
          "probability": 0.97265625
        },
        {
          "word": "ows",
          "start": 14.200000000000001,
          "end": 14.46,
          "probability": 0.99755859375
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.13841238822645813,
        "compression_ratio": 1.1423611111111112,
        "no_speech_prob": 2.5074888346243718e-11
      },
      "lid_status": "UNCERTAIN",
      "alternatives": [
        {
          "language": "zh",
          "text": "哈喽大家就等了同传字幕1.0正式版今天终于和大家见面了从最初的小工具到现在Android、Mac还有大家呼声超高的Windows三个平台这次全都齐了不管你是睡前抱着手机躺看直播还是在电脑前边干活边挂着生肉台现在都能用同一套伸手的体验让字母自己跟上"
        }
      ],
      "id": 4,
      "review_annotation": "",
      "row_id": "4",
      "bounded_row_id": 4,
      "parent_row_ids": [
        4
      ],
      "parent_row_id": 4,
      "packet_id": 0,
      "LID_evidence": [
        {
          "region_id": 3,
          "probabilities": {
            "zh": 0.9963101983070374
          },
          "status": "STABLE"
        },
        {
          "region_id": 4,
          "probabilities": {
            "zh": 0.9958563223481178
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        3,
        4
      ],
      "language_confidence": 0.9963101983070374,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 1.9000000000000001,
      "packet_end": 28.900000000000002,
      "raw_text": "Android、Mac还有大家呼声超高的Windows",
      "selected_text": "Android、Mac还有大家呼声超高的Windows",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            4
          ],
          "packet_id": 0,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        0
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "independent_ASR": {
        "engine": "whisper-large-v3",
        "model_path": "/Users/sagawa/AI_WORK/cache/huggingface/hub/models--mlx-community--whisper-large-v3",
        "language": "zh",
        "audio_interval": [
          9.26,
          16.46
        ],
        "target_interval": [
          11.26,
          14.46
        ],
        "text": "AndroidMac还有大家呼声超高的Windows",
        "words": [
          {
            "word": "And",
            "start": 11.1,
            "end": 11.68,
            "probability": 0.609375
          },
          {
            "word": "roid",
            "start": 11.68,
            "end": 12.0,
            "probability": 0.9990234375
          },
          {
            "word": " Mac",
            "start": 12.0,
            "end": 12.379999999999999,
            "probability": 0.392822265625
          },
          {
            "word": "还有",
            "start": 12.379999999999999,
            "end": 12.879999999999999,
            "probability": 0.9765625
          },
          {
            "word": "大家",
            "start": 12.879999999999999,
            "end": 13.08,
            "probability": 0.95166015625
          },
          {
            "word": "呼",
            "start": 13.08,
            "end": 13.3,
            "probability": 0.998046875
          },
          {
            "word": "声",
            "start": 13.3,
            "end": 13.52,
            "probability": 0.99755859375
          },
          {
            "word": "超",
            "start": 13.52,
            "end": 13.719999999999999,
            "probability": 0.99560546875
          },
          {
            "word": "高",
            "start": 13.719999999999999,
            "end": 13.899999999999999,
            "probability": 0.9951171875
          },
          {
            "word": "的",
            "start": 13.899999999999999,
            "end": 14.04,
            "probability": 0.99853515625
          },
          {
            "word": "Wind",
            "start": 14.04,
            "end": 14.14,
            "probability": 0.90673828125
          },
          {
            "word": "ows",
            "start": 14.14,
            "end": 14.36,
            "probability": 0.998046875
          }
        ],
        "raw_result": {
          "text": "从最初的小工具到现在Android Mac还有大家呼声超高的Windows三个平台这次",
          "segments": [
            {
              "id": 0,
              "seek": 0,
              "start": 0.0,
              "end": 1.84,
              "text": "从最初的小工具到现在",
              "tokens": [
                50365,
                35630,
                8661,
                28727,
                1546,
                7322,
                23323,
                39806,
                4511,
                25040,
                50467
              ],
              "temperature": 0.0,
              "avg_logprob": -0.15931052910654167,
              "compression_ratio": 0.8942307692307693,
              "no_speech_prob": 0.2348363995552063,
              "words": [
                {
                  "word": "从",
                  "start": 0.0,
                  "end": 0.28,
                  "probability": 0.87451171875
                },
                {
                  "word": "最",
                  "start": 0.28,
                  "end": 0.42,
                  "probability": 0.99755859375
                },
                {
                  "word": "初",
                  "start": 0.42,
                  "end": 0.64,
                  "probability": 0.998046875
                },
                {
                  "word": "的",
                  "start": 0.64,
                  "end": 0.74,
                  "probability": 0.9951171875
                },
                {
                  "word": "小",
                  "start": 0.74,
                  "end": 0.84,
                  "probability": 0.99365234375
                },
                {
                  "word": "工",
                  "start": 0.84,
                  "end": 1.02,
                  "probability": 0.998046875
                },
                {
                  "word": "具",
                  "start": 1.02,
                  "end": 1.26,
                  "probability": 1.0
                },
                {
                  "word": "到",
                  "start": 1.26,
                  "end": 1.54,
                  "probability": 0.91455078125
                },
                {
                  "word": "现在",
                  "start": 1.54,
                  "end": 1.84,
                  "probability": 0.998046875
                }
              ]
            },
            {
              "id": 1,
              "seek": 0,
              "start": 1.84,
              "end": 3.12,
              "text": "Android Mac",
              "tokens": [
                50467,
                5289,
                6490,
                5707,
                50532
              ],
              "temperature": 0.0,
              "avg_logprob": -0.15931052910654167,
              "compression_ratio": 0.8942307692307693,
              "no_speech_prob": 0.2348363995552063,
              "words": [
                {
                  "word": "And",
                  "start": 1.84,
                  "end": 2.42,
                  "probability": 0.609375
                },
                {
                  "word": "roid",
                  "start": 2.42,
                  "end": 2.74,
                  "probability": 0.9990234375
                },
                {
                  "word": " Mac",
                  "start": 2.74,
                  "end": 3.12,
                  "probability": 0.392822265625
                }
              ]
            },
            {
              "id": 2,
              "seek": 0,
              "start": 3.12,
              "end": 5.1,
              "text": "还有大家呼声超高的Windows",
              "tokens": [
                50532,
                35091,
                6868,
                38718,
                32045,
                19869,
                12979,
                1546,
                45813,
                1509,
                50627
              ],
              "temperature": 0.0,
              "avg_logprob": -0.15931052910654167,
              "compression_ratio": 0.8942307692307693,
              "no_speech_prob": 0.2348363995552063,
              "words": [
                {
                  "word": "还有",
                  "start": 3.12,
                  "end": 3.62,
                  "probability": 0.9765625
                },
                {
                  "word": "大家",
                  "start": 3.62,
                  "end": 3.82,
                  "probability": 0.95166015625
                },
                {
                  "word": "呼",
                  "start": 3.82,
                  "end": 4.04,
                  "probability": 0.998046875
                },
                {
                  "word": "声",
                  "start": 4.04,
                  "end": 4.26,
                  "probability": 0.99755859375
                },
                {
                  "word": "超",
                  "start": 4.26,
                  "end": 4.46,
                  "probability": 0.99560546875
                },
                {
                  "word": "高",
                  "start": 4.46,
                  "end": 4.64,
                  "probability": 0.9951171875
                },
                {
                  "word": "的",
                  "start": 4.64,
                  "end": 4.78,
                  "probability": 0.99853515625
                },
                {
                  "word": "Wind",
                  "start": 4.78,
                  "end": 4.88,
                  "probability": 0.90673828125
                },
                {
                  "word": "ows",
                  "start": 4.88,
                  "end": 5.1,
                  "probability": 0.998046875
                }
              ]
            },
            {
              "id": 3,
              "seek": 0,
              "start": 5.1,
              "end": 6.38,
              "text": "三个平台",
              "tokens": [
                50627,
                10960,
                7549,
                16716,
                15433,
                50689
              ],
              "temperature": 0.0,
              "avg_logprob": -0.15931052910654167,
              "compression_ratio": 0.8942307692307693,
              "no_speech_prob": 0.2348363995552063,
              "words": [
                {
                  "word": "三",
                  "start": 5.1,
                  "end": 5.84,
                  "probability": 0.9775390625
                },
                {
                  "word": "个",
                  "start": 5.84,
                  "end": 5.96,
                  "probability": 0.99951171875
                },
                {
                  "word": "平",
                  "start": 5.96,
                  "end": 6.1,
                  "probability": 0.9990234375
                },
                {
                  "word": "台",
                  "start": 6.1,
                  "end": 6.38,
                  "probability": 0.99951171875
                }
              ]
            },
            {
              "id": 4,
              "seek": 0,
              "start": 6.38,
              "end": 6.84,
              "text": "这次",
              "tokens": [
                50689,
                5562,
                9487,
                50715
              ],
              "temperature": 0.0,
              "avg_logprob": -0.15931052910654167,
              "compression_ratio": 0.8942307692307693,
              "no_speech_prob": 0.2348363995552063,
              "words": [
                {
                  "word": "这",
                  "start": 6.38,
                  "end": 6.72,
                  "probability": 0.939453125
                },
                {
                  "word": "次",
                  "start": 6.72,
                  "end": 6.84,
                  "probability": 0.99951171875
                }
              ]
            }
          ],
          "language": "zh",
          "duration": 7.2
        },
        "status": "COMPLETE",
        "requested_language": "zh"
      },
      "index": 4
    },
    {
      "start": 14.46,
      "end": 17.2,
      "text": "三个平台这次全都齐了",
      "raw_asr_text": "三个平台这次全都齐了",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 0,
      "words": [
        {
          "word": "三",
          "start": 14.46,
          "end": 15.18,
          "probability": 0.9921875
        },
        {
          "word": "个",
          "start": 15.18,
          "end": 15.34,
          "probability": 0.99951171875
        },
        {
          "word": "平",
          "start": 15.34,
          "end": 15.46,
          "probability": 1.0
        },
        {
          "word": "台",
          "start": 15.46,
          "end": 15.700000000000001,
          "probability": 0.99951171875
        },
        {
          "word": "这",
          "start": 15.700000000000001,
          "end": 16.06,
          "probability": 0.99072265625
        },
        {
          "word": "次",
          "start": 16.06,
          "end": 16.259999999999998,
          "probability": 1.0
        },
        {
          "word": "全",
          "start": 16.259999999999998,
          "end": 16.62,
          "probability": 0.99951171875
        },
        {
          "word": "都",
          "start": 16.62,
          "end": 16.84,
          "probability": 1.0
        },
        {
          "word": "齐",
          "start": 16.84,
          "end": 17.02,
          "probability": 0.9482421875
        },
        {
          "word": "了",
          "start": 17.02,
          "end": 17.2,
          "probability": 0.99951171875
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.13841238822645813,
        "compression_ratio": 1.1423611111111112,
        "no_speech_prob": 2.5074888346243718e-11
      },
      "lid_status": "UNCERTAIN",
      "alternatives": [
        {
          "language": "zh",
          "text": "哈喽大家就等了同传字幕1.0正式版今天终于和大家见面了从最初的小工具到现在Android、Mac还有大家呼声超高的Windows三个平台这次全都齐了不管你是睡前抱着手机躺看直播还是在电脑前边干活边挂着生肉台现在都能用同一套伸手的体验让字母自己跟上"
        }
      ],
      "id": 5,
      "review_annotation": "",
      "row_id": "5",
      "bounded_row_id": 5,
      "parent_row_ids": [
        5
      ],
      "parent_row_id": 5,
      "packet_id": 0,
      "LID_evidence": [
        {
          "region_id": 4,
          "probabilities": {
            "zh": 0.9958563223481178
          },
          "status": "STABLE"
        },
        {
          "region_id": 5,
          "probabilities": {
            "zh": 0.9980121321148343
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        4,
        5
      ],
      "language_confidence": 0.9980121321148343,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 1.9000000000000001,
      "packet_end": 28.900000000000002,
      "raw_text": "三个平台这次全都齐了",
      "selected_text": "三个平台这次全都齐了",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            5
          ],
          "packet_id": 0,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        0
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 5
    },
    {
      "start": 18.18,
      "end": 20.4,
      "text": "不管你是睡前抱着手机躺看直播",
      "raw_asr_text": "不管你是睡前抱着手机躺看直播",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 0,
      "words": [
        {
          "word": "不管",
          "start": 18.18,
          "end": 18.54,
          "probability": 0.9921875
        },
        {
          "word": "你是",
          "start": 18.54,
          "end": 18.779999999999998,
          "probability": 0.99951171875
        },
        {
          "word": "睡",
          "start": 18.779999999999998,
          "end": 18.959999999999997,
          "probability": 0.99853515625
        },
        {
          "word": "前",
          "start": 18.959999999999997,
          "end": 19.139999999999997,
          "probability": 1.0
        },
        {
          "word": "抱",
          "start": 19.139999999999997,
          "end": 19.32,
          "probability": 0.99951171875
        },
        {
          "word": "着",
          "start": 19.32,
          "end": 19.459999999999997,
          "probability": 0.99951171875
        },
        {
          "word": "手",
          "start": 19.459999999999997,
          "end": 19.599999999999998,
          "probability": 0.99951171875
        },
        {
          "word": "机",
          "start": 19.599999999999998,
          "end": 19.74,
          "probability": 0.99951171875
        },
        {
          "word": "躺",
          "start": 19.74,
          "end": 19.86,
          "probability": 0.9921875
        },
        {
          "word": "看",
          "start": 19.86,
          "end": 20.06,
          "probability": 0.97216796875
        },
        {
          "word": "直",
          "start": 20.06,
          "end": 20.2,
          "probability": 0.9970703125
        },
        {
          "word": "播",
          "start": 20.2,
          "end": 20.4,
          "probability": 1.0
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.13841238822645813,
        "compression_ratio": 1.1423611111111112,
        "no_speech_prob": 2.5074888346243718e-11
      },
      "lid_status": "UNCERTAIN",
      "alternatives": [
        {
          "language": "zh",
          "text": "哈喽大家就等了同传字幕1.0正式版今天终于和大家见面了从最初的小工具到现在Android、Mac还有大家呼声超高的Windows三个平台这次全都齐了不管你是睡前抱着手机躺看直播还是在电脑前边干活边挂着生肉台现在都能用同一套伸手的体验让字母自己跟上"
        }
      ],
      "id": 6,
      "review_annotation": "",
      "row_id": "6",
      "bounded_row_id": 6,
      "parent_row_ids": [
        6
      ],
      "parent_row_id": 6,
      "packet_id": 0,
      "LID_evidence": [
        {
          "region_id": 6,
          "probabilities": {
            "zh": 0.9978591246264321
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        6
      ],
      "language_confidence": 0.9978591246264321,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 1.9000000000000001,
      "packet_end": 28.900000000000002,
      "raw_text": "不管你是睡前抱着手机躺看直播",
      "selected_text": "不管你是睡前抱着手机躺看直播",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            6
          ],
          "packet_id": 0,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        0
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 6
    },
    {
      "start": 20.4,
      "end": 23.36,
      "text": "还是在电脑前边干活边挂着生肉台",
      "raw_asr_text": "还是在电脑前边干活边挂着生肉台",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 0,
      "words": [
        {
          "word": "还是",
          "start": 20.4,
          "end": 21.06,
          "probability": 0.998046875
        },
        {
          "word": "在",
          "start": 21.06,
          "end": 21.22,
          "probability": 1.0
        },
        {
          "word": "电",
          "start": 21.22,
          "end": 21.36,
          "probability": 0.99609375
        },
        {
          "word": "脑",
          "start": 21.36,
          "end": 21.52,
          "probability": 1.0
        },
        {
          "word": "前",
          "start": 21.52,
          "end": 21.68,
          "probability": 0.99951171875
        },
        {
          "word": "边",
          "start": 21.68,
          "end": 21.84,
          "probability": 0.99560546875
        },
        {
          "word": "干",
          "start": 21.84,
          "end": 22.0,
          "probability": 0.998046875
        },
        {
          "word": "活",
          "start": 22.0,
          "end": 22.24,
          "probability": 0.99951171875
        },
        {
          "word": "边",
          "start": 22.24,
          "end": 22.619999999999997,
          "probability": 0.99609375
        },
        {
          "word": "挂",
          "start": 22.619999999999997,
          "end": 22.759999999999998,
          "probability": 0.9990234375
        },
        {
          "word": "着",
          "start": 22.759999999999998,
          "end": 22.9,
          "probability": 0.9990234375
        },
        {
          "word": "生",
          "start": 22.9,
          "end": 23.02,
          "probability": 0.42822265625
        },
        {
          "word": "肉",
          "start": 23.02,
          "end": 23.18,
          "probability": 0.9814453125
        },
        {
          "word": "台",
          "start": 23.18,
          "end": 23.36,
          "probability": 0.96240234375
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.13841238822645813,
        "compression_ratio": 1.1423611111111112,
        "no_speech_prob": 2.5074888346243718e-11
      },
      "lid_status": "UNCERTAIN",
      "alternatives": [
        {
          "language": "zh",
          "text": "哈喽大家就等了同传字幕1.0正式版今天终于和大家见面了从最初的小工具到现在Android、Mac还有大家呼声超高的Windows三个平台这次全都齐了不管你是睡前抱着手机躺看直播还是在电脑前边干活边挂着生肉台现在都能用同一套伸手的体验让字母自己跟上"
        }
      ],
      "id": 7,
      "review_annotation": "",
      "row_id": "7",
      "bounded_row_id": 7,
      "parent_row_ids": [
        7
      ],
      "parent_row_id": 7,
      "packet_id": 0,
      "LID_evidence": [
        {
          "region_id": 6,
          "probabilities": {
            "zh": 0.9978591246264321
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        6
      ],
      "language_confidence": 0.9978591246264321,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 1.9000000000000001,
      "packet_end": 28.900000000000002,
      "raw_text": "还是在电脑前边干活边挂着生肉台",
      "selected_text": "还是在电脑前边干活边挂着生肉台",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            7
          ],
          "packet_id": 0,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        0
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 7
    },
    {
      "start": 24.16,
      "end": 26.299999999999997,
      "text": "伸手",
      "raw_asr_text": "现在都能用同一套伸手的体验",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 0,
      "words": [],
      "asr_metrics": {
        "avg_logprob": -0.13841238822645813,
        "compression_ratio": 1.1423611111111112,
        "no_speech_prob": 2.5074888346243718e-11
      },
      "lid_status": "UNCERTAIN",
      "alternatives": [
        {
          "language": "zh",
          "text": "哈喽大家就等了同传字幕1.0正式版今天终于和大家见面了从最初的小工具到现在Android、Mac还有大家呼声超高的Windows三个平台这次全都齐了不管你是睡前抱着手机躺看直播还是在电脑前边干活边挂着生肉台现在都能用同一套伸手的体验让字母自己跟上"
        }
      ],
      "id": 8,
      "review_annotation": "待確認",
      "row_id": "8",
      "bounded_row_id": 8,
      "parent_row_ids": [
        8
      ],
      "parent_row_id": 8,
      "packet_id": 0,
      "LID_evidence": [
        {
          "region_id": 7,
          "probabilities": {
            "zh": 0.997811065779792
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        7
      ],
      "language_confidence": 0.997811065779792,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 1.9000000000000001,
      "packet_end": 28.900000000000002,
      "raw_text": "现在都能用同一套伸手的体验",
      "selected_text": "伸手",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            8
          ],
          "packet_id": 0,
          "engine": "mlx-whisper"
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "8",
          "before": "现在都能用同一套伸手的体验",
          "after": "伸手",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "伸手体验语义不通，疑似同音误识",
          "context_evidence": "上下文为跨平台体验，应为'伸手'或'舒适'等",
          "independent_ASR": {
            "engine": "whisper-large-v3",
            "model_path": "/Users/sagawa/AI_WORK/cache/huggingface/hub/models--mlx-community--whisper-large-v3",
            "language": "zh",
            "audio_interval": [
              22.16,
              28.299999999999997
            ],
            "target_interval": [
              24.16,
              26.299999999999997
            ],
            "text": "现在都能用同一套身手的体验",
            "words": [
              {
                "word": "现在",
                "start": 24.16,
                "end": 24.48,
                "probability": 0.982421875
              },
              {
                "word": "都",
                "start": 24.48,
                "end": 24.8,
                "probability": 0.98193359375
              },
              {
                "word": "能",
                "start": 24.8,
                "end": 24.9,
                "probability": 0.9990234375
              },
              {
                "word": "用",
                "start": 24.9,
                "end": 25.06,
                "probability": 0.9990234375
              },
              {
                "word": "同",
                "start": 25.06,
                "end": 25.22,
                "probability": 0.99658203125
              },
              {
                "word": "一",
                "start": 25.22,
                "end": 25.4,
                "probability": 0.99951171875
              },
              {
                "word": "套",
                "start": 25.4,
                "end": 25.56,
                "probability": 1.0
              },
              {
                "word": "身",
                "start": 25.56,
                "end": 25.66,
                "probability": 0.5009765625
              },
              {
                "word": "手",
                "start": 25.66,
                "end": 25.82,
                "probability": 0.99853515625
              },
              {
                "word": "的",
                "start": 25.82,
                "end": 25.88,
                "probability": 0.998046875
              },
              {
                "word": "体",
                "start": 25.88,
                "end": 26.02,
                "probability": 0.99951171875
              },
              {
                "word": "验",
                "start": 26.02,
                "end": 26.240000000000002,
                "probability": 1.0
              }
            ],
            "raw_result": {
              "text": "边挂着生肉台现在都能用同一套身手的体验让子母自己跟上",
              "segments": [
                {
                  "id": 0,
                  "seek": 0,
                  "start": 0.0,
                  "end": 1.12,
                  "text": "边挂着生肉台",
                  "tokens": [
                    50365,
                    40503,
                    8501,
                    224,
                    20708,
                    8244,
                    31980,
                    15433,
                    50427
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.2176045950721292,
                  "compression_ratio": 0.8764044943820225,
                  "no_speech_prob": 0.6835574507713318,
                  "words": [
                    {
                      "word": "边",
                      "start": 0.0,
                      "end": 0.4,
                      "probability": 0.1392822265625
                    },
                    {
                      "word": "挂",
                      "start": 0.4,
                      "end": 0.62,
                      "probability": 0.9951171875
                    },
                    {
                      "word": "着",
                      "start": 0.62,
                      "end": 0.7,
                      "probability": 0.99365234375
                    },
                    {
                      "word": "生",
                      "start": 0.7,
                      "end": 0.82,
                      "probability": 0.59228515625
                    },
                    {
                      "word": "肉",
                      "start": 0.82,
                      "end": 0.98,
                      "probability": 0.99658203125
                    },
                    {
                      "word": "台",
                      "start": 0.98,
                      "end": 1.12,
                      "probability": 0.88671875
                    }
                  ]
                },
                {
                  "id": 1,
                  "seek": 0,
                  "start": 1.9999999999999996,
                  "end": 4.08,
                  "text": "现在都能用同一套身手的体验",
                  "tokens": [
                    50427,
                    25040,
                    7182,
                    8225,
                    9254,
                    13089,
                    2257,
                    1881,
                    245,
                    19847,
                    11389,
                    1546,
                    29485,
                    49657,
                    234,
                    50571
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.2176045950721292,
                  "compression_ratio": 0.8764044943820225,
                  "no_speech_prob": 0.6835574507713318,
                  "words": [
                    {
                      "word": "现在",
                      "start": 1.9999999999999996,
                      "end": 2.32,
                      "probability": 0.982421875
                    },
                    {
                      "word": "都",
                      "start": 2.32,
                      "end": 2.64,
                      "probability": 0.98193359375
                    },
                    {
                      "word": "能",
                      "start": 2.64,
                      "end": 2.74,
                      "probability": 0.9990234375
                    },
                    {
                      "word": "用",
                      "start": 2.74,
                      "end": 2.9,
                      "probability": 0.9990234375
                    },
                    {
                      "word": "同",
                      "start": 2.9,
                      "end": 3.06,
                      "probability": 0.99658203125
                    },
                    {
                      "word": "一",
                      "start": 3.06,
                      "end": 3.24,
                      "probability": 0.99951171875
                    },
                    {
                      "word": "套",
                      "start": 3.24,
                      "end": 3.4,
                      "probability": 1.0
                    },
                    {
                      "word": "身",
                      "start": 3.4,
                      "end": 3.5,
                      "probability": 0.5009765625
                    },
                    {
                      "word": "手",
                      "start": 3.5,
                      "end": 3.66,
                      "probability": 0.99853515625
                    },
                    {
                      "word": "的",
                      "start": 3.66,
                      "end": 3.72,
                      "probability": 0.998046875
                    },
                    {
                      "word": "体",
                      "start": 3.72,
                      "end": 3.86,
                      "probability": 0.99951171875
                    },
                    {
                      "word": "验",
                      "start": 3.86,
                      "end": 4.08,
                      "probability": 1.0
                    }
                  ]
                },
                {
                  "id": 2,
                  "seek": 0,
                  "start": 4.08,
                  "end": 5.62,
                  "text": "让子母自己跟上",
                  "tokens": [
                    50571,
                    33650,
                    7626,
                    35744,
                    17645,
                    9678,
                    5708,
                    50650
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.2176045950721292,
                  "compression_ratio": 0.8764044943820225,
                  "no_speech_prob": 0.6835574507713318,
                  "words": [
                    {
                      "word": "让",
                      "start": 4.08,
                      "end": 4.66,
                      "probability": 0.9970703125
                    },
                    {
                      "word": "子",
                      "start": 4.66,
                      "end": 4.78,
                      "probability": 0.59375
                    },
                    {
                      "word": "母",
                      "start": 4.78,
                      "end": 5.0,
                      "probability": 0.65966796875
                    },
                    {
                      "word": "自己",
                      "start": 5.0,
                      "end": 5.18,
                      "probability": 0.9990234375
                    },
                    {
                      "word": "跟",
                      "start": 5.18,
                      "end": 5.4,
                      "probability": 0.9990234375
                    },
                    {
                      "word": "上",
                      "start": 5.4,
                      "end": 5.62,
                      "probability": 1.0
                    }
                  ]
                }
              ],
              "language": "zh",
              "duration": 6.14
            },
            "status": "COMPLETE",
            "requested_language": "zh"
          },
          "suppression_proposed": null
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "8",
          "before": "伸手",
          "after": "伸手",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "Context mismatch: '伸手' (reach out) vs '同一套身手的体验' (same experience)",
          "context_evidence": "ASR: '现在都能用同一套身手的体验'",
          "independent_ASR": {
            "engine": "whisper-large-v3",
            "model_path": "/Users/sagawa/AI_WORK/cache/huggingface/hub/models--mlx-community--whisper-large-v3",
            "language": "zh",
            "audio_interval": [
              22.16,
              28.299999999999997
            ],
            "target_interval": [
              24.16,
              26.299999999999997
            ],
            "text": "现在都能用同一套身手的体验",
            "words": [
              {
                "word": "现在",
                "start": 24.16,
                "end": 24.48,
                "probability": 0.982421875
              },
              {
                "word": "都",
                "start": 24.48,
                "end": 24.8,
                "probability": 0.98193359375
              },
              {
                "word": "能",
                "start": 24.8,
                "end": 24.9,
                "probability": 0.9990234375
              },
              {
                "word": "用",
                "start": 24.9,
                "end": 25.06,
                "probability": 0.9990234375
              },
              {
                "word": "同",
                "start": 25.06,
                "end": 25.22,
                "probability": 0.99658203125
              },
              {
                "word": "一",
                "start": 25.22,
                "end": 25.4,
                "probability": 0.99951171875
              },
              {
                "word": "套",
                "start": 25.4,
                "end": 25.56,
                "probability": 1.0
              },
              {
                "word": "身",
                "start": 25.56,
                "end": 25.66,
                "probability": 0.5009765625
              },
              {
                "word": "手",
                "start": 25.66,
                "end": 25.82,
                "probability": 0.99853515625
              },
              {
                "word": "的",
                "start": 25.82,
                "end": 25.88,
                "probability": 0.998046875
              },
              {
                "word": "体",
                "start": 25.88,
                "end": 26.02,
                "probability": 0.99951171875
              },
              {
                "word": "验",
                "start": 26.02,
                "end": 26.240000000000002,
                "probability": 1.0
              }
            ],
            "raw_result": {
              "text": "边挂着生肉台现在都能用同一套身手的体验让子母自己跟上",
              "segments": [
                {
                  "id": 0,
                  "seek": 0,
                  "start": 0.0,
                  "end": 1.12,
                  "text": "边挂着生肉台",
                  "tokens": [
                    50365,
                    40503,
                    8501,
                    224,
                    20708,
                    8244,
                    31980,
                    15433,
                    50427
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.2176045950721292,
                  "compression_ratio": 0.8764044943820225,
                  "no_speech_prob": 0.6835574507713318,
                  "words": [
                    {
                      "word": "边",
                      "start": 0.0,
                      "end": 0.4,
                      "probability": 0.1392822265625
                    },
                    {
                      "word": "挂",
                      "start": 0.4,
                      "end": 0.62,
                      "probability": 0.9951171875
                    },
                    {
                      "word": "着",
                      "start": 0.62,
                      "end": 0.7,
                      "probability": 0.99365234375
                    },
                    {
                      "word": "生",
                      "start": 0.7,
                      "end": 0.82,
                      "probability": 0.59228515625
                    },
                    {
                      "word": "肉",
                      "start": 0.82,
                      "end": 0.98,
                      "probability": 0.99658203125
                    },
                    {
                      "word": "台",
                      "start": 0.98,
                      "end": 1.12,
                      "probability": 0.88671875
                    }
                  ]
                },
                {
                  "id": 1,
                  "seek": 0,
                  "start": 1.9999999999999996,
                  "end": 4.08,
                  "text": "现在都能用同一套身手的体验",
                  "tokens": [
                    50427,
                    25040,
                    7182,
                    8225,
                    9254,
                    13089,
                    2257,
                    1881,
                    245,
                    19847,
                    11389,
                    1546,
                    29485,
                    49657,
                    234,
                    50571
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.2176045950721292,
                  "compression_ratio": 0.8764044943820225,
                  "no_speech_prob": 0.6835574507713318,
                  "words": [
                    {
                      "word": "现在",
                      "start": 1.9999999999999996,
                      "end": 2.32,
                      "probability": 0.982421875
                    },
                    {
                      "word": "都",
                      "start": 2.32,
                      "end": 2.64,
                      "probability": 0.98193359375
                    },
                    {
                      "word": "能",
                      "start": 2.64,
                      "end": 2.74,
                      "probability": 0.9990234375
                    },
                    {
                      "word": "用",
                      "start": 2.74,
                      "end": 2.9,
                      "probability": 0.9990234375
                    },
                    {
                      "word": "同",
                      "start": 2.9,
                      "end": 3.06,
                      "probability": 0.99658203125
                    },
                    {
                      "word": "一",
                      "start": 3.06,
                      "end": 3.24,
                      "probability": 0.99951171875
                    },
                    {
                      "word": "套",
                      "start": 3.24,
                      "end": 3.4,
                      "probability": 1.0
                    },
                    {
                      "word": "身",
                      "start": 3.4,
                      "end": 3.5,
                      "probability": 0.5009765625
                    },
                    {
                      "word": "手",
                      "start": 3.5,
                      "end": 3.66,
                      "probability": 0.99853515625
                    },
                    {
                      "word": "的",
                      "start": 3.66,
                      "end": 3.72,
                      "probability": 0.998046875
                    },
                    {
                      "word": "体",
                      "start": 3.72,
                      "end": 3.86,
                      "probability": 0.99951171875
                    },
                    {
                      "word": "验",
                      "start": 3.86,
                      "end": 4.08,
                      "probability": 1.0
                    }
                  ]
                },
                {
                  "id": 2,
                  "seek": 0,
                  "start": 4.08,
                  "end": 5.62,
                  "text": "让子母自己跟上",
                  "tokens": [
                    50571,
                    33650,
                    7626,
                    35744,
                    17645,
                    9678,
                    5708,
                    50650
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.2176045950721292,
                  "compression_ratio": 0.8764044943820225,
                  "no_speech_prob": 0.6835574507713318,
                  "words": [
                    {
                      "word": "让",
                      "start": 4.08,
                      "end": 4.66,
                      "probability": 0.9970703125
                    },
                    {
                      "word": "子",
                      "start": 4.66,
                      "end": 4.78,
                      "probability": 0.59375
                    },
                    {
                      "word": "母",
                      "start": 4.78,
                      "end": 5.0,
                      "probability": 0.65966796875
                    },
                    {
                      "word": "自己",
                      "start": 5.0,
                      "end": 5.18,
                      "probability": 0.9990234375
                    },
                    {
                      "word": "跟",
                      "start": 5.18,
                      "end": 5.4,
                      "probability": 0.9990234375
                    },
                    {
                      "word": "上",
                      "start": 5.4,
                      "end": 5.62,
                      "probability": 1.0
                    }
                  ]
                }
              ],
              "language": "zh",
              "duration": 6.14
            },
            "status": "COMPLETE",
            "requested_language": "zh"
          },
          "suppression_proposed": null
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "8",
          "before": "伸手",
          "after": "伸手",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "Context mismatch: '伸手' (reach out) vs '同一套身手的体验' (same experience)",
          "context_evidence": "ASR: '现在都能用同一套身手的体验'",
          "independent_ASR": {
            "engine": "whisper-large-v3",
            "model_path": "/Users/sagawa/AI_WORK/cache/huggingface/hub/models--mlx-community--whisper-large-v3",
            "language": "zh",
            "audio_interval": [
              22.16,
              28.299999999999997
            ],
            "target_interval": [
              24.16,
              26.299999999999997
            ],
            "text": "现在都能用同一套身手的体验",
            "words": [
              {
                "word": "现在",
                "start": 24.16,
                "end": 24.48,
                "probability": 0.982421875
              },
              {
                "word": "都",
                "start": 24.48,
                "end": 24.8,
                "probability": 0.98193359375
              },
              {
                "word": "能",
                "start": 24.8,
                "end": 24.9,
                "probability": 0.9990234375
              },
              {
                "word": "用",
                "start": 24.9,
                "end": 25.06,
                "probability": 0.9990234375
              },
              {
                "word": "同",
                "start": 25.06,
                "end": 25.22,
                "probability": 0.99658203125
              },
              {
                "word": "一",
                "start": 25.22,
                "end": 25.4,
                "probability": 0.99951171875
              },
              {
                "word": "套",
                "start": 25.4,
                "end": 25.56,
                "probability": 1.0
              },
              {
                "word": "身",
                "start": 25.56,
                "end": 25.66,
                "probability": 0.5009765625
              },
              {
                "word": "手",
                "start": 25.66,
                "end": 25.82,
                "probability": 0.99853515625
              },
              {
                "word": "的",
                "start": 25.82,
                "end": 25.88,
                "probability": 0.998046875
              },
              {
                "word": "体",
                "start": 25.88,
                "end": 26.02,
                "probability": 0.99951171875
              },
              {
                "word": "验",
                "start": 26.02,
                "end": 26.240000000000002,
                "probability": 1.0
              }
            ],
            "raw_result": {
              "text": "边挂着生肉台现在都能用同一套身手的体验让子母自己跟上",
              "segments": [
                {
                  "id": 0,
                  "seek": 0,
                  "start": 0.0,
                  "end": 1.12,
                  "text": "边挂着生肉台",
                  "tokens": [
                    50365,
                    40503,
                    8501,
                    224,
                    20708,
                    8244,
                    31980,
                    15433,
                    50427
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.2176045950721292,
                  "compression_ratio": 0.8764044943820225,
                  "no_speech_prob": 0.6835574507713318,
                  "words": [
                    {
                      "word": "边",
                      "start": 0.0,
                      "end": 0.4,
                      "probability": 0.1392822265625
                    },
                    {
                      "word": "挂",
                      "start": 0.4,
                      "end": 0.62,
                      "probability": 0.9951171875
                    },
                    {
                      "word": "着",
                      "start": 0.62,
                      "end": 0.7,
                      "probability": 0.99365234375
                    },
                    {
                      "word": "生",
                      "start": 0.7,
                      "end": 0.82,
                      "probability": 0.59228515625
                    },
                    {
                      "word": "肉",
                      "start": 0.82,
                      "end": 0.98,
                      "probability": 0.99658203125
                    },
                    {
                      "word": "台",
                      "start": 0.98,
                      "end": 1.12,
                      "probability": 0.88671875
                    }
                  ]
                },
                {
                  "id": 1,
                  "seek": 0,
                  "start": 1.9999999999999996,
                  "end": 4.08,
                  "text": "现在都能用同一套身手的体验",
                  "tokens": [
                    50427,
                    25040,
                    7182,
                    8225,
                    9254,
                    13089,
                    2257,
                    1881,
                    245,
                    19847,
                    11389,
                    1546,
                    29485,
                    49657,
                    234,
                    50571
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.2176045950721292,
                  "compression_ratio": 0.8764044943820225,
                  "no_speech_prob": 0.6835574507713318,
                  "words": [
                    {
                      "word": "现在",
                      "start": 1.9999999999999996,
                      "end": 2.32,
                      "probability": 0.982421875
                    },
                    {
                      "word": "都",
                      "start": 2.32,
                      "end": 2.64,
                      "probability": 0.98193359375
                    },
                    {
                      "word": "能",
                      "start": 2.64,
                      "end": 2.74,
                      "probability": 0.9990234375
                    },
                    {
                      "word": "用",
                      "start": 2.74,
                      "end": 2.9,
                      "probability": 0.9990234375
                    },
                    {
                      "word": "同",
                      "start": 2.9,
                      "end": 3.06,
                      "probability": 0.99658203125
                    },
                    {
                      "word": "一",
                      "start": 3.06,
                      "end": 3.24,
                      "probability": 0.99951171875
                    },
                    {
                      "word": "套",
                      "start": 3.24,
                      "end": 3.4,
                      "probability": 1.0
                    },
                    {
                      "word": "身",
                      "start": 3.4,
                      "end": 3.5,
                      "probability": 0.5009765625
                    },
                    {
                      "word": "手",
                      "start": 3.5,
                      "end": 3.66,
                      "probability": 0.99853515625
                    },
                    {
                      "word": "的",
                      "start": 3.66,
                      "end": 3.72,
                      "probability": 0.998046875
                    },
                    {
                      "word": "体",
                      "start": 3.72,
                      "end": 3.86,
                      "probability": 0.99951171875
                    },
                    {
                      "word": "验",
                      "start": 3.86,
                      "end": 4.08,
                      "probability": 1.0
                    }
                  ]
                },
                {
                  "id": 2,
                  "seek": 0,
                  "start": 4.08,
                  "end": 5.62,
                  "text": "让子母自己跟上",
                  "tokens": [
                    50571,
                    33650,
                    7626,
                    35744,
                    17645,
                    9678,
                    5708,
                    50650
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.2176045950721292,
                  "compression_ratio": 0.8764044943820225,
                  "no_speech_prob": 0.6835574507713318,
                  "words": [
                    {
                      "word": "让",
                      "start": 4.08,
                      "end": 4.66,
                      "probability": 0.9970703125
                    },
                    {
                      "word": "子",
                      "start": 4.66,
                      "end": 4.78,
                      "probability": 0.59375
                    },
                    {
                      "word": "母",
                      "start": 4.78,
                      "end": 5.0,
                      "probability": 0.65966796875
                    },
                    {
                      "word": "自己",
                      "start": 5.0,
                      "end": 5.18,
                      "probability": 0.9990234375
                    },
                    {
                      "word": "跟",
                      "start": 5.18,
                      "end": 5.4,
                      "probability": 0.9990234375
                    },
                    {
                      "word": "上",
                      "start": 5.4,
                      "end": 5.62,
                      "probability": 1.0
                    }
                  ]
                }
              ],
              "language": "zh",
              "duration": 6.14
            },
            "status": "COMPLETE",
            "requested_language": "zh"
          },
          "suppression_proposed": null
        }
      ],
      "language": "zh",
      "packet_ids": [
        0
      ],
      "classification": "BEST_SUPPORTED_WITH_REVIEW",
      "independent_ASR": {
        "engine": "whisper-large-v3",
        "model_path": "/Users/sagawa/AI_WORK/cache/huggingface/hub/models--mlx-community--whisper-large-v3",
        "language": "zh",
        "audio_interval": [
          22.16,
          28.299999999999997
        ],
        "target_interval": [
          24.16,
          26.299999999999997
        ],
        "text": "现在都能用同一套身手的体验",
        "words": [
          {
            "word": "现在",
            "start": 24.16,
            "end": 24.48,
            "probability": 0.982421875
          },
          {
            "word": "都",
            "start": 24.48,
            "end": 24.8,
            "probability": 0.98193359375
          },
          {
            "word": "能",
            "start": 24.8,
            "end": 24.9,
            "probability": 0.9990234375
          },
          {
            "word": "用",
            "start": 24.9,
            "end": 25.06,
            "probability": 0.9990234375
          },
          {
            "word": "同",
            "start": 25.06,
            "end": 25.22,
            "probability": 0.99658203125
          },
          {
            "word": "一",
            "start": 25.22,
            "end": 25.4,
            "probability": 0.99951171875
          },
          {
            "word": "套",
            "start": 25.4,
            "end": 25.56,
            "probability": 1.0
          },
          {
            "word": "身",
            "start": 25.56,
            "end": 25.66,
            "probability": 0.5009765625
          },
          {
            "word": "手",
            "start": 25.66,
            "end": 25.82,
            "probability": 0.99853515625
          },
          {
            "word": "的",
            "start": 25.82,
            "end": 25.88,
            "probability": 0.998046875
          },
          {
            "word": "体",
            "start": 25.88,
            "end": 26.02,
            "probability": 0.99951171875
          },
          {
            "word": "验",
            "start": 26.02,
            "end": 26.240000000000002,
            "probability": 1.0
          }
        ],
        "raw_result": {
          "text": "边挂着生肉台现在都能用同一套身手的体验让子母自己跟上",
          "segments": [
            {
              "id": 0,
              "seek": 0,
              "start": 0.0,
              "end": 1.12,
              "text": "边挂着生肉台",
              "tokens": [
                50365,
                40503,
                8501,
                224,
                20708,
                8244,
                31980,
                15433,
                50427
              ],
              "temperature": 0.0,
              "avg_logprob": -0.2176045950721292,
              "compression_ratio": 0.8764044943820225,
              "no_speech_prob": 0.6835574507713318,
              "words": [
                {
                  "word": "边",
                  "start": 0.0,
                  "end": 0.4,
                  "probability": 0.1392822265625
                },
                {
                  "word": "挂",
                  "start": 0.4,
                  "end": 0.62,
                  "probability": 0.9951171875
                },
                {
                  "word": "着",
                  "start": 0.62,
                  "end": 0.7,
                  "probability": 0.99365234375
                },
                {
                  "word": "生",
                  "start": 0.7,
                  "end": 0.82,
                  "probability": 0.59228515625
                },
                {
                  "word": "肉",
                  "start": 0.82,
                  "end": 0.98,
                  "probability": 0.99658203125
                },
                {
                  "word": "台",
                  "start": 0.98,
                  "end": 1.12,
                  "probability": 0.88671875
                }
              ]
            },
            {
              "id": 1,
              "seek": 0,
              "start": 1.9999999999999996,
              "end": 4.08,
              "text": "现在都能用同一套身手的体验",
              "tokens": [
                50427,
                25040,
                7182,
                8225,
                9254,
                13089,
                2257,
                1881,
                245,
                19847,
                11389,
                1546,
                29485,
                49657,
                234,
                50571
              ],
              "temperature": 0.0,
              "avg_logprob": -0.2176045950721292,
              "compression_ratio": 0.8764044943820225,
              "no_speech_prob": 0.6835574507713318,
              "words": [
                {
                  "word": "现在",
                  "start": 1.9999999999999996,
                  "end": 2.32,
                  "probability": 0.982421875
                },
                {
                  "word": "都",
                  "start": 2.32,
                  "end": 2.64,
                  "probability": 0.98193359375
                },
                {
                  "word": "能",
                  "start": 2.64,
                  "end": 2.74,
                  "probability": 0.9990234375
                },
                {
                  "word": "用",
                  "start": 2.74,
                  "end": 2.9,
                  "probability": 0.9990234375
                },
                {
                  "word": "同",
                  "start": 2.9,
                  "end": 3.06,
                  "probability": 0.99658203125
                },
                {
                  "word": "一",
                  "start": 3.06,
                  "end": 3.24,
                  "probability": 0.99951171875
                },
                {
                  "word": "套",
                  "start": 3.24,
                  "end": 3.4,
                  "probability": 1.0
                },
                {
                  "word": "身",
                  "start": 3.4,
                  "end": 3.5,
                  "probability": 0.5009765625
                },
                {
                  "word": "手",
                  "start": 3.5,
                  "end": 3.66,
                  "probability": 0.99853515625
                },
                {
                  "word": "的",
                  "start": 3.66,
                  "end": 3.72,
                  "probability": 0.998046875
                },
                {
                  "word": "体",
                  "start": 3.72,
                  "end": 3.86,
                  "probability": 0.99951171875
                },
                {
                  "word": "验",
                  "start": 3.86,
                  "end": 4.08,
                  "probability": 1.0
                }
              ]
            },
            {
              "id": 2,
              "seek": 0,
              "start": 4.08,
              "end": 5.62,
              "text": "让子母自己跟上",
              "tokens": [
                50571,
                33650,
                7626,
                35744,
                17645,
                9678,
                5708,
                50650
              ],
              "temperature": 0.0,
              "avg_logprob": -0.2176045950721292,
              "compression_ratio": 0.8764044943820225,
              "no_speech_prob": 0.6835574507713318,
              "words": [
                {
                  "word": "让",
                  "start": 4.08,
                  "end": 4.66,
                  "probability": 0.9970703125
                },
                {
                  "word": "子",
                  "start": 4.66,
                  "end": 4.78,
                  "probability": 0.59375
                },
                {
                  "word": "母",
                  "start": 4.78,
                  "end": 5.0,
                  "probability": 0.65966796875
                },
                {
                  "word": "自己",
                  "start": 5.0,
                  "end": 5.18,
                  "probability": 0.9990234375
                },
                {
                  "word": "跟",
                  "start": 5.18,
                  "end": 5.4,
                  "probability": 0.9990234375
                },
                {
                  "word": "上",
                  "start": 5.4,
                  "end": 5.62,
                  "probability": 1.0
                }
              ]
            }
          ],
          "language": "zh",
          "duration": 6.14
        },
        "status": "COMPLETE",
        "requested_language": "zh"
      },
      "raw_word_timestamps": [
        {
          "word": "现在",
          "start": 24.16,
          "end": 24.52,
          "probability": 0.9990234375
        },
        {
          "word": "都",
          "start": 24.52,
          "end": 24.88,
          "probability": 0.998046875
        },
        {
          "word": "能",
          "start": 24.88,
          "end": 24.979999999999997,
          "probability": 1.0
        },
        {
          "word": "用",
          "start": 24.979999999999997,
          "end": 25.119999999999997,
          "probability": 1.0
        },
        {
          "word": "同",
          "start": 25.119999999999997,
          "end": 25.279999999999998,
          "probability": 0.9990234375
        },
        {
          "word": "一",
          "start": 25.279999999999998,
          "end": 25.419999999999998,
          "probability": 1.0
        },
        {
          "word": "套",
          "start": 25.419999999999998,
          "end": 25.58,
          "probability": 0.9990234375
        },
        {
          "word": "伸",
          "start": 25.58,
          "end": 25.72,
          "probability": 0.7099609375
        },
        {
          "word": "手",
          "start": 25.72,
          "end": 25.86,
          "probability": 0.9990234375
        },
        {
          "word": "的",
          "start": 25.86,
          "end": 25.959999999999997,
          "probability": 0.9990234375
        },
        {
          "word": "体",
          "start": 25.959999999999997,
          "end": 26.099999999999998,
          "probability": 0.99951171875
        },
        {
          "word": "验",
          "start": 26.099999999999998,
          "end": 26.299999999999997,
          "probability": 1.0
        }
      ],
      "word_timestamps": [],
      "source_text": "伸手",
      "review_status": "BEST_SUPPORTED_WITH_REVIEW",
      "canonical_timing_evidence": "CUE_ANCHORED",
      "index": 8
    },
    {
      "start": 26.299999999999997,
      "end": 27.82,
      "text": "让字母自己跟上",
      "raw_asr_text": "让字母自己跟上",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 0,
      "words": [
        {
          "word": "让",
          "start": 26.299999999999997,
          "end": 26.9,
          "probability": 0.99853515625
        },
        {
          "word": "字",
          "start": 26.9,
          "end": 27.08,
          "probability": 0.92626953125
        },
        {
          "word": "母",
          "start": 27.08,
          "end": 27.2,
          "probability": 0.6123046875
        },
        {
          "word": "自己",
          "start": 27.2,
          "end": 27.439999999999998,
          "probability": 1.0
        },
        {
          "word": "跟",
          "start": 27.439999999999998,
          "end": 27.619999999999997,
          "probability": 0.99951171875
        },
        {
          "word": "上",
          "start": 27.619999999999997,
          "end": 27.82,
          "probability": 1.0
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.13841238822645813,
        "compression_ratio": 1.1423611111111112,
        "no_speech_prob": 2.5074888346243718e-11
      },
      "lid_status": "UNCERTAIN",
      "alternatives": [
        {
          "language": "zh",
          "text": "哈喽大家就等了同传字幕1.0正式版今天终于和大家见面了从最初的小工具到现在Android、Mac还有大家呼声超高的Windows三个平台这次全都齐了不管你是睡前抱着手机躺看直播还是在电脑前边干活边挂着生肉台现在都能用同一套伸手的体验让字母自己跟上"
        }
      ],
      "id": 9,
      "review_annotation": "",
      "row_id": "9",
      "bounded_row_id": 9,
      "parent_row_ids": [
        9
      ],
      "parent_row_id": 9,
      "packet_id": 0,
      "LID_evidence": [
        {
          "region_id": 7,
          "probabilities": {
            "zh": 0.997811065779792
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        7
      ],
      "language_confidence": 0.997811065779792,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 1.9000000000000001,
      "packet_end": 28.900000000000002,
      "raw_text": "让字母自己跟上",
      "selected_text": "让字母自己跟上",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            9
          ],
          "packet_id": 0,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        0
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 9
    },
    {
      "start": 33.66,
      "end": 34.7,
      "text": "他现在能做什么",
      "raw_asr_text": "他现在能做什么",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 1,
      "words": [
        {
          "word": "他",
          "start": 33.66,
          "end": 33.96,
          "probability": 0.490478515625
        },
        {
          "word": "现在",
          "start": 33.96,
          "end": 34.14,
          "probability": 0.99755859375
        },
        {
          "word": "能",
          "start": 34.14,
          "end": 34.26,
          "probability": 0.998046875
        },
        {
          "word": "做",
          "start": 34.26,
          "end": 34.44,
          "probability": 0.9990234375
        },
        {
          "word": "什么",
          "start": 34.44,
          "end": 34.7,
          "probability": 0.99658203125
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.27883252963213856,
        "compression_ratio": 1.2687747035573123,
        "no_speech_prob": 5.346294587593725e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "现在新朋友快速看看他现在能做什么平时看日语或者英语的视频和直播他能实时听声音出字幕翻成中文或者你想看的任何语言列表里自带了中英日韩其他任何语言只要你自己填个名字它就能翻字幕会像这样浮在画面最上面随便拖动改大小换配色"
        }
      ],
      "id": 10,
      "review_annotation": "",
      "row_id": "10",
      "bounded_row_id": 10,
      "parent_row_ids": [
        10
      ],
      "parent_row_id": 10,
      "packet_id": 1,
      "LID_evidence": [
        {
          "region_id": 8,
          "probabilities": {
            "zh": 0.9978148804770576
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        8
      ],
      "language_confidence": 0.9978148804770576,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 32.5,
      "packet_end": 54.5,
      "raw_text": "他现在能做什么",
      "selected_text": "他现在能做什么",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            10
          ],
          "packet_id": 1,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        1
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 10
    },
    {
      "start": 35.26,
      "end": 37.34,
      "text": "平时看日语或者英语的视频和直播",
      "raw_asr_text": "平时看日语或者英语的视频和直播",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 1,
      "words": [
        {
          "word": "平",
          "start": 35.1,
          "end": 35.28,
          "probability": 0.9208984375
        },
        {
          "word": "时",
          "start": 35.28,
          "end": 35.46,
          "probability": 0.998046875
        },
        {
          "word": "看",
          "start": 35.46,
          "end": 35.62,
          "probability": 0.99853515625
        },
        {
          "word": "日",
          "start": 35.62,
          "end": 35.78,
          "probability": 0.99853515625
        },
        {
          "word": "语",
          "start": 35.78,
          "end": 35.96,
          "probability": 0.9990234375
        },
        {
          "word": "或者",
          "start": 35.96,
          "end": 36.2,
          "probability": 0.9873046875
        },
        {
          "word": "英",
          "start": 36.2,
          "end": 36.38,
          "probability": 0.9736328125
        },
        {
          "word": "语",
          "start": 36.38,
          "end": 36.5,
          "probability": 1.0
        },
        {
          "word": "的",
          "start": 36.5,
          "end": 36.6,
          "probability": 0.998046875
        },
        {
          "word": "视",
          "start": 36.6,
          "end": 36.74,
          "probability": 0.99951171875
        },
        {
          "word": "频",
          "start": 36.74,
          "end": 36.88,
          "probability": 1.0
        },
        {
          "word": "和",
          "start": 36.88,
          "end": 37.0,
          "probability": 0.998046875
        },
        {
          "word": "直",
          "start": 37.0,
          "end": 37.16,
          "probability": 1.0
        },
        {
          "word": "播",
          "start": 37.16,
          "end": 37.34,
          "probability": 1.0
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.27883252963213856,
        "compression_ratio": 1.2687747035573123,
        "no_speech_prob": 5.346294587593725e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "现在新朋友快速看看他现在能做什么平时看日语或者英语的视频和直播他能实时听声音出字幕翻成中文或者你想看的任何语言列表里自带了中英日韩其他任何语言只要你自己填个名字它就能翻字幕会像这样浮在画面最上面随便拖动改大小换配色"
        }
      ],
      "id": 11,
      "review_annotation": "",
      "row_id": "11",
      "bounded_row_id": 11,
      "parent_row_ids": [
        11
      ],
      "parent_row_id": 11,
      "packet_id": 1,
      "LID_evidence": [
        {
          "region_id": 8,
          "probabilities": {
            "zh": 0.9978148804770576
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        8
      ],
      "language_confidence": 0.9978148804770576,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 32.5,
      "packet_end": 54.5,
      "raw_text": "平时看日语或者英语的视频和直播",
      "selected_text": "平时看日语或者英语的视频和直播",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            11
          ],
          "packet_id": 1,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        1
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 11
    },
    {
      "start": 37.34,
      "end": 38.7,
      "text": "他能实时听声音",
      "raw_asr_text": "他能实时听声音",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 1,
      "words": [
        {
          "word": "他",
          "start": 37.34,
          "end": 37.78,
          "probability": 0.970703125
        },
        {
          "word": "能",
          "start": 37.78,
          "end": 37.92,
          "probability": 0.998046875
        },
        {
          "word": "实",
          "start": 37.92,
          "end": 38.06,
          "probability": 0.7939453125
        },
        {
          "word": "时",
          "start": 38.06,
          "end": 38.16,
          "probability": 0.94921875
        },
        {
          "word": "听",
          "start": 38.16,
          "end": 38.3,
          "probability": 0.99951171875
        },
        {
          "word": "声",
          "start": 38.3,
          "end": 38.519999999999996,
          "probability": 0.99853515625
        },
        {
          "word": "音",
          "start": 38.519999999999996,
          "end": 38.7,
          "probability": 0.99951171875
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.27883252963213856,
        "compression_ratio": 1.2687747035573123,
        "no_speech_prob": 5.346294587593725e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "现在新朋友快速看看他现在能做什么平时看日语或者英语的视频和直播他能实时听声音出字幕翻成中文或者你想看的任何语言列表里自带了中英日韩其他任何语言只要你自己填个名字它就能翻字幕会像这样浮在画面最上面随便拖动改大小换配色"
        }
      ],
      "id": 12,
      "review_annotation": "",
      "row_id": "12",
      "bounded_row_id": 12,
      "parent_row_ids": [
        12
      ],
      "parent_row_id": 12,
      "packet_id": 1,
      "LID_evidence": [
        {
          "region_id": 8,
          "probabilities": {
            "zh": 0.9978148804770576
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        8
      ],
      "language_confidence": 0.9978148804770576,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 32.5,
      "packet_end": 54.5,
      "raw_text": "他能实时听声音",
      "selected_text": "他能实时听声音",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            12
          ],
          "packet_id": 1,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        1
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 12
    },
    {
      "start": 38.7,
      "end": 39.4,
      "text": "出字幕",
      "raw_asr_text": "出字幕",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 1,
      "words": [
        {
          "word": "出",
          "start": 38.7,
          "end": 39.019999999999996,
          "probability": 0.81982421875
        },
        {
          "word": "字",
          "start": 39.019999999999996,
          "end": 39.18,
          "probability": 0.9970703125
        },
        {
          "word": "幕",
          "start": 39.18,
          "end": 39.4,
          "probability": 0.974609375
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.27883252963213856,
        "compression_ratio": 1.2687747035573123,
        "no_speech_prob": 5.346294587593725e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "现在新朋友快速看看他现在能做什么平时看日语或者英语的视频和直播他能实时听声音出字幕翻成中文或者你想看的任何语言列表里自带了中英日韩其他任何语言只要你自己填个名字它就能翻字幕会像这样浮在画面最上面随便拖动改大小换配色"
        }
      ],
      "id": 13,
      "review_annotation": "",
      "row_id": "13",
      "bounded_row_id": 13,
      "parent_row_ids": [
        13
      ],
      "parent_row_id": 13,
      "packet_id": 1,
      "LID_evidence": [
        {
          "region_id": 8,
          "probabilities": {
            "zh": 0.9978148804770576
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        8
      ],
      "language_confidence": 0.9978148804770576,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 32.5,
      "packet_end": 54.5,
      "raw_text": "出字幕",
      "selected_text": "出字幕",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            13
          ],
          "packet_id": 1,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        1
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 13
    },
    {
      "start": 39.4,
      "end": 41.96,
      "text": "翻成中文或者你想看的任何语言",
      "raw_asr_text": "翻成中文或者你想看的任何语言",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 1,
      "words": [
        {
          "word": "翻",
          "start": 39.4,
          "end": 39.9,
          "probability": 0.99560546875
        },
        {
          "word": "成",
          "start": 39.9,
          "end": 40.08,
          "probability": 0.99755859375
        },
        {
          "word": "中",
          "start": 40.08,
          "end": 40.24,
          "probability": 0.99853515625
        },
        {
          "word": "文",
          "start": 40.24,
          "end": 40.44,
          "probability": 1.0
        },
        {
          "word": "或者",
          "start": 40.44,
          "end": 40.74,
          "probability": 0.982421875
        },
        {
          "word": "你想",
          "start": 40.74,
          "end": 41.019999999999996,
          "probability": 0.9970703125
        },
        {
          "word": "看",
          "start": 41.019999999999996,
          "end": 41.18,
          "probability": 0.99560546875
        },
        {
          "word": "的",
          "start": 41.18,
          "end": 41.28,
          "probability": 0.998046875
        },
        {
          "word": "任",
          "start": 41.28,
          "end": 41.38,
          "probability": 0.99853515625
        },
        {
          "word": "何",
          "start": 41.38,
          "end": 41.56,
          "probability": 1.0
        },
        {
          "word": "语",
          "start": 41.56,
          "end": 41.74,
          "probability": 1.0
        },
        {
          "word": "言",
          "start": 41.74,
          "end": 41.96,
          "probability": 0.99853515625
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.27883252963213856,
        "compression_ratio": 1.2687747035573123,
        "no_speech_prob": 5.346294587593725e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "现在新朋友快速看看他现在能做什么平时看日语或者英语的视频和直播他能实时听声音出字幕翻成中文或者你想看的任何语言列表里自带了中英日韩其他任何语言只要你自己填个名字它就能翻字幕会像这样浮在画面最上面随便拖动改大小换配色"
        }
      ],
      "id": 14,
      "review_annotation": "",
      "row_id": "14",
      "bounded_row_id": 14,
      "parent_row_ids": [
        14
      ],
      "parent_row_id": 14,
      "packet_id": 1,
      "LID_evidence": [
        {
          "region_id": 8,
          "probabilities": {
            "zh": 0.9978148804770576
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        8
      ],
      "language_confidence": 0.9978148804770576,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 32.5,
      "packet_end": 54.5,
      "raw_text": "翻成中文或者你想看的任何语言",
      "selected_text": "翻成中文或者你想看的任何语言",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            14
          ],
          "packet_id": 1,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        1
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 14
    },
    {
      "start": 42.66,
      "end": 43.56,
      "text": "列表里自带了中",
      "raw_asr_text": "列表里自带了中",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 1,
      "words": [
        {
          "word": "列",
          "start": 42.5,
          "end": 42.68,
          "probability": 0.99609375
        },
        {
          "word": "表",
          "start": 42.68,
          "end": 42.84,
          "probability": 0.9912109375
        },
        {
          "word": "里",
          "start": 42.84,
          "end": 42.96,
          "probability": 0.98583984375
        },
        {
          "word": "自",
          "start": 42.96,
          "end": 43.14,
          "probability": 0.99072265625
        },
        {
          "word": "带",
          "start": 43.14,
          "end": 43.26,
          "probability": 0.9873046875
        },
        {
          "word": "了",
          "start": 43.26,
          "end": 43.38,
          "probability": 0.89599609375
        },
        {
          "word": "中",
          "start": 43.38,
          "end": 43.56,
          "probability": 0.99267578125
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.27883252963213856,
        "compression_ratio": 1.2687747035573123,
        "no_speech_prob": 5.346294587593725e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "现在新朋友快速看看他现在能做什么平时看日语或者英语的视频和直播他能实时听声音出字幕翻成中文或者你想看的任何语言列表里自带了中英日韩其他任何语言只要你自己填个名字它就能翻字幕会像这样浮在画面最上面随便拖动改大小换配色"
        }
      ],
      "id": 15,
      "review_annotation": "",
      "row_id": "15",
      "bounded_row_id": 15,
      "parent_row_ids": [
        15
      ],
      "parent_row_id": 15,
      "packet_id": 1,
      "LID_evidence": [
        {
          "region_id": 9,
          "probabilities": {
            "zh": 0.9953408539295197
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        9
      ],
      "language_confidence": 0.9953408539295197,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 32.5,
      "packet_end": 54.5,
      "raw_text": "列表里自带了中",
      "selected_text": "列表里自带了中",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            15
          ],
          "packet_id": 1,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        1
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 15
    },
    {
      "start": 43.56,
      "end": 44.0,
      "text": "英",
      "raw_asr_text": "英",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 1,
      "words": [
        {
          "word": "英",
          "start": 43.56,
          "end": 44.0,
          "probability": 0.404296875
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.27883252963213856,
        "compression_ratio": 1.2687747035573123,
        "no_speech_prob": 5.346294587593725e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "现在新朋友快速看看他现在能做什么平时看日语或者英语的视频和直播他能实时听声音出字幕翻成中文或者你想看的任何语言列表里自带了中英日韩其他任何语言只要你自己填个名字它就能翻字幕会像这样浮在画面最上面随便拖动改大小换配色"
        }
      ],
      "id": 16,
      "review_annotation": "",
      "row_id": "16",
      "bounded_row_id": 16,
      "parent_row_ids": [
        16
      ],
      "parent_row_id": 16,
      "packet_id": 1,
      "LID_evidence": [
        {
          "region_id": 9,
          "probabilities": {
            "zh": 0.9953408539295197
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        9
      ],
      "language_confidence": 0.9953408539295197,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 32.5,
      "packet_end": 54.5,
      "raw_text": "英",
      "selected_text": "英",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            16
          ],
          "packet_id": 1,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        1
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 16
    },
    {
      "start": 44.0,
      "end": 44.44,
      "text": "日",
      "raw_asr_text": "日",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 1,
      "words": [
        {
          "word": "日",
          "start": 44.0,
          "end": 44.44,
          "probability": 0.982421875
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.27883252963213856,
        "compression_ratio": 1.2687747035573123,
        "no_speech_prob": 5.346294587593725e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "现在新朋友快速看看他现在能做什么平时看日语或者英语的视频和直播他能实时听声音出字幕翻成中文或者你想看的任何语言列表里自带了中英日韩其他任何语言只要你自己填个名字它就能翻字幕会像这样浮在画面最上面随便拖动改大小换配色"
        }
      ],
      "id": 17,
      "review_annotation": "",
      "row_id": "17",
      "bounded_row_id": 17,
      "parent_row_ids": [
        17
      ],
      "parent_row_id": 17,
      "packet_id": 1,
      "LID_evidence": [
        {
          "region_id": 9,
          "probabilities": {
            "zh": 0.9953408539295197
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        9
      ],
      "language_confidence": 0.9953408539295197,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 32.5,
      "packet_end": 54.5,
      "raw_text": "日",
      "selected_text": "日",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            17
          ],
          "packet_id": 1,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        1
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 17
    },
    {
      "start": 44.44,
      "end": 44.86,
      "text": "韩",
      "raw_asr_text": "韩",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 1,
      "words": [
        {
          "word": "韩",
          "start": 44.44,
          "end": 44.86,
          "probability": 0.99267578125
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.27883252963213856,
        "compression_ratio": 1.2687747035573123,
        "no_speech_prob": 5.346294587593725e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "现在新朋友快速看看他现在能做什么平时看日语或者英语的视频和直播他能实时听声音出字幕翻成中文或者你想看的任何语言列表里自带了中英日韩其他任何语言只要你自己填个名字它就能翻字幕会像这样浮在画面最上面随便拖动改大小换配色"
        }
      ],
      "id": 18,
      "review_annotation": "",
      "row_id": "18",
      "bounded_row_id": 18,
      "parent_row_ids": [
        18
      ],
      "parent_row_id": 18,
      "packet_id": 1,
      "LID_evidence": [
        {
          "region_id": 9,
          "probabilities": {
            "zh": 0.9953408539295197
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        9
      ],
      "language_confidence": 0.9953408539295197,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 32.5,
      "packet_end": 54.5,
      "raw_text": "韩",
      "selected_text": "韩",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            18
          ],
          "packet_id": 1,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        1
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 18
    },
    {
      "start": 44.86,
      "end": 45.94,
      "text": "其他任何语言",
      "raw_asr_text": "其他任何语言",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 1,
      "words": [
        {
          "word": "其",
          "start": 44.86,
          "end": 45.32,
          "probability": 0.9990234375
        },
        {
          "word": "他",
          "start": 45.32,
          "end": 45.42,
          "probability": 0.966796875
        },
        {
          "word": "任",
          "start": 45.42,
          "end": 45.54,
          "probability": 0.9990234375
        },
        {
          "word": "何",
          "start": 45.54,
          "end": 45.7,
          "probability": 1.0
        },
        {
          "word": "语",
          "start": 45.7,
          "end": 45.82,
          "probability": 1.0
        },
        {
          "word": "言",
          "start": 45.82,
          "end": 45.94,
          "probability": 0.98486328125
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.27883252963213856,
        "compression_ratio": 1.2687747035573123,
        "no_speech_prob": 5.346294587593725e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "现在新朋友快速看看他现在能做什么平时看日语或者英语的视频和直播他能实时听声音出字幕翻成中文或者你想看的任何语言列表里自带了中英日韩其他任何语言只要你自己填个名字它就能翻字幕会像这样浮在画面最上面随便拖动改大小换配色"
        }
      ],
      "id": 19,
      "review_annotation": "",
      "row_id": "19",
      "bounded_row_id": 19,
      "parent_row_ids": [
        19
      ],
      "parent_row_id": 19,
      "packet_id": 1,
      "LID_evidence": [
        {
          "region_id": 9,
          "probabilities": {
            "zh": 0.9953408539295197
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        9
      ],
      "language_confidence": 0.9953408539295197,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 32.5,
      "packet_end": 54.5,
      "raw_text": "其他任何语言",
      "selected_text": "其他任何语言",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            19
          ],
          "packet_id": 1,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        1
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 19
    },
    {
      "start": 45.94,
      "end": 47.26,
      "text": "只要你自己填个名字",
      "raw_asr_text": "只要你自己填个名字",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 1,
      "words": [
        {
          "word": "只",
          "start": 45.94,
          "end": 46.22,
          "probability": 0.99755859375
        },
        {
          "word": "要",
          "start": 46.22,
          "end": 46.32,
          "probability": 0.998046875
        },
        {
          "word": "你",
          "start": 46.32,
          "end": 46.46,
          "probability": 0.99951171875
        },
        {
          "word": "自己",
          "start": 46.46,
          "end": 46.68,
          "probability": 0.99951171875
        },
        {
          "word": "填",
          "start": 46.68,
          "end": 46.82,
          "probability": 0.81494140625
        },
        {
          "word": "个",
          "start": 46.82,
          "end": 46.94,
          "probability": 0.9951171875
        },
        {
          "word": "名",
          "start": 46.94,
          "end": 47.06,
          "probability": 0.99951171875
        },
        {
          "word": "字",
          "start": 47.06,
          "end": 47.26,
          "probability": 0.99951171875
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.27883252963213856,
        "compression_ratio": 1.2687747035573123,
        "no_speech_prob": 5.346294587593725e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "现在新朋友快速看看他现在能做什么平时看日语或者英语的视频和直播他能实时听声音出字幕翻成中文或者你想看的任何语言列表里自带了中英日韩其他任何语言只要你自己填个名字它就能翻字幕会像这样浮在画面最上面随便拖动改大小换配色"
        }
      ],
      "id": 20,
      "review_annotation": "",
      "row_id": "20",
      "bounded_row_id": 20,
      "parent_row_ids": [
        20
      ],
      "parent_row_id": 20,
      "packet_id": 1,
      "LID_evidence": [
        {
          "region_id": 9,
          "probabilities": {
            "zh": 0.9953408539295197
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        9
      ],
      "language_confidence": 0.9953408539295197,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 32.5,
      "packet_end": 54.5,
      "raw_text": "只要你自己填个名字",
      "selected_text": "只要你自己填个名字",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            20
          ],
          "packet_id": 1,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        1
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 20
    },
    {
      "start": 47.26,
      "end": 48.0,
      "text": "它就能翻",
      "raw_asr_text": "它就能翻",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 1,
      "words": [
        {
          "word": "它",
          "start": 47.26,
          "end": 47.58,
          "probability": 0.454833984375
        },
        {
          "word": "就",
          "start": 47.58,
          "end": 47.72,
          "probability": 0.99853515625
        },
        {
          "word": "能",
          "start": 47.72,
          "end": 47.84,
          "probability": 0.9970703125
        },
        {
          "word": "翻",
          "start": 47.84,
          "end": 48.0,
          "probability": 0.99853515625
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.27883252963213856,
        "compression_ratio": 1.2687747035573123,
        "no_speech_prob": 5.346294587593725e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "现在新朋友快速看看他现在能做什么平时看日语或者英语的视频和直播他能实时听声音出字幕翻成中文或者你想看的任何语言列表里自带了中英日韩其他任何语言只要你自己填个名字它就能翻字幕会像这样浮在画面最上面随便拖动改大小换配色"
        }
      ],
      "id": 21,
      "review_annotation": "",
      "row_id": "21",
      "bounded_row_id": 21,
      "parent_row_ids": [
        21
      ],
      "parent_row_id": 21,
      "packet_id": 1,
      "LID_evidence": [
        {
          "region_id": 9,
          "probabilities": {
            "zh": 0.9953408539295197
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        9
      ],
      "language_confidence": 0.9953408539295197,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 32.5,
      "packet_end": 54.5,
      "raw_text": "它就能翻",
      "selected_text": "它就能翻",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            21
          ],
          "packet_id": 1,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        1
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 21
    },
    {
      "start": 49.06,
      "end": 51.2,
      "text": "字幕会像这样浮在画面最上面",
      "raw_asr_text": "字幕会像这样浮在画面最上面",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 1,
      "words": [
        {
          "word": "字",
          "start": 49.06,
          "end": 49.42,
          "probability": 0.99609375
        },
        {
          "word": "幕",
          "start": 49.42,
          "end": 49.58,
          "probability": 0.98583984375
        },
        {
          "word": "会",
          "start": 49.58,
          "end": 49.72,
          "probability": 0.998046875
        },
        {
          "word": "像",
          "start": 49.72,
          "end": 49.86,
          "probability": 0.9853515625
        },
        {
          "word": "这样",
          "start": 49.86,
          "end": 50.16,
          "probability": 0.99853515625
        },
        {
          "word": "浮",
          "start": 50.16,
          "end": 50.3,
          "probability": 0.7607421875
        },
        {
          "word": "在",
          "start": 50.3,
          "end": 50.42,
          "probability": 0.99560546875
        },
        {
          "word": "画",
          "start": 50.42,
          "end": 50.56,
          "probability": 0.9990234375
        },
        {
          "word": "面",
          "start": 50.56,
          "end": 50.739999999999995,
          "probability": 1.0
        },
        {
          "word": "最",
          "start": 50.739999999999995,
          "end": 50.9,
          "probability": 0.9990234375
        },
        {
          "word": "上面",
          "start": 50.9,
          "end": 51.2,
          "probability": 0.9970703125
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.27883252963213856,
        "compression_ratio": 1.2687747035573123,
        "no_speech_prob": 5.346294587593725e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "现在新朋友快速看看他现在能做什么平时看日语或者英语的视频和直播他能实时听声音出字幕翻成中文或者你想看的任何语言列表里自带了中英日韩其他任何语言只要你自己填个名字它就能翻字幕会像这样浮在画面最上面随便拖动改大小换配色"
        }
      ],
      "id": 22,
      "review_annotation": "",
      "row_id": "22",
      "bounded_row_id": 22,
      "parent_row_ids": [
        22
      ],
      "parent_row_id": 22,
      "packet_id": 1,
      "LID_evidence": [
        {
          "region_id": 10,
          "probabilities": {
            "zh": 0.9962855047649808
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        10
      ],
      "language_confidence": 0.9962855047649808,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 32.5,
      "packet_end": 54.5,
      "raw_text": "字幕会像这样浮在画面最上面",
      "selected_text": "字幕会像这样浮在画面最上面",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            22
          ],
          "packet_id": 1,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        1
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 22
    },
    {
      "start": 51.2,
      "end": 52.06,
      "text": "随便拖动",
      "raw_asr_text": "随便拖动",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 1,
      "words": [
        {
          "word": "随",
          "start": 51.2,
          "end": 51.6,
          "probability": 1.0
        },
        {
          "word": "便",
          "start": 51.6,
          "end": 51.72,
          "probability": 0.9970703125
        },
        {
          "word": "拖",
          "start": 51.72,
          "end": 51.9,
          "probability": 0.98583984375
        },
        {
          "word": "动",
          "start": 51.9,
          "end": 52.06,
          "probability": 0.99365234375
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.27883252963213856,
        "compression_ratio": 1.2687747035573123,
        "no_speech_prob": 5.346294587593725e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "现在新朋友快速看看他现在能做什么平时看日语或者英语的视频和直播他能实时听声音出字幕翻成中文或者你想看的任何语言列表里自带了中英日韩其他任何语言只要你自己填个名字它就能翻字幕会像这样浮在画面最上面随便拖动改大小换配色"
        }
      ],
      "id": 23,
      "review_annotation": "",
      "row_id": "23",
      "bounded_row_id": 23,
      "parent_row_ids": [
        23
      ],
      "parent_row_id": 23,
      "packet_id": 1,
      "LID_evidence": [
        {
          "region_id": 10,
          "probabilities": {
            "zh": 0.9962855047649808
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        10
      ],
      "language_confidence": 0.9962855047649808,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 32.5,
      "packet_end": 54.5,
      "raw_text": "随便拖动",
      "selected_text": "随便拖动",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            23
          ],
          "packet_id": 1,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        1
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 23
    },
    {
      "start": 52.06,
      "end": 52.739999999999995,
      "text": "改大小",
      "raw_asr_text": "改大小",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 1,
      "words": [
        {
          "word": "改",
          "start": 52.06,
          "end": 52.34,
          "probability": 0.85400390625
        },
        {
          "word": "大",
          "start": 52.34,
          "end": 52.519999999999996,
          "probability": 0.9990234375
        },
        {
          "word": "小",
          "start": 52.519999999999996,
          "end": 52.739999999999995,
          "probability": 0.99658203125
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.27883252963213856,
        "compression_ratio": 1.2687747035573123,
        "no_speech_prob": 5.346294587593725e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "现在新朋友快速看看他现在能做什么平时看日语或者英语的视频和直播他能实时听声音出字幕翻成中文或者你想看的任何语言列表里自带了中英日韩其他任何语言只要你自己填个名字它就能翻字幕会像这样浮在画面最上面随便拖动改大小换配色"
        }
      ],
      "id": 24,
      "review_annotation": "",
      "row_id": "24",
      "bounded_row_id": 24,
      "parent_row_ids": [
        24
      ],
      "parent_row_id": 24,
      "packet_id": 1,
      "LID_evidence": [
        {
          "region_id": 10,
          "probabilities": {
            "zh": 0.9962855047649808
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        10
      ],
      "language_confidence": 0.9962855047649808,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 32.5,
      "packet_end": 54.5,
      "raw_text": "改大小",
      "selected_text": "改大小",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            24
          ],
          "packet_id": 1,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        1
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 24
    },
    {
      "start": 52.739999999999995,
      "end": 53.379999999999995,
      "text": "换配色",
      "raw_asr_text": "换配色",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 1,
      "words": [
        {
          "word": "换",
          "start": 52.739999999999995,
          "end": 53.019999999999996,
          "probability": 0.9287109375
        },
        {
          "word": "配",
          "start": 53.019999999999996,
          "end": 53.16,
          "probability": 0.99462890625
        },
        {
          "word": "色",
          "start": 53.16,
          "end": 53.379999999999995,
          "probability": 1.0
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.27883252963213856,
        "compression_ratio": 1.2687747035573123,
        "no_speech_prob": 5.346294587593725e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "现在新朋友快速看看他现在能做什么平时看日语或者英语的视频和直播他能实时听声音出字幕翻成中文或者你想看的任何语言列表里自带了中英日韩其他任何语言只要你自己填个名字它就能翻字幕会像这样浮在画面最上面随便拖动改大小换配色"
        }
      ],
      "id": 25,
      "review_annotation": "",
      "row_id": "25",
      "bounded_row_id": 25,
      "parent_row_ids": [
        25
      ],
      "parent_row_id": 25,
      "packet_id": 1,
      "LID_evidence": [
        {
          "region_id": 10,
          "probabilities": {
            "zh": 0.9962855047649808
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        10
      ],
      "language_confidence": 0.9962855047649808,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 32.5,
      "packet_end": 54.5,
      "raw_text": "换配色",
      "selected_text": "换配色",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            25
          ],
          "packet_id": 1,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        1
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 25
    },
    {
      "start": 54.32000000000001,
      "end": 56.660000000000004,
      "text": "房屋处",
      "raw_asr_text": "调好位置还能一键锁定房屋处",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 2,
      "words": [],
      "asr_metrics": {
        "avg_logprob": -0.11145266908587831,
        "compression_ratio": 1.2433333333333334,
        "no_speech_prob": 6.322219137100404e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "调好位置还能一键锁定房屋处你可以选边听边出草稿随说随出一句说完立刻变正式字幕也可以切成稳稳地说完再出在手机上你可以单独选某一个会出声的APP来听到了电脑上它直接听整台电脑播放的声音哪怕关掉主窗口字幕也照样在Mac的菜单栏和Windows的右下角拖盘点一下就能随时开关因"
        }
      ],
      "id": 26,
      "review_annotation": "待確認",
      "row_id": "26",
      "bounded_row_id": 26,
      "parent_row_ids": [
        26
      ],
      "parent_row_id": 26,
      "packet_id": 2,
      "LID_evidence": [
        {
          "region_id": 11,
          "probabilities": {
            "zh": 0.997185610807859
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        11
      ],
      "language_confidence": 0.997185610807859,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 53.300000000000004,
      "packet_end": 81.2,
      "raw_text": "调好位置还能一键锁定房屋处",
      "selected_text": "房屋处",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            26
          ],
          "packet_id": 2,
          "engine": "mlx-whisper"
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "26",
          "before": "调好位置还能一键锁定房屋处",
          "after": "房屋处",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "锁定房屋处语义不明，疑似UI术语误识",
          "context_evidence": "应为'悬浮窗'或'固定位置'等UI相关词汇",
          "independent_ASR": {
            "engine": "whisper-large-v3",
            "model_path": "/Users/sagawa/AI_WORK/cache/huggingface/hub/models--mlx-community--whisper-large-v3",
            "language": "zh",
            "audio_interval": [
              52.32000000000001,
              58.660000000000004
            ],
            "target_interval": [
              54.32000000000001,
              56.660000000000004
            ],
            "text": "调好位置还能一键锁定防务处",
            "words": [
              {
                "word": "调",
                "start": 54.400000000000006,
                "end": 54.720000000000006,
                "probability": 0.947265625
              },
              {
                "word": "好",
                "start": 54.720000000000006,
                "end": 54.800000000000004,
                "probability": 0.998046875
              },
              {
                "word": "位",
                "start": 54.800000000000004,
                "end": 54.940000000000005,
                "probability": 0.9990234375
              },
              {
                "word": "置",
                "start": 54.940000000000005,
                "end": 55.14000000000001,
                "probability": 0.99951171875
              },
              {
                "word": "还",
                "start": 55.14000000000001,
                "end": 55.34000000000001,
                "probability": 0.83349609375
              },
              {
                "word": "能",
                "start": 55.34000000000001,
                "end": 55.480000000000004,
                "probability": 0.99951171875
              },
              {
                "word": "一",
                "start": 55.480000000000004,
                "end": 55.620000000000005,
                "probability": 0.9970703125
              },
              {
                "word": "键",
                "start": 55.620000000000005,
                "end": 55.82000000000001,
                "probability": 0.9990234375
              },
              {
                "word": "锁",
                "start": 55.82000000000001,
                "end": 56.00000000000001,
                "probability": 1.0
              },
              {
                "word": "定",
                "start": 56.00000000000001,
                "end": 56.10000000000001,
                "probability": 0.99755859375
              },
              {
                "word": "防",
                "start": 56.10000000000001,
                "end": 56.24000000000001,
                "probability": 0.75439453125
              },
              {
                "word": "务",
                "start": 56.24000000000001,
                "end": 56.440000000000005,
                "probability": 0.5986328125
              },
              {
                "word": "处",
                "start": 56.440000000000005,
                "end": 56.60000000000001,
                "probability": 0.998046875
              }
            ],
            "raw_result": {
              "text": "大小 光配色调好位置还能一键锁定防务处你可以选边听边出",
              "segments": [
                {
                  "id": 0,
                  "seek": 0,
                  "start": 0.0,
                  "end": 0.98,
                  "text": "大小 光配色",
                  "tokens": [
                    50365,
                    3582,
                    7322,
                    220,
                    20690,
                    38846,
                    17673,
                    50421
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.23005850315093995,
                  "compression_ratio": 0.8777777777777778,
                  "no_speech_prob": 0.9049589037895203,
                  "words": [
                    {
                      "word": "大",
                      "start": 0.0,
                      "end": 0.1,
                      "probability": 0.00997161865234375
                    },
                    {
                      "word": "小",
                      "start": 0.1,
                      "end": 0.4,
                      "probability": 0.98974609375
                    },
                    {
                      "word": " 光",
                      "start": 0.42,
                      "end": 0.66,
                      "probability": 0.91455078125
                    },
                    {
                      "word": "配",
                      "start": 0.66,
                      "end": 0.78,
                      "probability": 0.9501953125
                    },
                    {
                      "word": "色",
                      "start": 0.78,
                      "end": 0.98,
                      "probability": 1.0
                    }
                  ]
                },
                {
                  "id": 1,
                  "seek": 0,
                  "start": 2.0799999999999996,
                  "end": 2.82,
                  "text": "调好位置",
                  "tokens": [
                    50421,
                    8897,
                    225,
                    2131,
                    11160,
                    34719,
                    50509
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.23005850315093995,
                  "compression_ratio": 0.8777777777777778,
                  "no_speech_prob": 0.9049589037895203,
                  "words": [
                    {
                      "word": "调",
                      "start": 2.0799999999999996,
                      "end": 2.4,
                      "probability": 0.947265625
                    },
                    {
                      "word": "好",
                      "start": 2.4,
                      "end": 2.48,
                      "probability": 0.998046875
                    },
                    {
                      "word": "位",
                      "start": 2.48,
                      "end": 2.62,
                      "probability": 0.9990234375
                    },
                    {
                      "word": "置",
                      "start": 2.62,
                      "end": 2.82,
                      "probability": 0.99951171875
                    }
                  ]
                },
                {
                  "id": 2,
                  "seek": 0,
                  "start": 2.82,
                  "end": 4.28,
                  "text": "还能一键锁定防务处",
                  "tokens": [
                    50509,
                    14852,
                    8225,
                    2257,
                    23049,
                    106,
                    23049,
                    223,
                    12088,
                    35863,
                    5087,
                    94,
                    1787,
                    226,
                    50582
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.23005850315093995,
                  "compression_ratio": 0.8777777777777778,
                  "no_speech_prob": 0.9049589037895203,
                  "words": [
                    {
                      "word": "还",
                      "start": 2.82,
                      "end": 3.02,
                      "probability": 0.83349609375
                    },
                    {
                      "word": "能",
                      "start": 3.02,
                      "end": 3.16,
                      "probability": 0.99951171875
                    },
                    {
                      "word": "一",
                      "start": 3.16,
                      "end": 3.3,
                      "probability": 0.9970703125
                    },
                    {
                      "word": "键",
                      "start": 3.3,
                      "end": 3.5,
                      "probability": 0.9990234375
                    },
                    {
                      "word": "锁",
                      "start": 3.5,
                      "end": 3.68,
                      "probability": 1.0
                    },
                    {
                      "word": "定",
                      "start": 3.68,
                      "end": 3.78,
                      "probability": 0.99755859375
                    },
                    {
                      "word": "防",
                      "start": 3.78,
                      "end": 3.92,
                      "probability": 0.75439453125
                    },
                    {
                      "word": "务",
                      "start": 3.92,
                      "end": 4.12,
                      "probability": 0.5986328125
                    },
                    {
                      "word": "处",
                      "start": 4.12,
                      "end": 4.28,
                      "probability": 0.998046875
                    }
                  ]
                },
                {
                  "id": 3,
                  "seek": 0,
                  "start": 5.02,
                  "end": 6.22,
                  "text": "你可以选边听边出",
                  "tokens": [
                    50582,
                    42766,
                    2215,
                    231,
                    40503,
                    31022,
                    40503,
                    7781,
                    50683
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.23005850315093995,
                  "compression_ratio": 0.8777777777777778,
                  "no_speech_prob": 0.9049589037895203,
                  "words": [
                    {
                      "word": "你可以",
                      "start": 5.02,
                      "end": 5.34,
                      "probability": 0.99169921875
                    },
                    {
                      "word": "选",
                      "start": 5.34,
                      "end": 5.66,
                      "probability": 0.9990234375
                    },
                    {
                      "word": "边",
                      "start": 5.66,
                      "end": 5.8,
                      "probability": 0.99169921875
                    },
                    {
                      "word": "听",
                      "start": 5.8,
                      "end": 6.0,
                      "probability": 0.62158203125
                    },
                    {
                      "word": "边",
                      "start": 6.0,
                      "end": 6.16,
                      "probability": 0.998046875
                    },
                    {
                      "word": "出",
                      "start": 6.16,
                      "end": 6.22,
                      "probability": 0.67626953125
                    }
                  ]
                }
              ],
              "language": "zh",
              "duration": 6.34
            },
            "status": "COMPLETE",
            "requested_language": "zh"
          },
          "suppression_proposed": null
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "26",
          "before": "房屋处",
          "after": "房屋处",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "Nonsense: '房屋处' (house place) vs '调好位置' (adjust position)",
          "context_evidence": "ASR: '调好位置还能一键锁定防务处'",
          "independent_ASR": {
            "engine": "whisper-large-v3",
            "model_path": "/Users/sagawa/AI_WORK/cache/huggingface/hub/models--mlx-community--whisper-large-v3",
            "language": "zh",
            "audio_interval": [
              52.32000000000001,
              58.660000000000004
            ],
            "target_interval": [
              54.32000000000001,
              56.660000000000004
            ],
            "text": "调好位置还能一键锁定防务处",
            "words": [
              {
                "word": "调",
                "start": 54.400000000000006,
                "end": 54.720000000000006,
                "probability": 0.947265625
              },
              {
                "word": "好",
                "start": 54.720000000000006,
                "end": 54.800000000000004,
                "probability": 0.998046875
              },
              {
                "word": "位",
                "start": 54.800000000000004,
                "end": 54.940000000000005,
                "probability": 0.9990234375
              },
              {
                "word": "置",
                "start": 54.940000000000005,
                "end": 55.14000000000001,
                "probability": 0.99951171875
              },
              {
                "word": "还",
                "start": 55.14000000000001,
                "end": 55.34000000000001,
                "probability": 0.83349609375
              },
              {
                "word": "能",
                "start": 55.34000000000001,
                "end": 55.480000000000004,
                "probability": 0.99951171875
              },
              {
                "word": "一",
                "start": 55.480000000000004,
                "end": 55.620000000000005,
                "probability": 0.9970703125
              },
              {
                "word": "键",
                "start": 55.620000000000005,
                "end": 55.82000000000001,
                "probability": 0.9990234375
              },
              {
                "word": "锁",
                "start": 55.82000000000001,
                "end": 56.00000000000001,
                "probability": 1.0
              },
              {
                "word": "定",
                "start": 56.00000000000001,
                "end": 56.10000000000001,
                "probability": 0.99755859375
              },
              {
                "word": "防",
                "start": 56.10000000000001,
                "end": 56.24000000000001,
                "probability": 0.75439453125
              },
              {
                "word": "务",
                "start": 56.24000000000001,
                "end": 56.440000000000005,
                "probability": 0.5986328125
              },
              {
                "word": "处",
                "start": 56.440000000000005,
                "end": 56.60000000000001,
                "probability": 0.998046875
              }
            ],
            "raw_result": {
              "text": "大小 光配色调好位置还能一键锁定防务处你可以选边听边出",
              "segments": [
                {
                  "id": 0,
                  "seek": 0,
                  "start": 0.0,
                  "end": 0.98,
                  "text": "大小 光配色",
                  "tokens": [
                    50365,
                    3582,
                    7322,
                    220,
                    20690,
                    38846,
                    17673,
                    50421
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.23005850315093995,
                  "compression_ratio": 0.8777777777777778,
                  "no_speech_prob": 0.9049589037895203,
                  "words": [
                    {
                      "word": "大",
                      "start": 0.0,
                      "end": 0.1,
                      "probability": 0.00997161865234375
                    },
                    {
                      "word": "小",
                      "start": 0.1,
                      "end": 0.4,
                      "probability": 0.98974609375
                    },
                    {
                      "word": " 光",
                      "start": 0.42,
                      "end": 0.66,
                      "probability": 0.91455078125
                    },
                    {
                      "word": "配",
                      "start": 0.66,
                      "end": 0.78,
                      "probability": 0.9501953125
                    },
                    {
                      "word": "色",
                      "start": 0.78,
                      "end": 0.98,
                      "probability": 1.0
                    }
                  ]
                },
                {
                  "id": 1,
                  "seek": 0,
                  "start": 2.0799999999999996,
                  "end": 2.82,
                  "text": "调好位置",
                  "tokens": [
                    50421,
                    8897,
                    225,
                    2131,
                    11160,
                    34719,
                    50509
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.23005850315093995,
                  "compression_ratio": 0.8777777777777778,
                  "no_speech_prob": 0.9049589037895203,
                  "words": [
                    {
                      "word": "调",
                      "start": 2.0799999999999996,
                      "end": 2.4,
                      "probability": 0.947265625
                    },
                    {
                      "word": "好",
                      "start": 2.4,
                      "end": 2.48,
                      "probability": 0.998046875
                    },
                    {
                      "word": "位",
                      "start": 2.48,
                      "end": 2.62,
                      "probability": 0.9990234375
                    },
                    {
                      "word": "置",
                      "start": 2.62,
                      "end": 2.82,
                      "probability": 0.99951171875
                    }
                  ]
                },
                {
                  "id": 2,
                  "seek": 0,
                  "start": 2.82,
                  "end": 4.28,
                  "text": "还能一键锁定防务处",
                  "tokens": [
                    50509,
                    14852,
                    8225,
                    2257,
                    23049,
                    106,
                    23049,
                    223,
                    12088,
                    35863,
                    5087,
                    94,
                    1787,
                    226,
                    50582
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.23005850315093995,
                  "compression_ratio": 0.8777777777777778,
                  "no_speech_prob": 0.9049589037895203,
                  "words": [
                    {
                      "word": "还",
                      "start": 2.82,
                      "end": 3.02,
                      "probability": 0.83349609375
                    },
                    {
                      "word": "能",
                      "start": 3.02,
                      "end": 3.16,
                      "probability": 0.99951171875
                    },
                    {
                      "word": "一",
                      "start": 3.16,
                      "end": 3.3,
                      "probability": 0.9970703125
                    },
                    {
                      "word": "键",
                      "start": 3.3,
                      "end": 3.5,
                      "probability": 0.9990234375
                    },
                    {
                      "word": "锁",
                      "start": 3.5,
                      "end": 3.68,
                      "probability": 1.0
                    },
                    {
                      "word": "定",
                      "start": 3.68,
                      "end": 3.78,
                      "probability": 0.99755859375
                    },
                    {
                      "word": "防",
                      "start": 3.78,
                      "end": 3.92,
                      "probability": 0.75439453125
                    },
                    {
                      "word": "务",
                      "start": 3.92,
                      "end": 4.12,
                      "probability": 0.5986328125
                    },
                    {
                      "word": "处",
                      "start": 4.12,
                      "end": 4.28,
                      "probability": 0.998046875
                    }
                  ]
                },
                {
                  "id": 3,
                  "seek": 0,
                  "start": 5.02,
                  "end": 6.22,
                  "text": "你可以选边听边出",
                  "tokens": [
                    50582,
                    42766,
                    2215,
                    231,
                    40503,
                    31022,
                    40503,
                    7781,
                    50683
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.23005850315093995,
                  "compression_ratio": 0.8777777777777778,
                  "no_speech_prob": 0.9049589037895203,
                  "words": [
                    {
                      "word": "你可以",
                      "start": 5.02,
                      "end": 5.34,
                      "probability": 0.99169921875
                    },
                    {
                      "word": "选",
                      "start": 5.34,
                      "end": 5.66,
                      "probability": 0.9990234375
                    },
                    {
                      "word": "边",
                      "start": 5.66,
                      "end": 5.8,
                      "probability": 0.99169921875
                    },
                    {
                      "word": "听",
                      "start": 5.8,
                      "end": 6.0,
                      "probability": 0.62158203125
                    },
                    {
                      "word": "边",
                      "start": 6.0,
                      "end": 6.16,
                      "probability": 0.998046875
                    },
                    {
                      "word": "出",
                      "start": 6.16,
                      "end": 6.22,
                      "probability": 0.67626953125
                    }
                  ]
                }
              ],
              "language": "zh",
              "duration": 6.34
            },
            "status": "COMPLETE",
            "requested_language": "zh"
          },
          "suppression_proposed": null
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "26",
          "before": "房屋处",
          "after": "房屋处",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "Nonsense: '房屋处' (house place) vs '调好位置' (adjust position)",
          "context_evidence": "ASR: '调好位置还能一键锁定防务处'",
          "independent_ASR": {
            "engine": "whisper-large-v3",
            "model_path": "/Users/sagawa/AI_WORK/cache/huggingface/hub/models--mlx-community--whisper-large-v3",
            "language": "zh",
            "audio_interval": [
              52.32000000000001,
              58.660000000000004
            ],
            "target_interval": [
              54.32000000000001,
              56.660000000000004
            ],
            "text": "调好位置还能一键锁定防务处",
            "words": [
              {
                "word": "调",
                "start": 54.400000000000006,
                "end": 54.720000000000006,
                "probability": 0.947265625
              },
              {
                "word": "好",
                "start": 54.720000000000006,
                "end": 54.800000000000004,
                "probability": 0.998046875
              },
              {
                "word": "位",
                "start": 54.800000000000004,
                "end": 54.940000000000005,
                "probability": 0.9990234375
              },
              {
                "word": "置",
                "start": 54.940000000000005,
                "end": 55.14000000000001,
                "probability": 0.99951171875
              },
              {
                "word": "还",
                "start": 55.14000000000001,
                "end": 55.34000000000001,
                "probability": 0.83349609375
              },
              {
                "word": "能",
                "start": 55.34000000000001,
                "end": 55.480000000000004,
                "probability": 0.99951171875
              },
              {
                "word": "一",
                "start": 55.480000000000004,
                "end": 55.620000000000005,
                "probability": 0.9970703125
              },
              {
                "word": "键",
                "start": 55.620000000000005,
                "end": 55.82000000000001,
                "probability": 0.9990234375
              },
              {
                "word": "锁",
                "start": 55.82000000000001,
                "end": 56.00000000000001,
                "probability": 1.0
              },
              {
                "word": "定",
                "start": 56.00000000000001,
                "end": 56.10000000000001,
                "probability": 0.99755859375
              },
              {
                "word": "防",
                "start": 56.10000000000001,
                "end": 56.24000000000001,
                "probability": 0.75439453125
              },
              {
                "word": "务",
                "start": 56.24000000000001,
                "end": 56.440000000000005,
                "probability": 0.5986328125
              },
              {
                "word": "处",
                "start": 56.440000000000005,
                "end": 56.60000000000001,
                "probability": 0.998046875
              }
            ],
            "raw_result": {
              "text": "大小 光配色调好位置还能一键锁定防务处你可以选边听边出",
              "segments": [
                {
                  "id": 0,
                  "seek": 0,
                  "start": 0.0,
                  "end": 0.98,
                  "text": "大小 光配色",
                  "tokens": [
                    50365,
                    3582,
                    7322,
                    220,
                    20690,
                    38846,
                    17673,
                    50421
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.23005850315093995,
                  "compression_ratio": 0.8777777777777778,
                  "no_speech_prob": 0.9049589037895203,
                  "words": [
                    {
                      "word": "大",
                      "start": 0.0,
                      "end": 0.1,
                      "probability": 0.00997161865234375
                    },
                    {
                      "word": "小",
                      "start": 0.1,
                      "end": 0.4,
                      "probability": 0.98974609375
                    },
                    {
                      "word": " 光",
                      "start": 0.42,
                      "end": 0.66,
                      "probability": 0.91455078125
                    },
                    {
                      "word": "配",
                      "start": 0.66,
                      "end": 0.78,
                      "probability": 0.9501953125
                    },
                    {
                      "word": "色",
                      "start": 0.78,
                      "end": 0.98,
                      "probability": 1.0
                    }
                  ]
                },
                {
                  "id": 1,
                  "seek": 0,
                  "start": 2.0799999999999996,
                  "end": 2.82,
                  "text": "调好位置",
                  "tokens": [
                    50421,
                    8897,
                    225,
                    2131,
                    11160,
                    34719,
                    50509
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.23005850315093995,
                  "compression_ratio": 0.8777777777777778,
                  "no_speech_prob": 0.9049589037895203,
                  "words": [
                    {
                      "word": "调",
                      "start": 2.0799999999999996,
                      "end": 2.4,
                      "probability": 0.947265625
                    },
                    {
                      "word": "好",
                      "start": 2.4,
                      "end": 2.48,
                      "probability": 0.998046875
                    },
                    {
                      "word": "位",
                      "start": 2.48,
                      "end": 2.62,
                      "probability": 0.9990234375
                    },
                    {
                      "word": "置",
                      "start": 2.62,
                      "end": 2.82,
                      "probability": 0.99951171875
                    }
                  ]
                },
                {
                  "id": 2,
                  "seek": 0,
                  "start": 2.82,
                  "end": 4.28,
                  "text": "还能一键锁定防务处",
                  "tokens": [
                    50509,
                    14852,
                    8225,
                    2257,
                    23049,
                    106,
                    23049,
                    223,
                    12088,
                    35863,
                    5087,
                    94,
                    1787,
                    226,
                    50582
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.23005850315093995,
                  "compression_ratio": 0.8777777777777778,
                  "no_speech_prob": 0.9049589037895203,
                  "words": [
                    {
                      "word": "还",
                      "start": 2.82,
                      "end": 3.02,
                      "probability": 0.83349609375
                    },
                    {
                      "word": "能",
                      "start": 3.02,
                      "end": 3.16,
                      "probability": 0.99951171875
                    },
                    {
                      "word": "一",
                      "start": 3.16,
                      "end": 3.3,
                      "probability": 0.9970703125
                    },
                    {
                      "word": "键",
                      "start": 3.3,
                      "end": 3.5,
                      "probability": 0.9990234375
                    },
                    {
                      "word": "锁",
                      "start": 3.5,
                      "end": 3.68,
                      "probability": 1.0
                    },
                    {
                      "word": "定",
                      "start": 3.68,
                      "end": 3.78,
                      "probability": 0.99755859375
                    },
                    {
                      "word": "防",
                      "start": 3.78,
                      "end": 3.92,
                      "probability": 0.75439453125
                    },
                    {
                      "word": "务",
                      "start": 3.92,
                      "end": 4.12,
                      "probability": 0.5986328125
                    },
                    {
                      "word": "处",
                      "start": 4.12,
                      "end": 4.28,
                      "probability": 0.998046875
                    }
                  ]
                },
                {
                  "id": 3,
                  "seek": 0,
                  "start": 5.02,
                  "end": 6.22,
                  "text": "你可以选边听边出",
                  "tokens": [
                    50582,
                    42766,
                    2215,
                    231,
                    40503,
                    31022,
                    40503,
                    7781,
                    50683
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.23005850315093995,
                  "compression_ratio": 0.8777777777777778,
                  "no_speech_prob": 0.9049589037895203,
                  "words": [
                    {
                      "word": "你可以",
                      "start": 5.02,
                      "end": 5.34,
                      "probability": 0.99169921875
                    },
                    {
                      "word": "选",
                      "start": 5.34,
                      "end": 5.66,
                      "probability": 0.9990234375
                    },
                    {
                      "word": "边",
                      "start": 5.66,
                      "end": 5.8,
                      "probability": 0.99169921875
                    },
                    {
                      "word": "听",
                      "start": 5.8,
                      "end": 6.0,
                      "probability": 0.62158203125
                    },
                    {
                      "word": "边",
                      "start": 6.0,
                      "end": 6.16,
                      "probability": 0.998046875
                    },
                    {
                      "word": "出",
                      "start": 6.16,
                      "end": 6.22,
                      "probability": 0.67626953125
                    }
                  ]
                }
              ],
              "language": "zh",
              "duration": 6.34
            },
            "status": "COMPLETE",
            "requested_language": "zh"
          },
          "suppression_proposed": null
        }
      ],
      "language": "zh",
      "packet_ids": [
        2
      ],
      "classification": "BEST_SUPPORTED_WITH_REVIEW",
      "independent_ASR": {
        "engine": "whisper-large-v3",
        "model_path": "/Users/sagawa/AI_WORK/cache/huggingface/hub/models--mlx-community--whisper-large-v3",
        "language": "zh",
        "audio_interval": [
          52.32000000000001,
          58.660000000000004
        ],
        "target_interval": [
          54.32000000000001,
          56.660000000000004
        ],
        "text": "调好位置还能一键锁定防务处",
        "words": [
          {
            "word": "调",
            "start": 54.400000000000006,
            "end": 54.720000000000006,
            "probability": 0.947265625
          },
          {
            "word": "好",
            "start": 54.720000000000006,
            "end": 54.800000000000004,
            "probability": 0.998046875
          },
          {
            "word": "位",
            "start": 54.800000000000004,
            "end": 54.940000000000005,
            "probability": 0.9990234375
          },
          {
            "word": "置",
            "start": 54.940000000000005,
            "end": 55.14000000000001,
            "probability": 0.99951171875
          },
          {
            "word": "还",
            "start": 55.14000000000001,
            "end": 55.34000000000001,
            "probability": 0.83349609375
          },
          {
            "word": "能",
            "start": 55.34000000000001,
            "end": 55.480000000000004,
            "probability": 0.99951171875
          },
          {
            "word": "一",
            "start": 55.480000000000004,
            "end": 55.620000000000005,
            "probability": 0.9970703125
          },
          {
            "word": "键",
            "start": 55.620000000000005,
            "end": 55.82000000000001,
            "probability": 0.9990234375
          },
          {
            "word": "锁",
            "start": 55.82000000000001,
            "end": 56.00000000000001,
            "probability": 1.0
          },
          {
            "word": "定",
            "start": 56.00000000000001,
            "end": 56.10000000000001,
            "probability": 0.99755859375
          },
          {
            "word": "防",
            "start": 56.10000000000001,
            "end": 56.24000000000001,
            "probability": 0.75439453125
          },
          {
            "word": "务",
            "start": 56.24000000000001,
            "end": 56.440000000000005,
            "probability": 0.5986328125
          },
          {
            "word": "处",
            "start": 56.440000000000005,
            "end": 56.60000000000001,
            "probability": 0.998046875
          }
        ],
        "raw_result": {
          "text": "大小 光配色调好位置还能一键锁定防务处你可以选边听边出",
          "segments": [
            {
              "id": 0,
              "seek": 0,
              "start": 0.0,
              "end": 0.98,
              "text": "大小 光配色",
              "tokens": [
                50365,
                3582,
                7322,
                220,
                20690,
                38846,
                17673,
                50421
              ],
              "temperature": 0.0,
              "avg_logprob": -0.23005850315093995,
              "compression_ratio": 0.8777777777777778,
              "no_speech_prob": 0.9049589037895203,
              "words": [
                {
                  "word": "大",
                  "start": 0.0,
                  "end": 0.1,
                  "probability": 0.00997161865234375
                },
                {
                  "word": "小",
                  "start": 0.1,
                  "end": 0.4,
                  "probability": 0.98974609375
                },
                {
                  "word": " 光",
                  "start": 0.42,
                  "end": 0.66,
                  "probability": 0.91455078125
                },
                {
                  "word": "配",
                  "start": 0.66,
                  "end": 0.78,
                  "probability": 0.9501953125
                },
                {
                  "word": "色",
                  "start": 0.78,
                  "end": 0.98,
                  "probability": 1.0
                }
              ]
            },
            {
              "id": 1,
              "seek": 0,
              "start": 2.0799999999999996,
              "end": 2.82,
              "text": "调好位置",
              "tokens": [
                50421,
                8897,
                225,
                2131,
                11160,
                34719,
                50509
              ],
              "temperature": 0.0,
              "avg_logprob": -0.23005850315093995,
              "compression_ratio": 0.8777777777777778,
              "no_speech_prob": 0.9049589037895203,
              "words": [
                {
                  "word": "调",
                  "start": 2.0799999999999996,
                  "end": 2.4,
                  "probability": 0.947265625
                },
                {
                  "word": "好",
                  "start": 2.4,
                  "end": 2.48,
                  "probability": 0.998046875
                },
                {
                  "word": "位",
                  "start": 2.48,
                  "end": 2.62,
                  "probability": 0.9990234375
                },
                {
                  "word": "置",
                  "start": 2.62,
                  "end": 2.82,
                  "probability": 0.99951171875
                }
              ]
            },
            {
              "id": 2,
              "seek": 0,
              "start": 2.82,
              "end": 4.28,
              "text": "还能一键锁定防务处",
              "tokens": [
                50509,
                14852,
                8225,
                2257,
                23049,
                106,
                23049,
                223,
                12088,
                35863,
                5087,
                94,
                1787,
                226,
                50582
              ],
              "temperature": 0.0,
              "avg_logprob": -0.23005850315093995,
              "compression_ratio": 0.8777777777777778,
              "no_speech_prob": 0.9049589037895203,
              "words": [
                {
                  "word": "还",
                  "start": 2.82,
                  "end": 3.02,
                  "probability": 0.83349609375
                },
                {
                  "word": "能",
                  "start": 3.02,
                  "end": 3.16,
                  "probability": 0.99951171875
                },
                {
                  "word": "一",
                  "start": 3.16,
                  "end": 3.3,
                  "probability": 0.9970703125
                },
                {
                  "word": "键",
                  "start": 3.3,
                  "end": 3.5,
                  "probability": 0.9990234375
                },
                {
                  "word": "锁",
                  "start": 3.5,
                  "end": 3.68,
                  "probability": 1.0
                },
                {
                  "word": "定",
                  "start": 3.68,
                  "end": 3.78,
                  "probability": 0.99755859375
                },
                {
                  "word": "防",
                  "start": 3.78,
                  "end": 3.92,
                  "probability": 0.75439453125
                },
                {
                  "word": "务",
                  "start": 3.92,
                  "end": 4.12,
                  "probability": 0.5986328125
                },
                {
                  "word": "处",
                  "start": 4.12,
                  "end": 4.28,
                  "probability": 0.998046875
                }
              ]
            },
            {
              "id": 3,
              "seek": 0,
              "start": 5.02,
              "end": 6.22,
              "text": "你可以选边听边出",
              "tokens": [
                50582,
                42766,
                2215,
                231,
                40503,
                31022,
                40503,
                7781,
                50683
              ],
              "temperature": 0.0,
              "avg_logprob": -0.23005850315093995,
              "compression_ratio": 0.8777777777777778,
              "no_speech_prob": 0.9049589037895203,
              "words": [
                {
                  "word": "你可以",
                  "start": 5.02,
                  "end": 5.34,
                  "probability": 0.99169921875
                },
                {
                  "word": "选",
                  "start": 5.34,
                  "end": 5.66,
                  "probability": 0.9990234375
                },
                {
                  "word": "边",
                  "start": 5.66,
                  "end": 5.8,
                  "probability": 0.99169921875
                },
                {
                  "word": "听",
                  "start": 5.8,
                  "end": 6.0,
                  "probability": 0.62158203125
                },
                {
                  "word": "边",
                  "start": 6.0,
                  "end": 6.16,
                  "probability": 0.998046875
                },
                {
                  "word": "出",
                  "start": 6.16,
                  "end": 6.22,
                  "probability": 0.67626953125
                }
              ]
            }
          ],
          "language": "zh",
          "duration": 6.34
        },
        "status": "COMPLETE",
        "requested_language": "zh"
      },
      "raw_word_timestamps": [
        {
          "word": "调",
          "start": 54.32000000000001,
          "end": 54.68000000000001,
          "probability": 0.8076171875
        },
        {
          "word": "好",
          "start": 54.68000000000001,
          "end": 54.86000000000001,
          "probability": 0.96044921875
        },
        {
          "word": "位",
          "start": 54.86000000000001,
          "end": 55.02,
          "probability": 0.99755859375
        },
        {
          "word": "置",
          "start": 55.02,
          "end": 55.18000000000001,
          "probability": 0.99951171875
        },
        {
          "word": "还",
          "start": 55.18000000000001,
          "end": 55.400000000000006,
          "probability": 0.9501953125
        },
        {
          "word": "能",
          "start": 55.400000000000006,
          "end": 55.540000000000006,
          "probability": 0.9970703125
        },
        {
          "word": "一",
          "start": 55.540000000000006,
          "end": 55.68000000000001,
          "probability": 0.99609375
        },
        {
          "word": "键",
          "start": 55.68000000000001,
          "end": 55.82000000000001,
          "probability": 0.998046875
        },
        {
          "word": "锁",
          "start": 55.82000000000001,
          "end": 56.02,
          "probability": 0.99658203125
        },
        {
          "word": "定",
          "start": 56.02,
          "end": 56.160000000000004,
          "probability": 0.98876953125
        },
        {
          "word": "房",
          "start": 56.160000000000004,
          "end": 56.28,
          "probability": 0.50830078125
        },
        {
          "word": "屋",
          "start": 56.28,
          "end": 56.480000000000004,
          "probability": 0.9287109375
        },
        {
          "word": "处",
          "start": 56.480000000000004,
          "end": 56.660000000000004,
          "probability": 0.98046875
        }
      ],
      "word_timestamps": [],
      "source_text": "房屋处",
      "review_status": "BEST_SUPPORTED_WITH_REVIEW",
      "canonical_timing_evidence": "CUE_ANCHORED",
      "index": 26
    },
    {
      "start": 57.32000000000001,
      "end": 58.74,
      "text": "你可以选边听边出",
      "raw_asr_text": "你可以选边听边出",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 2,
      "words": [
        {
          "word": "你可以",
          "start": 57.32000000000001,
          "end": 57.68000000000001,
          "probability": 0.951171875
        },
        {
          "word": "选",
          "start": 57.68000000000001,
          "end": 58.00000000000001,
          "probability": 0.99853515625
        },
        {
          "word": "边",
          "start": 58.00000000000001,
          "end": 58.160000000000004,
          "probability": 0.9560546875
        },
        {
          "word": "听",
          "start": 58.160000000000004,
          "end": 58.34,
          "probability": 0.99755859375
        },
        {
          "word": "边",
          "start": 58.34,
          "end": 58.50000000000001,
          "probability": 0.994140625
        },
        {
          "word": "出",
          "start": 58.50000000000001,
          "end": 58.74,
          "probability": 0.990234375
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.11145266908587831,
        "compression_ratio": 1.2433333333333334,
        "no_speech_prob": 6.322219137100404e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "调好位置还能一键锁定房屋处你可以选边听边出草稿随说随出一句说完立刻变正式字幕也可以切成稳稳地说完再出在手机上你可以单独选某一个会出声的APP来听到了电脑上它直接听整台电脑播放的声音哪怕关掉主窗口字幕也照样在Mac的菜单栏和Windows的右下角拖盘点一下就能随时开关因"
        }
      ],
      "id": 27,
      "review_annotation": "",
      "row_id": "27",
      "bounded_row_id": 27,
      "parent_row_ids": [
        27
      ],
      "parent_row_id": 27,
      "packet_id": 2,
      "LID_evidence": [
        {
          "region_id": 11,
          "probabilities": {
            "zh": 0.997185610807859
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        11
      ],
      "language_confidence": 0.997185610807859,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 53.300000000000004,
      "packet_end": 81.2,
      "raw_text": "你可以选边听边出",
      "selected_text": "你可以选边听边出",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            27
          ],
          "packet_id": 2,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        2
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 27
    },
    {
      "start": 58.74,
      "end": 60.2,
      "text": "草稿随说随出",
      "raw_asr_text": "草稿随说随出",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 2,
      "words": [
        {
          "word": "草",
          "start": 58.74,
          "end": 59.28,
          "probability": 0.978515625
        },
        {
          "word": "稿",
          "start": 59.28,
          "end": 59.440000000000005,
          "probability": 0.99609375
        },
        {
          "word": "随",
          "start": 59.440000000000005,
          "end": 59.720000000000006,
          "probability": 0.9697265625
        },
        {
          "word": "说",
          "start": 59.720000000000006,
          "end": 59.84,
          "probability": 0.9970703125
        },
        {
          "word": "随",
          "start": 59.84,
          "end": 60.02,
          "probability": 0.994140625
        },
        {
          "word": "出",
          "start": 60.02,
          "end": 60.2,
          "probability": 0.98828125
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.11145266908587831,
        "compression_ratio": 1.2433333333333334,
        "no_speech_prob": 6.322219137100404e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "调好位置还能一键锁定房屋处你可以选边听边出草稿随说随出一句说完立刻变正式字幕也可以切成稳稳地说完再出在手机上你可以单独选某一个会出声的APP来听到了电脑上它直接听整台电脑播放的声音哪怕关掉主窗口字幕也照样在Mac的菜单栏和Windows的右下角拖盘点一下就能随时开关因"
        }
      ],
      "id": 28,
      "review_annotation": "",
      "row_id": "28",
      "bounded_row_id": 28,
      "parent_row_ids": [
        28
      ],
      "parent_row_id": 28,
      "packet_id": 2,
      "LID_evidence": [
        {
          "region_id": 11,
          "probabilities": {
            "zh": 0.997185610807859
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        11
      ],
      "language_confidence": 0.997185610807859,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 53.300000000000004,
      "packet_end": 81.2,
      "raw_text": "草稿随说随出",
      "selected_text": "草稿随说随出",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            28
          ],
          "packet_id": 2,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        2
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 28
    },
    {
      "start": 60.2,
      "end": 62.38,
      "text": "一句说完立刻变正式字幕",
      "raw_asr_text": "一句说完立刻变正式字幕",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 2,
      "words": [
        {
          "word": "一",
          "start": 60.2,
          "end": 60.660000000000004,
          "probability": 0.9560546875
        },
        {
          "word": "句",
          "start": 60.660000000000004,
          "end": 60.78,
          "probability": 0.998046875
        },
        {
          "word": "说",
          "start": 60.78,
          "end": 60.92,
          "probability": 0.99755859375
        },
        {
          "word": "完",
          "start": 60.92,
          "end": 61.1,
          "probability": 0.98583984375
        },
        {
          "word": "立",
          "start": 61.1,
          "end": 61.400000000000006,
          "probability": 0.96875
        },
        {
          "word": "刻",
          "start": 61.400000000000006,
          "end": 61.540000000000006,
          "probability": 0.90966796875
        },
        {
          "word": "变",
          "start": 61.540000000000006,
          "end": 61.720000000000006,
          "probability": 0.8720703125
        },
        {
          "word": "正",
          "start": 61.720000000000006,
          "end": 61.92,
          "probability": 0.97314453125
        },
        {
          "word": "式",
          "start": 61.92,
          "end": 62.06,
          "probability": 0.97216796875
        },
        {
          "word": "字",
          "start": 62.06,
          "end": 62.18000000000001,
          "probability": 0.9921875
        },
        {
          "word": "幕",
          "start": 62.18000000000001,
          "end": 62.38,
          "probability": 0.966796875
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.11145266908587831,
        "compression_ratio": 1.2433333333333334,
        "no_speech_prob": 6.322219137100404e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "调好位置还能一键锁定房屋处你可以选边听边出草稿随说随出一句说完立刻变正式字幕也可以切成稳稳地说完再出在手机上你可以单独选某一个会出声的APP来听到了电脑上它直接听整台电脑播放的声音哪怕关掉主窗口字幕也照样在Mac的菜单栏和Windows的右下角拖盘点一下就能随时开关因"
        }
      ],
      "id": 29,
      "review_annotation": "",
      "row_id": "29",
      "bounded_row_id": 29,
      "parent_row_ids": [
        29
      ],
      "parent_row_id": 29,
      "packet_id": 2,
      "LID_evidence": [
        {
          "region_id": 11,
          "probabilities": {
            "zh": 0.997185610807859
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        11
      ],
      "language_confidence": 0.997185610807859,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 53.300000000000004,
      "packet_end": 81.2,
      "raw_text": "一句说完立刻变正式字幕",
      "selected_text": "一句说完立刻变正式字幕",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            29
          ],
          "packet_id": 2,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        2
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 29
    },
    {
      "start": 62.38,
      "end": 64.7,
      "text": "也可以切成稳稳地说完再出",
      "raw_asr_text": "也可以切成稳稳地说完再出",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 2,
      "words": [
        {
          "word": "也可以",
          "start": 62.38,
          "end": 63.0,
          "probability": 0.98046875
        },
        {
          "word": "切",
          "start": 63.0,
          "end": 63.28,
          "probability": 0.9765625
        },
        {
          "word": "成",
          "start": 63.28,
          "end": 63.480000000000004,
          "probability": 0.994140625
        },
        {
          "word": "稳",
          "start": 63.480000000000004,
          "end": 63.82000000000001,
          "probability": 0.9873046875
        },
        {
          "word": "稳",
          "start": 63.82000000000001,
          "end": 63.940000000000005,
          "probability": 0.99951171875
        },
        {
          "word": "地",
          "start": 63.940000000000005,
          "end": 64.04,
          "probability": 0.40625
        },
        {
          "word": "说",
          "start": 64.04,
          "end": 64.18,
          "probability": 0.998046875
        },
        {
          "word": "完",
          "start": 64.18,
          "end": 64.30000000000001,
          "probability": 0.98974609375
        },
        {
          "word": "再",
          "start": 64.30000000000001,
          "end": 64.46000000000001,
          "probability": 0.994140625
        },
        {
          "word": "出",
          "start": 64.46000000000001,
          "end": 64.7,
          "probability": 0.99658203125
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.11145266908587831,
        "compression_ratio": 1.2433333333333334,
        "no_speech_prob": 6.322219137100404e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "调好位置还能一键锁定房屋处你可以选边听边出草稿随说随出一句说完立刻变正式字幕也可以切成稳稳地说完再出在手机上你可以单独选某一个会出声的APP来听到了电脑上它直接听整台电脑播放的声音哪怕关掉主窗口字幕也照样在Mac的菜单栏和Windows的右下角拖盘点一下就能随时开关因"
        }
      ],
      "id": 30,
      "review_annotation": "",
      "row_id": "30",
      "bounded_row_id": 30,
      "parent_row_ids": [
        30
      ],
      "parent_row_id": 30,
      "packet_id": 2,
      "LID_evidence": [
        {
          "region_id": 11,
          "probabilities": {
            "zh": 0.997185610807859
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        11
      ],
      "language_confidence": 0.997185610807859,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 53.300000000000004,
      "packet_end": 81.2,
      "raw_text": "也可以切成稳稳地说完再出",
      "selected_text": "也可以切成稳稳地说完再出",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            30
          ],
          "packet_id": 2,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        2
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 30
    },
    {
      "start": 65.58000000000001,
      "end": 66.4,
      "text": "在手机上",
      "raw_asr_text": "在手机上",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 2,
      "words": [
        {
          "word": "在",
          "start": 65.58000000000001,
          "end": 65.94,
          "probability": 0.98486328125
        },
        {
          "word": "手",
          "start": 65.94,
          "end": 66.10000000000001,
          "probability": 0.9990234375
        },
        {
          "word": "机",
          "start": 66.10000000000001,
          "end": 66.22,
          "probability": 0.9990234375
        },
        {
          "word": "上",
          "start": 66.22,
          "end": 66.4,
          "probability": 0.99853515625
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.11145266908587831,
        "compression_ratio": 1.2433333333333334,
        "no_speech_prob": 6.322219137100404e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "调好位置还能一键锁定房屋处你可以选边听边出草稿随说随出一句说完立刻变正式字幕也可以切成稳稳地说完再出在手机上你可以单独选某一个会出声的APP来听到了电脑上它直接听整台电脑播放的声音哪怕关掉主窗口字幕也照样在Mac的菜单栏和Windows的右下角拖盘点一下就能随时开关因"
        }
      ],
      "id": 31,
      "review_annotation": "",
      "row_id": "31",
      "bounded_row_id": 31,
      "parent_row_ids": [
        31
      ],
      "parent_row_id": 31,
      "packet_id": 2,
      "LID_evidence": [
        {
          "region_id": 12,
          "probabilities": {
            "zh": 0.9983474769090351
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        12
      ],
      "language_confidence": 0.9983474769090351,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 53.300000000000004,
      "packet_end": 81.2,
      "raw_text": "在手机上",
      "selected_text": "在手机上",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            31
          ],
          "packet_id": 2,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        2
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 31
    },
    {
      "start": 66.4,
      "end": 69.10000000000001,
      "text": "你可以单独选某一个会出声的APP来听",
      "raw_asr_text": "你可以单独选某一个会出声的APP来听",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 2,
      "words": [
        {
          "word": "你可以",
          "start": 66.4,
          "end": 66.9,
          "probability": 0.951171875
        },
        {
          "word": "单",
          "start": 66.9,
          "end": 67.18,
          "probability": 0.9990234375
        },
        {
          "word": "独",
          "start": 67.18,
          "end": 67.34,
          "probability": 0.998046875
        },
        {
          "word": "选",
          "start": 67.34,
          "end": 67.52000000000001,
          "probability": 0.9794921875
        },
        {
          "word": "某",
          "start": 67.52000000000001,
          "end": 67.72,
          "probability": 0.99951171875
        },
        {
          "word": "一个",
          "start": 67.72,
          "end": 67.92,
          "probability": 0.99609375
        },
        {
          "word": "会",
          "start": 67.92,
          "end": 68.08,
          "probability": 0.94189453125
        },
        {
          "word": "出",
          "start": 68.08,
          "end": 68.24000000000001,
          "probability": 0.99462890625
        },
        {
          "word": "声",
          "start": 68.24000000000001,
          "end": 68.38000000000001,
          "probability": 0.9677734375
        },
        {
          "word": "的",
          "start": 68.38000000000001,
          "end": 68.5,
          "probability": 0.99755859375
        },
        {
          "word": "APP",
          "start": 68.5,
          "end": 68.68,
          "probability": 0.457275390625
        },
        {
          "word": "来",
          "start": 68.68,
          "end": 68.88000000000001,
          "probability": 0.98828125
        },
        {
          "word": "听",
          "start": 68.88000000000001,
          "end": 69.10000000000001,
          "probability": 0.99951171875
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.11145266908587831,
        "compression_ratio": 1.2433333333333334,
        "no_speech_prob": 6.322219137100404e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "调好位置还能一键锁定房屋处你可以选边听边出草稿随说随出一句说完立刻变正式字幕也可以切成稳稳地说完再出在手机上你可以单独选某一个会出声的APP来听到了电脑上它直接听整台电脑播放的声音哪怕关掉主窗口字幕也照样在Mac的菜单栏和Windows的右下角拖盘点一下就能随时开关因"
        }
      ],
      "id": 32,
      "review_annotation": "",
      "row_id": "32",
      "bounded_row_id": 32,
      "parent_row_ids": [
        32
      ],
      "parent_row_id": 32,
      "packet_id": 2,
      "LID_evidence": [
        {
          "region_id": 12,
          "probabilities": {
            "zh": 0.9983474769090351
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        12
      ],
      "language_confidence": 0.9983474769090351,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 53.300000000000004,
      "packet_end": 81.2,
      "raw_text": "你可以单独选某一个会出声的APP来听",
      "selected_text": "你可以单独选某一个会出声的APP来听",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            32
          ],
          "packet_id": 2,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        2
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 32
    },
    {
      "start": 69.10000000000001,
      "end": 70.22,
      "text": "到了电脑上",
      "raw_asr_text": "到了电脑上",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 2,
      "words": [
        {
          "word": "到了",
          "start": 69.10000000000001,
          "end": 69.7,
          "probability": 0.986328125
        },
        {
          "word": "电",
          "start": 69.7,
          "end": 69.9,
          "probability": 0.998046875
        },
        {
          "word": "脑",
          "start": 69.9,
          "end": 70.02000000000001,
          "probability": 1.0
        },
        {
          "word": "上",
          "start": 70.02000000000001,
          "end": 70.22,
          "probability": 0.9990234375
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.11145266908587831,
        "compression_ratio": 1.2433333333333334,
        "no_speech_prob": 6.322219137100404e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "调好位置还能一键锁定房屋处你可以选边听边出草稿随说随出一句说完立刻变正式字幕也可以切成稳稳地说完再出在手机上你可以单独选某一个会出声的APP来听到了电脑上它直接听整台电脑播放的声音哪怕关掉主窗口字幕也照样在Mac的菜单栏和Windows的右下角拖盘点一下就能随时开关因"
        }
      ],
      "id": 33,
      "review_annotation": "",
      "row_id": "33",
      "bounded_row_id": 33,
      "parent_row_ids": [
        33
      ],
      "parent_row_id": 33,
      "packet_id": 2,
      "LID_evidence": [
        {
          "region_id": 12,
          "probabilities": {
            "zh": 0.9983474769090351
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        12
      ],
      "language_confidence": 0.9983474769090351,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 53.300000000000004,
      "packet_end": 81.2,
      "raw_text": "到了电脑上",
      "selected_text": "到了电脑上",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            33
          ],
          "packet_id": 2,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        2
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 33
    },
    {
      "start": 70.22,
      "end": 72.64,
      "text": "它直接听整台电脑播放的声音",
      "raw_asr_text": "它直接听整台电脑播放的声音",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 2,
      "words": [
        {
          "word": "它",
          "start": 70.22,
          "end": 70.66,
          "probability": 0.82470703125
        },
        {
          "word": "直接",
          "start": 70.66,
          "end": 70.94,
          "probability": 0.99462890625
        },
        {
          "word": "听",
          "start": 70.94,
          "end": 71.16,
          "probability": 0.99853515625
        },
        {
          "word": "整",
          "start": 71.16,
          "end": 71.36,
          "probability": 0.98779296875
        },
        {
          "word": "台",
          "start": 71.36,
          "end": 71.52000000000001,
          "probability": 0.99462890625
        },
        {
          "word": "电",
          "start": 71.52000000000001,
          "end": 71.7,
          "probability": 0.99755859375
        },
        {
          "word": "脑",
          "start": 71.7,
          "end": 71.84,
          "probability": 1.0
        },
        {
          "word": "播",
          "start": 71.84,
          "end": 72.0,
          "probability": 0.99951171875
        },
        {
          "word": "放",
          "start": 72.0,
          "end": 72.16,
          "probability": 0.99951171875
        },
        {
          "word": "的",
          "start": 72.16,
          "end": 72.26,
          "probability": 0.99951171875
        },
        {
          "word": "声",
          "start": 72.26,
          "end": 72.42,
          "probability": 0.9990234375
        },
        {
          "word": "音",
          "start": 72.42,
          "end": 72.64,
          "probability": 1.0
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.11145266908587831,
        "compression_ratio": 1.2433333333333334,
        "no_speech_prob": 6.322219137100404e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "调好位置还能一键锁定房屋处你可以选边听边出草稿随说随出一句说完立刻变正式字幕也可以切成稳稳地说完再出在手机上你可以单独选某一个会出声的APP来听到了电脑上它直接听整台电脑播放的声音哪怕关掉主窗口字幕也照样在Mac的菜单栏和Windows的右下角拖盘点一下就能随时开关因"
        }
      ],
      "id": 34,
      "review_annotation": "",
      "row_id": "34",
      "bounded_row_id": 34,
      "parent_row_ids": [
        34
      ],
      "parent_row_id": 34,
      "packet_id": 2,
      "LID_evidence": [
        {
          "region_id": 12,
          "probabilities": {
            "zh": 0.9983474769090351
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        12
      ],
      "language_confidence": 0.9983474769090351,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 53.300000000000004,
      "packet_end": 81.2,
      "raw_text": "它直接听整台电脑播放的声音",
      "selected_text": "它直接听整台电脑播放的声音",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            34
          ],
          "packet_id": 2,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        2
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 34
    },
    {
      "start": 72.64,
      "end": 74.56,
      "text": "哪怕关掉主窗口字幕",
      "raw_asr_text": "哪怕关掉主窗口字幕",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 2,
      "words": [
        {
          "word": "哪",
          "start": 72.64,
          "end": 73.32000000000001,
          "probability": 0.9931640625
        },
        {
          "word": "怕",
          "start": 73.32000000000001,
          "end": 73.44,
          "probability": 1.0
        },
        {
          "word": "关",
          "start": 73.44,
          "end": 73.58000000000001,
          "probability": 0.99951171875
        },
        {
          "word": "掉",
          "start": 73.58000000000001,
          "end": 73.74000000000001,
          "probability": 0.9990234375
        },
        {
          "word": "主",
          "start": 73.74000000000001,
          "end": 73.9,
          "probability": 0.99267578125
        },
        {
          "word": "窗",
          "start": 73.9,
          "end": 74.04,
          "probability": 0.9619140625
        },
        {
          "word": "口",
          "start": 74.04,
          "end": 74.2,
          "probability": 0.9990234375
        },
        {
          "word": "字",
          "start": 74.2,
          "end": 74.36,
          "probability": 0.9912109375
        },
        {
          "word": "幕",
          "start": 74.36,
          "end": 74.56,
          "probability": 0.998046875
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.11145266908587831,
        "compression_ratio": 1.2433333333333334,
        "no_speech_prob": 6.322219137100404e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "调好位置还能一键锁定房屋处你可以选边听边出草稿随说随出一句说完立刻变正式字幕也可以切成稳稳地说完再出在手机上你可以单独选某一个会出声的APP来听到了电脑上它直接听整台电脑播放的声音哪怕关掉主窗口字幕也照样在Mac的菜单栏和Windows的右下角拖盘点一下就能随时开关因"
        }
      ],
      "id": 35,
      "review_annotation": "",
      "row_id": "35",
      "bounded_row_id": 35,
      "parent_row_ids": [
        35
      ],
      "parent_row_id": 35,
      "packet_id": 2,
      "LID_evidence": [
        {
          "region_id": 12,
          "probabilities": {
            "zh": 0.9983474769090351
          },
          "status": "STABLE"
        },
        {
          "region_id": 13,
          "probabilities": {
            "zh": 0.9956205354796516
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        12,
        13
      ],
      "language_confidence": 0.9983474769090351,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 53.300000000000004,
      "packet_end": 81.2,
      "raw_text": "哪怕关掉主窗口字幕",
      "selected_text": "哪怕关掉主窗口字幕",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            35
          ],
          "packet_id": 2,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        2
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 35
    },
    {
      "start": 74.56,
      "end": 75.36,
      "text": "也照样在",
      "raw_asr_text": "也照样在",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 2,
      "words": [
        {
          "word": "也",
          "start": 74.56,
          "end": 74.9,
          "probability": 0.96875
        },
        {
          "word": "照",
          "start": 74.9,
          "end": 75.04,
          "probability": 0.951171875
        },
        {
          "word": "样",
          "start": 75.04,
          "end": 75.18,
          "probability": 0.9921875
        },
        {
          "word": "在",
          "start": 75.18,
          "end": 75.36,
          "probability": 0.99169921875
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.11145266908587831,
        "compression_ratio": 1.2433333333333334,
        "no_speech_prob": 6.322219137100404e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "调好位置还能一键锁定房屋处你可以选边听边出草稿随说随出一句说完立刻变正式字幕也可以切成稳稳地说完再出在手机上你可以单独选某一个会出声的APP来听到了电脑上它直接听整台电脑播放的声音哪怕关掉主窗口字幕也照样在Mac的菜单栏和Windows的右下角拖盘点一下就能随时开关因"
        }
      ],
      "id": 36,
      "review_annotation": "",
      "row_id": "36",
      "bounded_row_id": 36,
      "parent_row_ids": [
        36
      ],
      "parent_row_id": 36,
      "packet_id": 2,
      "LID_evidence": [
        {
          "region_id": 13,
          "probabilities": {
            "zh": 0.9956205354796516
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        13
      ],
      "language_confidence": 0.9956205354796516,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 53.300000000000004,
      "packet_end": 81.2,
      "raw_text": "也照样在",
      "selected_text": "也照样在",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            36
          ],
          "packet_id": 2,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        2
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 36
    },
    {
      "start": 75.36,
      "end": 78.7,
      "text": "Mac的菜单栏和Windows的右下角拖盘点一下",
      "raw_asr_text": "Mac的菜单栏和Windows的右下角拖盘点一下",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 2,
      "words": [
        {
          "word": "Mac",
          "start": 75.36,
          "end": 76.06,
          "probability": 0.9501953125
        },
        {
          "word": "的",
          "start": 76.06,
          "end": 76.22,
          "probability": 0.98974609375
        },
        {
          "word": "菜",
          "start": 76.22,
          "end": 76.30000000000001,
          "probability": 0.947265625
        },
        {
          "word": "单",
          "start": 76.30000000000001,
          "end": 76.48,
          "probability": 0.9990234375
        },
        {
          "word": "栏",
          "start": 76.48,
          "end": 76.66,
          "probability": 0.99267578125
        },
        {
          "word": "和",
          "start": 76.66,
          "end": 76.96000000000001,
          "probability": 0.99072265625
        },
        {
          "word": "Wind",
          "start": 76.96000000000001,
          "end": 77.14,
          "probability": 0.943359375
        },
        {
          "word": "ows",
          "start": 77.14,
          "end": 77.30000000000001,
          "probability": 0.99658203125
        },
        {
          "word": "的",
          "start": 77.30000000000001,
          "end": 77.42,
          "probability": 0.97900390625
        },
        {
          "word": "右",
          "start": 77.42,
          "end": 77.56,
          "probability": 0.99755859375
        },
        {
          "word": "下",
          "start": 77.56,
          "end": 77.72,
          "probability": 0.998046875
        },
        {
          "word": "角",
          "start": 77.72,
          "end": 77.9,
          "probability": 0.994140625
        },
        {
          "word": "拖",
          "start": 77.9,
          "end": 78.08000000000001,
          "probability": 0.8271484375
        },
        {
          "word": "盘",
          "start": 78.08000000000001,
          "end": 78.26,
          "probability": 0.9384765625
        },
        {
          "word": "点",
          "start": 78.26,
          "end": 78.42,
          "probability": 0.994140625
        },
        {
          "word": "一下",
          "start": 78.42,
          "end": 78.7,
          "probability": 0.998046875
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.11145266908587831,
        "compression_ratio": 1.2433333333333334,
        "no_speech_prob": 6.322219137100404e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "调好位置还能一键锁定房屋处你可以选边听边出草稿随说随出一句说完立刻变正式字幕也可以切成稳稳地说完再出在手机上你可以单独选某一个会出声的APP来听到了电脑上它直接听整台电脑播放的声音哪怕关掉主窗口字幕也照样在Mac的菜单栏和Windows的右下角拖盘点一下就能随时开关因"
        }
      ],
      "id": 37,
      "review_annotation": "",
      "row_id": "37",
      "bounded_row_id": 37,
      "parent_row_ids": [
        37
      ],
      "parent_row_id": 37,
      "packet_id": 2,
      "LID_evidence": [
        {
          "region_id": 13,
          "probabilities": {
            "zh": 0.9956205354796516
          },
          "status": "STABLE"
        },
        {
          "region_id": 14,
          "probabilities": {
            "zh": 0.9878684282302856
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        13,
        14
      ],
      "language_confidence": 0.9956205354796516,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 53.300000000000004,
      "packet_end": 81.2,
      "raw_text": "Mac的菜单栏和Windows的右下角拖盘点一下",
      "selected_text": "Mac的菜单栏和Windows的右下角拖盘点一下",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            37
          ],
          "packet_id": 2,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        2
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 37
    },
    {
      "start": 78.7,
      "end": 79.76,
      "text": "就能随时开关",
      "raw_asr_text": "就能随时开关",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 2,
      "words": [
        {
          "word": "就",
          "start": 78.7,
          "end": 78.94,
          "probability": 0.994140625
        },
        {
          "word": "能",
          "start": 78.94,
          "end": 79.08000000000001,
          "probability": 0.99951171875
        },
        {
          "word": "随",
          "start": 79.08000000000001,
          "end": 79.28,
          "probability": 1.0
        },
        {
          "word": "时",
          "start": 79.28,
          "end": 79.4,
          "probability": 0.9990234375
        },
        {
          "word": "开",
          "start": 79.4,
          "end": 79.54,
          "probability": 0.9990234375
        },
        {
          "word": "关",
          "start": 79.54,
          "end": 79.76,
          "probability": 0.99560546875
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.11145266908587831,
        "compression_ratio": 1.2433333333333334,
        "no_speech_prob": 6.322219137100404e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "调好位置还能一键锁定房屋处你可以选边听边出草稿随说随出一句说完立刻变正式字幕也可以切成稳稳地说完再出在手机上你可以单独选某一个会出声的APP来听到了电脑上它直接听整台电脑播放的声音哪怕关掉主窗口字幕也照样在Mac的菜单栏和Windows的右下角拖盘点一下就能随时开关因"
        }
      ],
      "id": 38,
      "review_annotation": "",
      "row_id": "38",
      "bounded_row_id": 38,
      "parent_row_ids": [
        38
      ],
      "parent_row_id": 38,
      "packet_id": 2,
      "LID_evidence": [
        {
          "region_id": 14,
          "probabilities": {
            "zh": 0.9878684282302856
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        14
      ],
      "language_confidence": 0.9878684282302856,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 53.300000000000004,
      "packet_end": 81.2,
      "raw_text": "就能随时开关",
      "selected_text": "就能随时开关",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            38
          ],
          "packet_id": 2,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        2
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 38
    },
    {
      "start": 80.9,
      "end": 83.02000000000001,
      "text": "隐私方面也完全不用担心",
      "raw_asr_text": "隐私方面也完全不用担心",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 3,
      "words": [
        {
          "word": "隐",
          "start": 80.82000000000001,
          "end": 81.18,
          "probability": 0.99609375
        },
        {
          "word": "私",
          "start": 81.18,
          "end": 81.36000000000001,
          "probability": 0.99658203125
        },
        {
          "word": "方面",
          "start": 81.36000000000001,
          "end": 81.72000000000001,
          "probability": 0.9990234375
        },
        {
          "word": "也",
          "start": 81.72000000000001,
          "end": 82.10000000000001,
          "probability": 0.9833984375
        },
        {
          "word": "完全",
          "start": 82.10000000000001,
          "end": 82.38000000000001,
          "probability": 0.99853515625
        },
        {
          "word": "不用",
          "start": 82.38000000000001,
          "end": 82.60000000000001,
          "probability": 0.99951171875
        },
        {
          "word": "担",
          "start": 82.60000000000001,
          "end": 82.80000000000001,
          "probability": 0.99609375
        },
        {
          "word": "心",
          "start": 82.80000000000001,
          "end": 83.02000000000001,
          "probability": 1.0
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.13635246753692626,
        "compression_ratio": 1.2902097902097902,
        "no_speech_prob": 1.730680702916043e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "隐私方面也完全不用担心声音指摘你自己的设备本地识别绝对不会上传而且语音识别模型已经直接内置在安装包里装好就能用不用另外下载翻译的时候也只是把识别出来的纯文字发给你自己填的大模型借口为了让直播和视频翻得更抵挡你可以按内容类型写好背景说明和数语表甚至"
        }
      ],
      "id": 39,
      "review_annotation": "",
      "row_id": "39",
      "bounded_row_id": 39,
      "parent_row_ids": [
        39
      ],
      "parent_row_id": 39,
      "packet_id": 3,
      "LID_evidence": [
        {
          "region_id": 15,
          "probabilities": {
            "zh": 0.985863983631134
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        15
      ],
      "language_confidence": 0.985863983631134,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 80.10000000000001,
      "packet_end": 106.2,
      "raw_text": "隐私方面也完全不用担心",
      "selected_text": "隐私方面也完全不用担心",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            39
          ],
          "packet_id": 3,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        3
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 39
    },
    {
      "start": 83.4,
      "end": 85.78,
      "text": "声音指摘你自己的设备本地识别",
      "raw_asr_text": "声音指摘你自己的设备本地识别",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 3,
      "words": [
        {
          "word": "声",
          "start": 83.4,
          "end": 83.76,
          "probability": 0.98779296875
        },
        {
          "word": "音",
          "start": 83.76,
          "end": 83.94000000000001,
          "probability": 0.99951171875
        },
        {
          "word": "指",
          "start": 83.94000000000001,
          "end": 84.02000000000001,
          "probability": 0.9140625
        },
        {
          "word": "摘",
          "start": 84.02000000000001,
          "end": 84.14000000000001,
          "probability": 0.55810546875
        },
        {
          "word": "你",
          "start": 84.14000000000001,
          "end": 84.28,
          "probability": 0.99658203125
        },
        {
          "word": "自己",
          "start": 84.28,
          "end": 84.48,
          "probability": 0.99951171875
        },
        {
          "word": "的",
          "start": 84.48,
          "end": 84.60000000000001,
          "probability": 0.99951171875
        },
        {
          "word": "设",
          "start": 84.60000000000001,
          "end": 84.78,
          "probability": 0.9990234375
        },
        {
          "word": "备",
          "start": 84.78,
          "end": 84.96000000000001,
          "probability": 1.0
        },
        {
          "word": "本",
          "start": 84.96000000000001,
          "end": 85.28,
          "probability": 0.9873046875
        },
        {
          "word": "地",
          "start": 85.28,
          "end": 85.46000000000001,
          "probability": 0.998046875
        },
        {
          "word": "识",
          "start": 85.46000000000001,
          "end": 85.62,
          "probability": 0.9951171875
        },
        {
          "word": "别",
          "start": 85.62,
          "end": 85.78,
          "probability": 0.99560546875
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.13635246753692626,
        "compression_ratio": 1.2902097902097902,
        "no_speech_prob": 1.730680702916043e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "隐私方面也完全不用担心声音指摘你自己的设备本地识别绝对不会上传而且语音识别模型已经直接内置在安装包里装好就能用不用另外下载翻译的时候也只是把识别出来的纯文字发给你自己填的大模型借口为了让直播和视频翻得更抵挡你可以按内容类型写好背景说明和数语表甚至"
        }
      ],
      "id": 40,
      "review_annotation": "",
      "row_id": "40",
      "bounded_row_id": 40,
      "parent_row_ids": [
        40
      ],
      "parent_row_id": 40,
      "packet_id": 3,
      "LID_evidence": [
        {
          "region_id": 16,
          "probabilities": {
            "zh": 0.9969189708883112
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        16
      ],
      "language_confidence": 0.9969189708883112,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 80.10000000000001,
      "packet_end": 106.2,
      "raw_text": "声音指摘你自己的设备本地识别",
      "selected_text": "声音指摘你自己的设备本地识别",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            40
          ],
          "packet_id": 3,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        3
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 40
    },
    {
      "start": 85.78,
      "end": 86.96000000000001,
      "text": "绝对不会上传",
      "raw_asr_text": "绝对不会上传",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 3,
      "words": [
        {
          "word": "绝",
          "start": 85.78,
          "end": 86.28,
          "probability": 0.98388671875
        },
        {
          "word": "对",
          "start": 86.28,
          "end": 86.4,
          "probability": 0.99951171875
        },
        {
          "word": "不会",
          "start": 86.4,
          "end": 86.58000000000001,
          "probability": 0.99951171875
        },
        {
          "word": "上",
          "start": 86.58000000000001,
          "end": 86.74000000000001,
          "probability": 0.99951171875
        },
        {
          "word": "传",
          "start": 86.74000000000001,
          "end": 86.96000000000001,
          "probability": 0.99951171875
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.13635246753692626,
        "compression_ratio": 1.2902097902097902,
        "no_speech_prob": 1.730680702916043e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "隐私方面也完全不用担心声音指摘你自己的设备本地识别绝对不会上传而且语音识别模型已经直接内置在安装包里装好就能用不用另外下载翻译的时候也只是把识别出来的纯文字发给你自己填的大模型借口为了让直播和视频翻得更抵挡你可以按内容类型写好背景说明和数语表甚至"
        }
      ],
      "id": 41,
      "review_annotation": "",
      "row_id": "41",
      "bounded_row_id": 41,
      "parent_row_ids": [
        41
      ],
      "parent_row_id": 41,
      "packet_id": 3,
      "LID_evidence": [
        {
          "region_id": 16,
          "probabilities": {
            "zh": 0.9969189708883112
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        16
      ],
      "language_confidence": 0.9969189708883112,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 80.10000000000001,
      "packet_end": 106.2,
      "raw_text": "绝对不会上传",
      "selected_text": "绝对不会上传",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            41
          ],
          "packet_id": 3,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        3
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 41
    },
    {
      "start": 86.96000000000001,
      "end": 88.48,
      "text": "而且语音识别模型",
      "raw_asr_text": "而且语音识别模型",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 3,
      "words": [
        {
          "word": "而且",
          "start": 86.96000000000001,
          "end": 87.50000000000001,
          "probability": 0.98974609375
        },
        {
          "word": "语",
          "start": 87.50000000000001,
          "end": 87.7,
          "probability": 0.9990234375
        },
        {
          "word": "音",
          "start": 87.7,
          "end": 87.84,
          "probability": 0.9990234375
        },
        {
          "word": "识",
          "start": 87.84,
          "end": 87.98,
          "probability": 0.998046875
        },
        {
          "word": "别",
          "start": 87.98,
          "end": 88.10000000000001,
          "probability": 0.99365234375
        },
        {
          "word": "模",
          "start": 88.10000000000001,
          "end": 88.26,
          "probability": 0.99951171875
        },
        {
          "word": "型",
          "start": 88.26,
          "end": 88.48,
          "probability": 0.998046875
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.13635246753692626,
        "compression_ratio": 1.2902097902097902,
        "no_speech_prob": 1.730680702916043e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "隐私方面也完全不用担心声音指摘你自己的设备本地识别绝对不会上传而且语音识别模型已经直接内置在安装包里装好就能用不用另外下载翻译的时候也只是把识别出来的纯文字发给你自己填的大模型借口为了让直播和视频翻得更抵挡你可以按内容类型写好背景说明和数语表甚至"
        }
      ],
      "id": 42,
      "review_annotation": "",
      "row_id": "42",
      "bounded_row_id": 42,
      "parent_row_ids": [
        42
      ],
      "parent_row_id": 42,
      "packet_id": 3,
      "LID_evidence": [
        {
          "region_id": 16,
          "probabilities": {
            "zh": 0.9969189708883112
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        16
      ],
      "language_confidence": 0.9969189708883112,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 80.10000000000001,
      "packet_end": 106.2,
      "raw_text": "而且语音识别模型",
      "selected_text": "而且语音识别模型",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            42
          ],
          "packet_id": 3,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        3
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 42
    },
    {
      "start": 88.48,
      "end": 90.16000000000001,
      "text": "已经直接内置在安装包里",
      "raw_asr_text": "已经直接内置在安装包里",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 3,
      "words": [
        {
          "word": "已经",
          "start": 88.48,
          "end": 88.84,
          "probability": 0.99658203125
        },
        {
          "word": "直接",
          "start": 88.84,
          "end": 89.16000000000001,
          "probability": 0.99853515625
        },
        {
          "word": "内",
          "start": 89.16000000000001,
          "end": 89.30000000000001,
          "probability": 0.91162109375
        },
        {
          "word": "置",
          "start": 89.30000000000001,
          "end": 89.42000000000002,
          "probability": 0.9853515625
        },
        {
          "word": "在",
          "start": 89.42000000000002,
          "end": 89.54,
          "probability": 0.9970703125
        },
        {
          "word": "安",
          "start": 89.54,
          "end": 89.64000000000001,
          "probability": 0.9970703125
        },
        {
          "word": "装",
          "start": 89.64000000000001,
          "end": 89.82000000000001,
          "probability": 1.0
        },
        {
          "word": "包",
          "start": 89.82000000000001,
          "end": 89.98,
          "probability": 0.99755859375
        },
        {
          "word": "里",
          "start": 89.98,
          "end": 90.16000000000001,
          "probability": 0.99951171875
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.13635246753692626,
        "compression_ratio": 1.2902097902097902,
        "no_speech_prob": 1.730680702916043e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "隐私方面也完全不用担心声音指摘你自己的设备本地识别绝对不会上传而且语音识别模型已经直接内置在安装包里装好就能用不用另外下载翻译的时候也只是把识别出来的纯文字发给你自己填的大模型借口为了让直播和视频翻得更抵挡你可以按内容类型写好背景说明和数语表甚至"
        }
      ],
      "id": 43,
      "review_annotation": "",
      "row_id": "43",
      "bounded_row_id": 43,
      "parent_row_ids": [
        43
      ],
      "parent_row_id": 43,
      "packet_id": 3,
      "LID_evidence": [
        {
          "region_id": 16,
          "probabilities": {
            "zh": 0.9969189708883112
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        16
      ],
      "language_confidence": 0.9969189708883112,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 80.10000000000001,
      "packet_end": 106.2,
      "raw_text": "已经直接内置在安装包里",
      "selected_text": "已经直接内置在安装包里",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            43
          ],
          "packet_id": 3,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        3
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 43
    },
    {
      "start": 90.16000000000001,
      "end": 91.2,
      "text": "装好就能用",
      "raw_asr_text": "装好就能用",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 3,
      "words": [
        {
          "word": "装",
          "start": 90.16000000000001,
          "end": 90.58000000000001,
          "probability": 0.99169921875
        },
        {
          "word": "好",
          "start": 90.58000000000001,
          "end": 90.72000000000001,
          "probability": 0.99755859375
        },
        {
          "word": "就",
          "start": 90.72000000000001,
          "end": 90.86000000000001,
          "probability": 1.0
        },
        {
          "word": "能",
          "start": 90.86000000000001,
          "end": 91.02000000000001,
          "probability": 1.0
        },
        {
          "word": "用",
          "start": 91.02000000000001,
          "end": 91.2,
          "probability": 1.0
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.13635246753692626,
        "compression_ratio": 1.2902097902097902,
        "no_speech_prob": 1.730680702916043e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "隐私方面也完全不用担心声音指摘你自己的设备本地识别绝对不会上传而且语音识别模型已经直接内置在安装包里装好就能用不用另外下载翻译的时候也只是把识别出来的纯文字发给你自己填的大模型借口为了让直播和视频翻得更抵挡你可以按内容类型写好背景说明和数语表甚至"
        }
      ],
      "id": 44,
      "review_annotation": "",
      "row_id": "44",
      "bounded_row_id": 44,
      "parent_row_ids": [
        44
      ],
      "parent_row_id": 44,
      "packet_id": 3,
      "LID_evidence": [
        {
          "region_id": 16,
          "probabilities": {
            "zh": 0.9969189708883112
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        16
      ],
      "language_confidence": 0.9969189708883112,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 80.10000000000001,
      "packet_end": 106.2,
      "raw_text": "装好就能用",
      "selected_text": "装好就能用",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            44
          ],
          "packet_id": 3,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        3
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 44
    },
    {
      "start": 91.2,
      "end": 92.52000000000001,
      "text": "不用另外下载",
      "raw_asr_text": "不用另外下载",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 3,
      "words": [
        {
          "word": "不用",
          "start": 91.2,
          "end": 91.84,
          "probability": 0.99072265625
        },
        {
          "word": "另外",
          "start": 91.84,
          "end": 91.98,
          "probability": 0.89453125
        },
        {
          "word": "下",
          "start": 91.98,
          "end": 92.28,
          "probability": 0.99462890625
        },
        {
          "word": "载",
          "start": 92.28,
          "end": 92.52000000000001,
          "probability": 0.998046875
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.13635246753692626,
        "compression_ratio": 1.2902097902097902,
        "no_speech_prob": 1.730680702916043e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "隐私方面也完全不用担心声音指摘你自己的设备本地识别绝对不会上传而且语音识别模型已经直接内置在安装包里装好就能用不用另外下载翻译的时候也只是把识别出来的纯文字发给你自己填的大模型借口为了让直播和视频翻得更抵挡你可以按内容类型写好背景说明和数语表甚至"
        }
      ],
      "id": 45,
      "review_annotation": "",
      "row_id": "45",
      "bounded_row_id": 45,
      "parent_row_ids": [
        45
      ],
      "parent_row_id": 45,
      "packet_id": 3,
      "LID_evidence": [
        {
          "region_id": 16,
          "probabilities": {
            "zh": 0.9969189708883112
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        16
      ],
      "language_confidence": 0.9969189708883112,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 80.10000000000001,
      "packet_end": 106.2,
      "raw_text": "不用另外下载",
      "selected_text": "不用另外下载",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            45
          ],
          "packet_id": 3,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        3
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 45
    },
    {
      "start": 92.52000000000001,
      "end": 93.60000000000001,
      "text": "翻译的时候",
      "raw_asr_text": "翻译的时候",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 3,
      "words": [
        {
          "word": "翻",
          "start": 92.52000000000001,
          "end": 93.10000000000001,
          "probability": 0.998046875
        },
        {
          "word": "译",
          "start": 93.10000000000001,
          "end": 93.24000000000001,
          "probability": 0.9990234375
        },
        {
          "word": "的时候",
          "start": 93.24000000000001,
          "end": 93.60000000000001,
          "probability": 0.99951171875
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.13635246753692626,
        "compression_ratio": 1.2902097902097902,
        "no_speech_prob": 1.730680702916043e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "隐私方面也完全不用担心声音指摘你自己的设备本地识别绝对不会上传而且语音识别模型已经直接内置在安装包里装好就能用不用另外下载翻译的时候也只是把识别出来的纯文字发给你自己填的大模型借口为了让直播和视频翻得更抵挡你可以按内容类型写好背景说明和数语表甚至"
        }
      ],
      "id": 46,
      "review_annotation": "",
      "row_id": "46",
      "bounded_row_id": 46,
      "parent_row_ids": [
        46
      ],
      "parent_row_id": 46,
      "packet_id": 3,
      "LID_evidence": [
        {
          "region_id": 16,
          "probabilities": {
            "zh": 0.9969189708883112
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        16
      ],
      "language_confidence": 0.9969189708883112,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 80.10000000000001,
      "packet_end": 106.2,
      "raw_text": "翻译的时候",
      "selected_text": "翻译的时候",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            46
          ],
          "packet_id": 3,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        3
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 46
    },
    {
      "start": 93.60000000000001,
      "end": 95.50000000000001,
      "text": "也只是把识别出来的纯文字",
      "raw_asr_text": "也只是把识别出来的纯文字",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 3,
      "words": [
        {
          "word": "也",
          "start": 93.60000000000001,
          "end": 94.04,
          "probability": 0.9990234375
        },
        {
          "word": "只是",
          "start": 94.04,
          "end": 94.22000000000001,
          "probability": 0.99951171875
        },
        {
          "word": "把",
          "start": 94.22000000000001,
          "end": 94.36000000000001,
          "probability": 0.99951171875
        },
        {
          "word": "识",
          "start": 94.36000000000001,
          "end": 94.54,
          "probability": 1.0
        },
        {
          "word": "别",
          "start": 94.54,
          "end": 94.66000000000001,
          "probability": 0.9990234375
        },
        {
          "word": "出来",
          "start": 94.66000000000001,
          "end": 94.86000000000001,
          "probability": 0.9990234375
        },
        {
          "word": "的",
          "start": 94.86000000000001,
          "end": 95.00000000000001,
          "probability": 1.0
        },
        {
          "word": "纯",
          "start": 95.00000000000001,
          "end": 95.14000000000001,
          "probability": 0.9892578125
        },
        {
          "word": "文",
          "start": 95.14000000000001,
          "end": 95.30000000000001,
          "probability": 0.99267578125
        },
        {
          "word": "字",
          "start": 95.30000000000001,
          "end": 95.50000000000001,
          "probability": 0.99951171875
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.13635246753692626,
        "compression_ratio": 1.2902097902097902,
        "no_speech_prob": 1.730680702916043e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "隐私方面也完全不用担心声音指摘你自己的设备本地识别绝对不会上传而且语音识别模型已经直接内置在安装包里装好就能用不用另外下载翻译的时候也只是把识别出来的纯文字发给你自己填的大模型借口为了让直播和视频翻得更抵挡你可以按内容类型写好背景说明和数语表甚至"
        }
      ],
      "id": 47,
      "review_annotation": "",
      "row_id": "47",
      "bounded_row_id": 47,
      "parent_row_ids": [
        47
      ],
      "parent_row_id": 47,
      "packet_id": 3,
      "LID_evidence": [
        {
          "region_id": 16,
          "probabilities": {
            "zh": 0.9969189708883112
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        16
      ],
      "language_confidence": 0.9969189708883112,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 80.10000000000001,
      "packet_end": 106.2,
      "raw_text": "也只是把识别出来的纯文字",
      "selected_text": "也只是把识别出来的纯文字",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            47
          ],
          "packet_id": 3,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        3
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 47
    },
    {
      "start": 95.50000000000001,
      "end": 97.68,
      "text": "借口",
      "raw_asr_text": "发给你自己填的大模型借口",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 3,
      "words": [],
      "asr_metrics": {
        "avg_logprob": -0.13635246753692626,
        "compression_ratio": 1.2902097902097902,
        "no_speech_prob": 1.730680702916043e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "隐私方面也完全不用担心声音指摘你自己的设备本地识别绝对不会上传而且语音识别模型已经直接内置在安装包里装好就能用不用另外下载翻译的时候也只是把识别出来的纯文字发给你自己填的大模型借口为了让直播和视频翻得更抵挡你可以按内容类型写好背景说明和数语表甚至"
        }
      ],
      "id": 48,
      "review_annotation": "",
      "row_id": "48",
      "bounded_row_id": 48,
      "parent_row_ids": [
        48
      ],
      "parent_row_id": 48,
      "packet_id": 3,
      "LID_evidence": [
        {
          "region_id": 16,
          "probabilities": {
            "zh": 0.9969189708883112
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        16
      ],
      "language_confidence": 0.9969189708883112,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 80.10000000000001,
      "packet_end": 106.2,
      "raw_text": "发给你自己填的大模型借口",
      "selected_text": "借口",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            48
          ],
          "packet_id": 3,
          "engine": "mlx-whisper"
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "48",
          "before": "发给你自己填的大模型借口",
          "after": "借口",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "CERTAIN_RESTORATION",
          "reason": "大模型借口语义错误，应为接口",
          "context_evidence": "上下文为API调用，'借口'为'接口'同音误识",
          "independent_ASR": {},
          "suppression_proposed": null
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "48",
          "before": "借口",
          "after": "借口",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "CERTAIN_RESTORATION",
          "reason": "Correction: '借口' (excuse) vs '接出' (output/connect)",
          "context_evidence": "Context: '把识别出来的纯文字接出'",
          "independent_ASR": {},
          "suppression_proposed": null
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "48",
          "before": "借口",
          "after": "借口",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "CERTAIN_RESTORATION",
          "reason": "Correction: '借口' (excuse) vs '接出' (output/connect)",
          "context_evidence": "Context: '把识别出来的纯文字接出'",
          "independent_ASR": {},
          "suppression_proposed": null
        }
      ],
      "language": "zh",
      "packet_ids": [
        3
      ],
      "classification": "CERTAIN_RESTORATION",
      "raw_word_timestamps": [
        {
          "word": "发",
          "start": 95.50000000000001,
          "end": 95.9,
          "probability": 0.98876953125
        },
        {
          "word": "给",
          "start": 95.9,
          "end": 96.04,
          "probability": 1.0
        },
        {
          "word": "你",
          "start": 96.04,
          "end": 96.18,
          "probability": 1.0
        },
        {
          "word": "自己",
          "start": 96.18,
          "end": 96.4,
          "probability": 0.99951171875
        },
        {
          "word": "填",
          "start": 96.4,
          "end": 96.54,
          "probability": 0.9990234375
        },
        {
          "word": "的",
          "start": 96.54,
          "end": 96.74000000000001,
          "probability": 0.99853515625
        },
        {
          "word": "大",
          "start": 96.74000000000001,
          "end": 97.06,
          "probability": 0.97802734375
        },
        {
          "word": "模",
          "start": 97.06,
          "end": 97.20000000000002,
          "probability": 0.99755859375
        },
        {
          "word": "型",
          "start": 97.20000000000002,
          "end": 97.34,
          "probability": 0.9931640625
        },
        {
          "word": "借",
          "start": 97.34,
          "end": 97.52000000000001,
          "probability": 0.953125
        },
        {
          "word": "口",
          "start": 97.52000000000001,
          "end": 97.68,
          "probability": 0.99755859375
        }
      ],
      "word_timestamps": [],
      "source_text": "借口",
      "review_status": "CERTAIN_RESTORATION",
      "canonical_timing_evidence": "CUE_ANCHORED",
      "index": 48
    },
    {
      "start": 98.56,
      "end": 100.82000000000001,
      "text": "抵挡",
      "raw_asr_text": "为了让直播和视频翻得更抵挡",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 3,
      "words": [],
      "asr_metrics": {
        "avg_logprob": -0.13635246753692626,
        "compression_ratio": 1.2902097902097902,
        "no_speech_prob": 1.730680702916043e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "隐私方面也完全不用担心声音指摘你自己的设备本地识别绝对不会上传而且语音识别模型已经直接内置在安装包里装好就能用不用另外下载翻译的时候也只是把识别出来的纯文字发给你自己填的大模型借口为了让直播和视频翻得更抵挡你可以按内容类型写好背景说明和数语表甚至"
        }
      ],
      "id": 49,
      "review_annotation": "待確認",
      "row_id": "49",
      "bounded_row_id": 49,
      "parent_row_ids": [
        49
      ],
      "parent_row_id": 49,
      "packet_id": 3,
      "LID_evidence": [
        {
          "region_id": 17,
          "probabilities": {
            "zh": 0.989954320823445
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        17
      ],
      "language_confidence": 0.989954320823445,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 80.10000000000001,
      "packet_end": 106.2,
      "raw_text": "为了让直播和视频翻得更抵挡",
      "selected_text": "抵挡",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            49
          ],
          "packet_id": 3,
          "engine": "mlx-whisper"
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "49",
          "before": "为了让直播和视频翻得更抵挡",
          "after": "抵挡",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "翻得更抵挡语义不通，疑似同音误识",
          "context_evidence": "应为'流畅'或'地道'等描述翻译质量的词",
          "independent_ASR": {
            "engine": "whisper-large-v3",
            "model_path": "/Users/sagawa/AI_WORK/cache/huggingface/hub/models--mlx-community--whisper-large-v3",
            "language": "zh",
            "audio_interval": [
              96.56,
              102.82000000000001
            ],
            "target_interval": [
              98.56,
              100.82000000000001
            ],
            "text": "请不吝点赞订阅转发打赏",
            "words": [
              {
                "word": "请",
                "start": 98.5,
                "end": 98.9,
                "probability": 0.140625
              },
              {
                "word": "不",
                "start": 98.9,
                "end": 98.9,
                "probability": 0.98681640625
              },
              {
                "word": "吝",
                "start": 98.9,
                "end": 99.08,
                "probability": 0.99951171875
              },
              {
                "word": "点",
                "start": 99.08,
                "end": 99.18,
                "probability": 1.0
              },
              {
                "word": "赞",
                "start": 99.18,
                "end": 99.32000000000001,
                "probability": 1.0
              },
              {
                "word": " 订",
                "start": 99.32000000000001,
                "end": 99.34,
                "probability": 1.0
              },
              {
                "word": "阅",
                "start": 99.34,
                "end": 99.5,
                "probability": 1.0
              },
              {
                "word": " 转",
                "start": 99.5,
                "end": 99.7,
                "probability": 1.0
              },
              {
                "word": "发",
                "start": 99.7,
                "end": 99.7,
                "probability": 0.99853515625
              },
              {
                "word": " 打",
                "start": 100.16,
                "end": 100.4,
                "probability": 0.99951171875
              },
              {
                "word": "赏",
                "start": 100.4,
                "end": 100.76,
                "probability": 1.0
              }
            ],
            "raw_result": {
              "text": "请不吝点赞 订阅 转发 打赏支持明镜与点点栏目",
              "segments": [
                {
                  "id": 0,
                  "seek": 0,
                  "start": 1.9399999999999995,
                  "end": 5.86,
                  "text": "请不吝点赞 订阅 转发 打赏支持明镜与点点栏目",
                  "tokens": [
                    50365,
                    27908,
                    1960,
                    2392,
                    251,
                    12579,
                    5266,
                    252,
                    220,
                    7422,
                    95,
                    10034,
                    227,
                    220,
                    17819,
                    105,
                    28926,
                    220,
                    12467,
                    5266,
                    237,
                    35488,
                    11100,
                    12373,
                    250,
                    940,
                    236,
                    12579,
                    12579,
                    8987,
                    237,
                    11386,
                    51864
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.11699073454912971,
                  "compression_ratio": 0.8513513513513513,
                  "no_speech_prob": 0.8844096660614014,
                  "words": [
                    {
                      "word": "请",
                      "start": 1.9399999999999995,
                      "end": 2.34,
                      "probability": 0.140625
                    },
                    {
                      "word": "不",
                      "start": 2.34,
                      "end": 2.34,
                      "probability": 0.98681640625
                    },
                    {
                      "word": "吝",
                      "start": 2.34,
                      "end": 2.52,
                      "probability": 0.99951171875
                    },
                    {
                      "word": "点",
                      "start": 2.52,
                      "end": 2.62,
                      "probability": 1.0
                    },
                    {
                      "word": "赞",
                      "start": 2.62,
                      "end": 2.76,
                      "probability": 1.0
                    },
                    {
                      "word": " 订",
                      "start": 2.76,
                      "end": 2.78,
                      "probability": 1.0
                    },
                    {
                      "word": "阅",
                      "start": 2.78,
                      "end": 2.94,
                      "probability": 1.0
                    },
                    {
                      "word": " 转",
                      "start": 2.94,
                      "end": 3.14,
                      "probability": 1.0
                    },
                    {
                      "word": "发",
                      "start": 3.14,
                      "end": 3.14,
                      "probability": 0.99853515625
                    },
                    {
                      "word": " 打",
                      "start": 3.6,
                      "end": 3.84,
                      "probability": 0.99951171875
                    },
                    {
                      "word": "赏",
                      "start": 3.84,
                      "end": 4.2,
                      "probability": 1.0
                    },
                    {
                      "word": "支持",
                      "start": 4.2,
                      "end": 4.36,
                      "probability": 1.0
                    },
                    {
                      "word": "明",
                      "start": 4.36,
                      "end": 5.14,
                      "probability": 0.99951171875
                    },
                    {
                      "word": "镜",
                      "start": 5.14,
                      "end": 5.34,
                      "probability": 1.0
                    },
                    {
                      "word": "与",
                      "start": 5.34,
                      "end": 5.4,
                      "probability": 0.99951171875
                    },
                    {
                      "word": "点",
                      "start": 5.4,
                      "end": 5.4,
                      "probability": 0.99951171875
                    },
                    {
                      "word": "点",
                      "start": 5.4,
                      "end": 5.4,
                      "probability": 0.99951171875
                    },
                    {
                      "word": "栏",
                      "start": 5.4,
                      "end": 5.4,
                      "probability": 1.0
                    },
                    {
                      "word": "目",
                      "start": 5.4,
                      "end": 5.86,
                      "probability": 1.0
                    }
                  ]
                }
              ],
              "language": "zh",
              "duration": 6.26
            },
            "status": "COMPLETE",
            "requested_language": "zh"
          },
          "suppression_proposed": null
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "49",
          "before": "抵挡",
          "after": "抵挡",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "Hallucination: '抵挡' (resist) vs '请不吝点赞' (please like)",
          "context_evidence": "ASR: '请不吝点赞订阅转发打赏'",
          "independent_ASR": {
            "engine": "whisper-large-v3",
            "model_path": "/Users/sagawa/AI_WORK/cache/huggingface/hub/models--mlx-community--whisper-large-v3",
            "language": "zh",
            "audio_interval": [
              96.56,
              102.82000000000001
            ],
            "target_interval": [
              98.56,
              100.82000000000001
            ],
            "text": "请不吝点赞订阅转发打赏",
            "words": [
              {
                "word": "请",
                "start": 98.5,
                "end": 98.9,
                "probability": 0.140625
              },
              {
                "word": "不",
                "start": 98.9,
                "end": 98.9,
                "probability": 0.98681640625
              },
              {
                "word": "吝",
                "start": 98.9,
                "end": 99.08,
                "probability": 0.99951171875
              },
              {
                "word": "点",
                "start": 99.08,
                "end": 99.18,
                "probability": 1.0
              },
              {
                "word": "赞",
                "start": 99.18,
                "end": 99.32000000000001,
                "probability": 1.0
              },
              {
                "word": " 订",
                "start": 99.32000000000001,
                "end": 99.34,
                "probability": 1.0
              },
              {
                "word": "阅",
                "start": 99.34,
                "end": 99.5,
                "probability": 1.0
              },
              {
                "word": " 转",
                "start": 99.5,
                "end": 99.7,
                "probability": 1.0
              },
              {
                "word": "发",
                "start": 99.7,
                "end": 99.7,
                "probability": 0.99853515625
              },
              {
                "word": " 打",
                "start": 100.16,
                "end": 100.4,
                "probability": 0.99951171875
              },
              {
                "word": "赏",
                "start": 100.4,
                "end": 100.76,
                "probability": 1.0
              }
            ],
            "raw_result": {
              "text": "请不吝点赞 订阅 转发 打赏支持明镜与点点栏目",
              "segments": [
                {
                  "id": 0,
                  "seek": 0,
                  "start": 1.9399999999999995,
                  "end": 5.86,
                  "text": "请不吝点赞 订阅 转发 打赏支持明镜与点点栏目",
                  "tokens": [
                    50365,
                    27908,
                    1960,
                    2392,
                    251,
                    12579,
                    5266,
                    252,
                    220,
                    7422,
                    95,
                    10034,
                    227,
                    220,
                    17819,
                    105,
                    28926,
                    220,
                    12467,
                    5266,
                    237,
                    35488,
                    11100,
                    12373,
                    250,
                    940,
                    236,
                    12579,
                    12579,
                    8987,
                    237,
                    11386,
                    51864
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.11699073454912971,
                  "compression_ratio": 0.8513513513513513,
                  "no_speech_prob": 0.8844096660614014,
                  "words": [
                    {
                      "word": "请",
                      "start": 1.9399999999999995,
                      "end": 2.34,
                      "probability": 0.140625
                    },
                    {
                      "word": "不",
                      "start": 2.34,
                      "end": 2.34,
                      "probability": 0.98681640625
                    },
                    {
                      "word": "吝",
                      "start": 2.34,
                      "end": 2.52,
                      "probability": 0.99951171875
                    },
                    {
                      "word": "点",
                      "start": 2.52,
                      "end": 2.62,
                      "probability": 1.0
                    },
                    {
                      "word": "赞",
                      "start": 2.62,
                      "end": 2.76,
                      "probability": 1.0
                    },
                    {
                      "word": " 订",
                      "start": 2.76,
                      "end": 2.78,
                      "probability": 1.0
                    },
                    {
                      "word": "阅",
                      "start": 2.78,
                      "end": 2.94,
                      "probability": 1.0
                    },
                    {
                      "word": " 转",
                      "start": 2.94,
                      "end": 3.14,
                      "probability": 1.0
                    },
                    {
                      "word": "发",
                      "start": 3.14,
                      "end": 3.14,
                      "probability": 0.99853515625
                    },
                    {
                      "word": " 打",
                      "start": 3.6,
                      "end": 3.84,
                      "probability": 0.99951171875
                    },
                    {
                      "word": "赏",
                      "start": 3.84,
                      "end": 4.2,
                      "probability": 1.0
                    },
                    {
                      "word": "支持",
                      "start": 4.2,
                      "end": 4.36,
                      "probability": 1.0
                    },
                    {
                      "word": "明",
                      "start": 4.36,
                      "end": 5.14,
                      "probability": 0.99951171875
                    },
                    {
                      "word": "镜",
                      "start": 5.14,
                      "end": 5.34,
                      "probability": 1.0
                    },
                    {
                      "word": "与",
                      "start": 5.34,
                      "end": 5.4,
                      "probability": 0.99951171875
                    },
                    {
                      "word": "点",
                      "start": 5.4,
                      "end": 5.4,
                      "probability": 0.99951171875
                    },
                    {
                      "word": "点",
                      "start": 5.4,
                      "end": 5.4,
                      "probability": 0.99951171875
                    },
                    {
                      "word": "栏",
                      "start": 5.4,
                      "end": 5.4,
                      "probability": 1.0
                    },
                    {
                      "word": "目",
                      "start": 5.4,
                      "end": 5.86,
                      "probability": 1.0
                    }
                  ]
                }
              ],
              "language": "zh",
              "duration": 6.26
            },
            "status": "COMPLETE",
            "requested_language": "zh"
          },
          "suppression_proposed": null
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "49",
          "before": "抵挡",
          "after": "抵挡",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "Hallucination: '抵挡' (resist) vs '请不吝点赞' (please like)",
          "context_evidence": "ASR: '请不吝点赞订阅转发打赏'",
          "independent_ASR": {
            "engine": "whisper-large-v3",
            "model_path": "/Users/sagawa/AI_WORK/cache/huggingface/hub/models--mlx-community--whisper-large-v3",
            "language": "zh",
            "audio_interval": [
              96.56,
              102.82000000000001
            ],
            "target_interval": [
              98.56,
              100.82000000000001
            ],
            "text": "请不吝点赞订阅转发打赏",
            "words": [
              {
                "word": "请",
                "start": 98.5,
                "end": 98.9,
                "probability": 0.140625
              },
              {
                "word": "不",
                "start": 98.9,
                "end": 98.9,
                "probability": 0.98681640625
              },
              {
                "word": "吝",
                "start": 98.9,
                "end": 99.08,
                "probability": 0.99951171875
              },
              {
                "word": "点",
                "start": 99.08,
                "end": 99.18,
                "probability": 1.0
              },
              {
                "word": "赞",
                "start": 99.18,
                "end": 99.32000000000001,
                "probability": 1.0
              },
              {
                "word": " 订",
                "start": 99.32000000000001,
                "end": 99.34,
                "probability": 1.0
              },
              {
                "word": "阅",
                "start": 99.34,
                "end": 99.5,
                "probability": 1.0
              },
              {
                "word": " 转",
                "start": 99.5,
                "end": 99.7,
                "probability": 1.0
              },
              {
                "word": "发",
                "start": 99.7,
                "end": 99.7,
                "probability": 0.99853515625
              },
              {
                "word": " 打",
                "start": 100.16,
                "end": 100.4,
                "probability": 0.99951171875
              },
              {
                "word": "赏",
                "start": 100.4,
                "end": 100.76,
                "probability": 1.0
              }
            ],
            "raw_result": {
              "text": "请不吝点赞 订阅 转发 打赏支持明镜与点点栏目",
              "segments": [
                {
                  "id": 0,
                  "seek": 0,
                  "start": 1.9399999999999995,
                  "end": 5.86,
                  "text": "请不吝点赞 订阅 转发 打赏支持明镜与点点栏目",
                  "tokens": [
                    50365,
                    27908,
                    1960,
                    2392,
                    251,
                    12579,
                    5266,
                    252,
                    220,
                    7422,
                    95,
                    10034,
                    227,
                    220,
                    17819,
                    105,
                    28926,
                    220,
                    12467,
                    5266,
                    237,
                    35488,
                    11100,
                    12373,
                    250,
                    940,
                    236,
                    12579,
                    12579,
                    8987,
                    237,
                    11386,
                    51864
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.11699073454912971,
                  "compression_ratio": 0.8513513513513513,
                  "no_speech_prob": 0.8844096660614014,
                  "words": [
                    {
                      "word": "请",
                      "start": 1.9399999999999995,
                      "end": 2.34,
                      "probability": 0.140625
                    },
                    {
                      "word": "不",
                      "start": 2.34,
                      "end": 2.34,
                      "probability": 0.98681640625
                    },
                    {
                      "word": "吝",
                      "start": 2.34,
                      "end": 2.52,
                      "probability": 0.99951171875
                    },
                    {
                      "word": "点",
                      "start": 2.52,
                      "end": 2.62,
                      "probability": 1.0
                    },
                    {
                      "word": "赞",
                      "start": 2.62,
                      "end": 2.76,
                      "probability": 1.0
                    },
                    {
                      "word": " 订",
                      "start": 2.76,
                      "end": 2.78,
                      "probability": 1.0
                    },
                    {
                      "word": "阅",
                      "start": 2.78,
                      "end": 2.94,
                      "probability": 1.0
                    },
                    {
                      "word": " 转",
                      "start": 2.94,
                      "end": 3.14,
                      "probability": 1.0
                    },
                    {
                      "word": "发",
                      "start": 3.14,
                      "end": 3.14,
                      "probability": 0.99853515625
                    },
                    {
                      "word": " 打",
                      "start": 3.6,
                      "end": 3.84,
                      "probability": 0.99951171875
                    },
                    {
                      "word": "赏",
                      "start": 3.84,
                      "end": 4.2,
                      "probability": 1.0
                    },
                    {
                      "word": "支持",
                      "start": 4.2,
                      "end": 4.36,
                      "probability": 1.0
                    },
                    {
                      "word": "明",
                      "start": 4.36,
                      "end": 5.14,
                      "probability": 0.99951171875
                    },
                    {
                      "word": "镜",
                      "start": 5.14,
                      "end": 5.34,
                      "probability": 1.0
                    },
                    {
                      "word": "与",
                      "start": 5.34,
                      "end": 5.4,
                      "probability": 0.99951171875
                    },
                    {
                      "word": "点",
                      "start": 5.4,
                      "end": 5.4,
                      "probability": 0.99951171875
                    },
                    {
                      "word": "点",
                      "start": 5.4,
                      "end": 5.4,
                      "probability": 0.99951171875
                    },
                    {
                      "word": "栏",
                      "start": 5.4,
                      "end": 5.4,
                      "probability": 1.0
                    },
                    {
                      "word": "目",
                      "start": 5.4,
                      "end": 5.86,
                      "probability": 1.0
                    }
                  ]
                }
              ],
              "language": "zh",
              "duration": 6.26
            },
            "status": "COMPLETE",
            "requested_language": "zh"
          },
          "suppression_proposed": null
        }
      ],
      "language": "zh",
      "packet_ids": [
        3
      ],
      "classification": "BEST_SUPPORTED_WITH_REVIEW",
      "independent_ASR": {
        "engine": "whisper-large-v3",
        "model_path": "/Users/sagawa/AI_WORK/cache/huggingface/hub/models--mlx-community--whisper-large-v3",
        "language": "zh",
        "audio_interval": [
          96.56,
          102.82000000000001
        ],
        "target_interval": [
          98.56,
          100.82000000000001
        ],
        "text": "请不吝点赞订阅转发打赏",
        "words": [
          {
            "word": "请",
            "start": 98.5,
            "end": 98.9,
            "probability": 0.140625
          },
          {
            "word": "不",
            "start": 98.9,
            "end": 98.9,
            "probability": 0.98681640625
          },
          {
            "word": "吝",
            "start": 98.9,
            "end": 99.08,
            "probability": 0.99951171875
          },
          {
            "word": "点",
            "start": 99.08,
            "end": 99.18,
            "probability": 1.0
          },
          {
            "word": "赞",
            "start": 99.18,
            "end": 99.32000000000001,
            "probability": 1.0
          },
          {
            "word": " 订",
            "start": 99.32000000000001,
            "end": 99.34,
            "probability": 1.0
          },
          {
            "word": "阅",
            "start": 99.34,
            "end": 99.5,
            "probability": 1.0
          },
          {
            "word": " 转",
            "start": 99.5,
            "end": 99.7,
            "probability": 1.0
          },
          {
            "word": "发",
            "start": 99.7,
            "end": 99.7,
            "probability": 0.99853515625
          },
          {
            "word": " 打",
            "start": 100.16,
            "end": 100.4,
            "probability": 0.99951171875
          },
          {
            "word": "赏",
            "start": 100.4,
            "end": 100.76,
            "probability": 1.0
          }
        ],
        "raw_result": {
          "text": "请不吝点赞 订阅 转发 打赏支持明镜与点点栏目",
          "segments": [
            {
              "id": 0,
              "seek": 0,
              "start": 1.9399999999999995,
              "end": 5.86,
              "text": "请不吝点赞 订阅 转发 打赏支持明镜与点点栏目",
              "tokens": [
                50365,
                27908,
                1960,
                2392,
                251,
                12579,
                5266,
                252,
                220,
                7422,
                95,
                10034,
                227,
                220,
                17819,
                105,
                28926,
                220,
                12467,
                5266,
                237,
                35488,
                11100,
                12373,
                250,
                940,
                236,
                12579,
                12579,
                8987,
                237,
                11386,
                51864
              ],
              "temperature": 0.0,
              "avg_logprob": -0.11699073454912971,
              "compression_ratio": 0.8513513513513513,
              "no_speech_prob": 0.8844096660614014,
              "words": [
                {
                  "word": "请",
                  "start": 1.9399999999999995,
                  "end": 2.34,
                  "probability": 0.140625
                },
                {
                  "word": "不",
                  "start": 2.34,
                  "end": 2.34,
                  "probability": 0.98681640625
                },
                {
                  "word": "吝",
                  "start": 2.34,
                  "end": 2.52,
                  "probability": 0.99951171875
                },
                {
                  "word": "点",
                  "start": 2.52,
                  "end": 2.62,
                  "probability": 1.0
                },
                {
                  "word": "赞",
                  "start": 2.62,
                  "end": 2.76,
                  "probability": 1.0
                },
                {
                  "word": " 订",
                  "start": 2.76,
                  "end": 2.78,
                  "probability": 1.0
                },
                {
                  "word": "阅",
                  "start": 2.78,
                  "end": 2.94,
                  "probability": 1.0
                },
                {
                  "word": " 转",
                  "start": 2.94,
                  "end": 3.14,
                  "probability": 1.0
                },
                {
                  "word": "发",
                  "start": 3.14,
                  "end": 3.14,
                  "probability": 0.99853515625
                },
                {
                  "word": " 打",
                  "start": 3.6,
                  "end": 3.84,
                  "probability": 0.99951171875
                },
                {
                  "word": "赏",
                  "start": 3.84,
                  "end": 4.2,
                  "probability": 1.0
                },
                {
                  "word": "支持",
                  "start": 4.2,
                  "end": 4.36,
                  "probability": 1.0
                },
                {
                  "word": "明",
                  "start": 4.36,
                  "end": 5.14,
                  "probability": 0.99951171875
                },
                {
                  "word": "镜",
                  "start": 5.14,
                  "end": 5.34,
                  "probability": 1.0
                },
                {
                  "word": "与",
                  "start": 5.34,
                  "end": 5.4,
                  "probability": 0.99951171875
                },
                {
                  "word": "点",
                  "start": 5.4,
                  "end": 5.4,
                  "probability": 0.99951171875
                },
                {
                  "word": "点",
                  "start": 5.4,
                  "end": 5.4,
                  "probability": 0.99951171875
                },
                {
                  "word": "栏",
                  "start": 5.4,
                  "end": 5.4,
                  "probability": 1.0
                },
                {
                  "word": "目",
                  "start": 5.4,
                  "end": 5.86,
                  "probability": 1.0
                }
              ]
            }
          ],
          "language": "zh",
          "duration": 6.26
        },
        "status": "COMPLETE",
        "requested_language": "zh"
      },
      "raw_word_timestamps": [
        {
          "word": "为",
          "start": 98.56,
          "end": 98.92000000000002,
          "probability": 0.99853515625
        },
        {
          "word": "了",
          "start": 98.92000000000002,
          "end": 99.04,
          "probability": 0.99951171875
        },
        {
          "word": "让",
          "start": 99.04,
          "end": 99.18,
          "probability": 0.99951171875
        },
        {
          "word": "直",
          "start": 99.18,
          "end": 99.36000000000001,
          "probability": 0.99951171875
        },
        {
          "word": "播",
          "start": 99.36000000000001,
          "end": 99.52000000000001,
          "probability": 1.0
        },
        {
          "word": "和",
          "start": 99.52000000000001,
          "end": 99.68,
          "probability": 0.9990234375
        },
        {
          "word": "视",
          "start": 99.68,
          "end": 99.84,
          "probability": 0.99951171875
        },
        {
          "word": "频",
          "start": 99.84,
          "end": 100.0,
          "probability": 1.0
        },
        {
          "word": "翻",
          "start": 100.0,
          "end": 100.16000000000001,
          "probability": 0.7783203125
        },
        {
          "word": "得",
          "start": 100.16000000000001,
          "end": 100.26,
          "probability": 0.58837890625
        },
        {
          "word": "更",
          "start": 100.26,
          "end": 100.44000000000001,
          "probability": 0.99951171875
        },
        {
          "word": "抵",
          "start": 100.44000000000001,
          "end": 100.64000000000001,
          "probability": 0.6572265625
        },
        {
          "word": "挡",
          "start": 100.64000000000001,
          "end": 100.82000000000001,
          "probability": 0.955078125
        }
      ],
      "word_timestamps": [],
      "source_text": "抵挡",
      "review_status": "BEST_SUPPORTED_WITH_REVIEW",
      "canonical_timing_evidence": "CUE_ANCHORED",
      "index": 49
    },
    {
      "start": 100.82000000000001,
      "end": 102.28,
      "text": "你可以按内容类型",
      "raw_asr_text": "你可以按内容类型",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 3,
      "words": [
        {
          "word": "你可以",
          "start": 100.82000000000001,
          "end": 101.36000000000001,
          "probability": 0.99609375
        },
        {
          "word": "按",
          "start": 101.36000000000001,
          "end": 101.62,
          "probability": 0.9921875
        },
        {
          "word": "内",
          "start": 101.62,
          "end": 101.78,
          "probability": 0.97607421875
        },
        {
          "word": "容",
          "start": 101.78,
          "end": 101.92000000000002,
          "probability": 1.0
        },
        {
          "word": "类",
          "start": 101.92000000000002,
          "end": 102.10000000000001,
          "probability": 0.96337890625
        },
        {
          "word": "型",
          "start": 102.10000000000001,
          "end": 102.28,
          "probability": 0.99853515625
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.13635246753692626,
        "compression_ratio": 1.2902097902097902,
        "no_speech_prob": 1.730680702916043e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "隐私方面也完全不用担心声音指摘你自己的设备本地识别绝对不会上传而且语音识别模型已经直接内置在安装包里装好就能用不用另外下载翻译的时候也只是把识别出来的纯文字发给你自己填的大模型借口为了让直播和视频翻得更抵挡你可以按内容类型写好背景说明和数语表甚至"
        }
      ],
      "id": 50,
      "review_annotation": "",
      "row_id": "50",
      "bounded_row_id": 50,
      "parent_row_ids": [
        50
      ],
      "parent_row_id": 50,
      "packet_id": 3,
      "LID_evidence": [
        {
          "region_id": 17,
          "probabilities": {
            "zh": 0.989954320823445
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        17
      ],
      "language_confidence": 0.989954320823445,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 80.10000000000001,
      "packet_end": 106.2,
      "raw_text": "你可以按内容类型",
      "selected_text": "你可以按内容类型",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            50
          ],
          "packet_id": 3,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        3
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 50
    },
    {
      "start": 102.28,
      "end": 103.4,
      "text": "写好背景",
      "raw_asr_text": "写好背景",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 3,
      "words": [
        {
          "word": "写",
          "start": 102.28,
          "end": 102.88000000000001,
          "probability": 0.9990234375
        },
        {
          "word": "好",
          "start": 102.88000000000001,
          "end": 103.02000000000001,
          "probability": 1.0
        },
        {
          "word": "背",
          "start": 103.02000000000001,
          "end": 103.18,
          "probability": 0.99951171875
        },
        {
          "word": "景",
          "start": 103.18,
          "end": 103.4,
          "probability": 0.99951171875
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.13635246753692626,
        "compression_ratio": 1.2902097902097902,
        "no_speech_prob": 1.730680702916043e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "隐私方面也完全不用担心声音指摘你自己的设备本地识别绝对不会上传而且语音识别模型已经直接内置在安装包里装好就能用不用另外下载翻译的时候也只是把识别出来的纯文字发给你自己填的大模型借口为了让直播和视频翻得更抵挡你可以按内容类型写好背景说明和数语表甚至"
        }
      ],
      "id": 51,
      "review_annotation": "",
      "row_id": "51",
      "bounded_row_id": 51,
      "parent_row_ids": [
        51
      ],
      "parent_row_id": 51,
      "packet_id": 3,
      "LID_evidence": [
        {
          "region_id": 17,
          "probabilities": {
            "zh": 0.989954320823445
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        17
      ],
      "language_confidence": 0.989954320823445,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 80.10000000000001,
      "packet_end": 106.2,
      "raw_text": "写好背景",
      "selected_text": "写好背景",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            51
          ],
          "packet_id": 3,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        3
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 51
    },
    {
      "start": 103.4,
      "end": 105.10000000000001,
      "text": "数语表",
      "raw_asr_text": "说明和数语表",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 3,
      "words": [],
      "asr_metrics": {
        "avg_logprob": -0.13635246753692626,
        "compression_ratio": 1.2902097902097902,
        "no_speech_prob": 1.730680702916043e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "隐私方面也完全不用担心声音指摘你自己的设备本地识别绝对不会上传而且语音识别模型已经直接内置在安装包里装好就能用不用另外下载翻译的时候也只是把识别出来的纯文字发给你自己填的大模型借口为了让直播和视频翻得更抵挡你可以按内容类型写好背景说明和数语表甚至"
        }
      ],
      "id": 52,
      "review_annotation": "待確認",
      "row_id": "52",
      "bounded_row_id": 52,
      "parent_row_ids": [
        52
      ],
      "parent_row_id": 52,
      "packet_id": 3,
      "LID_evidence": [
        {
          "region_id": 17,
          "probabilities": {
            "zh": 0.989954320823445
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        17
      ],
      "language_confidence": 0.989954320823445,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 80.10000000000001,
      "packet_end": 106.2,
      "raw_text": "说明和数语表",
      "selected_text": "数语表",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            52
          ],
          "packet_id": 3,
          "engine": "mlx-whisper"
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "52",
          "before": "说明和数语表",
          "after": "数语表",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "数语表疑似术语表误识",
          "context_evidence": "上下文为翻译配置，'术语表'更符合逻辑",
          "independent_ASR": {
            "engine": "whisper-large-v3",
            "model_path": "/Users/sagawa/AI_WORK/cache/huggingface/hub/models--mlx-community--whisper-large-v3",
            "language": "zh",
            "audio_interval": [
              101.4,
              107.10000000000001
            ],
            "target_interval": [
              103.4,
              105.10000000000001
            ],
            "text": "说明和数据表,",
            "words": [
              {
                "word": "说",
                "start": 103.42,
                "end": 103.86,
                "probability": 0.99951171875
              },
              {
                "word": "明",
                "start": 103.86,
                "end": 104.18,
                "probability": 0.998046875
              },
              {
                "word": "和",
                "start": 104.18,
                "end": 104.5,
                "probability": 0.986328125
              },
              {
                "word": "数",
                "start": 104.5,
                "end": 104.68,
                "probability": 0.47998046875
              },
              {
                "word": "据",
                "start": 104.68,
                "end": 104.82000000000001,
                "probability": 0.7392578125
              },
              {
                "word": "表,",
                "start": 104.82000000000001,
                "end": 105.0,
                "probability": 1.0
              }
            ],
            "raw_result": {
              "text": "可以按内容类型,写好背景,说明和数据表,甚至直接点一下。",
              "segments": [
                {
                  "id": 0,
                  "seek": 0,
                  "start": 0.0,
                  "end": 5.58,
                  "text": "可以按内容类型,写好背景,说明和数据表,甚至直接点一下。",
                  "tokens": [
                    50365,
                    6723,
                    26613,
                    34742,
                    25750,
                    22113,
                    119,
                    39823,
                    11,
                    5676,
                    247,
                    2131,
                    46329,
                    50218,
                    11,
                    8090,
                    11100,
                    12565,
                    33188,
                    26075,
                    106,
                    17571,
                    11,
                    17556,
                    20844,
                    43297,
                    12579,
                    8861,
                    1543,
                    50650
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.1863205663619503,
                  "compression_ratio": 0.8764044943820225,
                  "no_speech_prob": 0.8773801326751709,
                  "words": [
                    {
                      "word": "可以",
                      "start": 0.0,
                      "end": 0.1,
                      "probability": 0.444580078125
                    },
                    {
                      "word": "按",
                      "start": 0.1,
                      "end": 0.16,
                      "probability": 0.99560546875
                    },
                    {
                      "word": "内",
                      "start": 0.16,
                      "end": 0.32,
                      "probability": 0.9375
                    },
                    {
                      "word": "容",
                      "start": 0.32,
                      "end": 0.48,
                      "probability": 1.0
                    },
                    {
                      "word": "类",
                      "start": 0.48,
                      "end": 0.66,
                      "probability": 0.9990234375
                    },
                    {
                      "word": "型,",
                      "start": 0.66,
                      "end": 0.82,
                      "probability": 0.99853515625
                    },
                    {
                      "word": "写",
                      "start": 0.9,
                      "end": 1.48,
                      "probability": 0.9990234375
                    },
                    {
                      "word": "好",
                      "start": 1.48,
                      "end": 1.56,
                      "probability": 1.0
                    },
                    {
                      "word": "背",
                      "start": 1.56,
                      "end": 1.68,
                      "probability": 0.99951171875
                    },
                    {
                      "word": "景,",
                      "start": 1.68,
                      "end": 1.96,
                      "probability": 1.0
                    },
                    {
                      "word": "说",
                      "start": 2.02,
                      "end": 2.46,
                      "probability": 0.99951171875
                    },
                    {
                      "word": "明",
                      "start": 2.46,
                      "end": 2.78,
                      "probability": 0.998046875
                    },
                    {
                      "word": "和",
                      "start": 2.78,
                      "end": 3.1,
                      "probability": 0.986328125
                    },
                    {
                      "word": "数",
                      "start": 3.1,
                      "end": 3.28,
                      "probability": 0.47998046875
                    },
                    {
                      "word": "据",
                      "start": 3.28,
                      "end": 3.42,
                      "probability": 0.7392578125
                    },
                    {
                      "word": "表,",
                      "start": 3.42,
                      "end": 3.6,
                      "probability": 1.0
                    },
                    {
                      "word": "甚",
                      "start": 3.72,
                      "end": 4.28,
                      "probability": 0.99951171875
                    },
                    {
                      "word": "至",
                      "start": 4.28,
                      "end": 4.54,
                      "probability": 1.0
                    },
                    {
                      "word": "直接",
                      "start": 4.54,
                      "end": 5.1,
                      "probability": 0.978515625
                    },
                    {
                      "word": "点",
                      "start": 5.1,
                      "end": 5.4,
                      "probability": 0.99951171875
                    },
                    {
                      "word": "一下。",
                      "start": 5.4,
                      "end": 5.58,
                      "probability": 0.84423828125
                    }
                  ]
                }
              ],
              "language": "zh",
              "duration": 5.7
            },
            "status": "COMPLETE",
            "requested_language": "zh"
          },
          "suppression_proposed": null
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "52",
          "before": "数语表",
          "after": "数语表",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "Nonsense: '数语表' vs '术语表' (terminology list)",
          "context_evidence": "Context: '写好背景术语表'",
          "independent_ASR": {
            "engine": "whisper-large-v3",
            "model_path": "/Users/sagawa/AI_WORK/cache/huggingface/hub/models--mlx-community--whisper-large-v3",
            "language": "zh",
            "audio_interval": [
              101.4,
              107.10000000000001
            ],
            "target_interval": [
              103.4,
              105.10000000000001
            ],
            "text": "说明和数据表,",
            "words": [
              {
                "word": "说",
                "start": 103.42,
                "end": 103.86,
                "probability": 0.99951171875
              },
              {
                "word": "明",
                "start": 103.86,
                "end": 104.18,
                "probability": 0.998046875
              },
              {
                "word": "和",
                "start": 104.18,
                "end": 104.5,
                "probability": 0.986328125
              },
              {
                "word": "数",
                "start": 104.5,
                "end": 104.68,
                "probability": 0.47998046875
              },
              {
                "word": "据",
                "start": 104.68,
                "end": 104.82000000000001,
                "probability": 0.7392578125
              },
              {
                "word": "表,",
                "start": 104.82000000000001,
                "end": 105.0,
                "probability": 1.0
              }
            ],
            "raw_result": {
              "text": "可以按内容类型,写好背景,说明和数据表,甚至直接点一下。",
              "segments": [
                {
                  "id": 0,
                  "seek": 0,
                  "start": 0.0,
                  "end": 5.58,
                  "text": "可以按内容类型,写好背景,说明和数据表,甚至直接点一下。",
                  "tokens": [
                    50365,
                    6723,
                    26613,
                    34742,
                    25750,
                    22113,
                    119,
                    39823,
                    11,
                    5676,
                    247,
                    2131,
                    46329,
                    50218,
                    11,
                    8090,
                    11100,
                    12565,
                    33188,
                    26075,
                    106,
                    17571,
                    11,
                    17556,
                    20844,
                    43297,
                    12579,
                    8861,
                    1543,
                    50650
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.1863205663619503,
                  "compression_ratio": 0.8764044943820225,
                  "no_speech_prob": 0.8773801326751709,
                  "words": [
                    {
                      "word": "可以",
                      "start": 0.0,
                      "end": 0.1,
                      "probability": 0.444580078125
                    },
                    {
                      "word": "按",
                      "start": 0.1,
                      "end": 0.16,
                      "probability": 0.99560546875
                    },
                    {
                      "word": "内",
                      "start": 0.16,
                      "end": 0.32,
                      "probability": 0.9375
                    },
                    {
                      "word": "容",
                      "start": 0.32,
                      "end": 0.48,
                      "probability": 1.0
                    },
                    {
                      "word": "类",
                      "start": 0.48,
                      "end": 0.66,
                      "probability": 0.9990234375
                    },
                    {
                      "word": "型,",
                      "start": 0.66,
                      "end": 0.82,
                      "probability": 0.99853515625
                    },
                    {
                      "word": "写",
                      "start": 0.9,
                      "end": 1.48,
                      "probability": 0.9990234375
                    },
                    {
                      "word": "好",
                      "start": 1.48,
                      "end": 1.56,
                      "probability": 1.0
                    },
                    {
                      "word": "背",
                      "start": 1.56,
                      "end": 1.68,
                      "probability": 0.99951171875
                    },
                    {
                      "word": "景,",
                      "start": 1.68,
                      "end": 1.96,
                      "probability": 1.0
                    },
                    {
                      "word": "说",
                      "start": 2.02,
                      "end": 2.46,
                      "probability": 0.99951171875
                    },
                    {
                      "word": "明",
                      "start": 2.46,
                      "end": 2.78,
                      "probability": 0.998046875
                    },
                    {
                      "word": "和",
                      "start": 2.78,
                      "end": 3.1,
                      "probability": 0.986328125
                    },
                    {
                      "word": "数",
                      "start": 3.1,
                      "end": 3.28,
                      "probability": 0.47998046875
                    },
                    {
                      "word": "据",
                      "start": 3.28,
                      "end": 3.42,
                      "probability": 0.7392578125
                    },
                    {
                      "word": "表,",
                      "start": 3.42,
                      "end": 3.6,
                      "probability": 1.0
                    },
                    {
                      "word": "甚",
                      "start": 3.72,
                      "end": 4.28,
                      "probability": 0.99951171875
                    },
                    {
                      "word": "至",
                      "start": 4.28,
                      "end": 4.54,
                      "probability": 1.0
                    },
                    {
                      "word": "直接",
                      "start": 4.54,
                      "end": 5.1,
                      "probability": 0.978515625
                    },
                    {
                      "word": "点",
                      "start": 5.1,
                      "end": 5.4,
                      "probability": 0.99951171875
                    },
                    {
                      "word": "一下。",
                      "start": 5.4,
                      "end": 5.58,
                      "probability": 0.84423828125
                    }
                  ]
                }
              ],
              "language": "zh",
              "duration": 5.7
            },
            "status": "COMPLETE",
            "requested_language": "zh"
          },
          "suppression_proposed": null
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "52",
          "before": "数语表",
          "after": "数语表",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "Nonsense: '数语表' vs '术语表' (terminology list)",
          "context_evidence": "Context: '写好背景术语表'",
          "independent_ASR": {
            "engine": "whisper-large-v3",
            "model_path": "/Users/sagawa/AI_WORK/cache/huggingface/hub/models--mlx-community--whisper-large-v3",
            "language": "zh",
            "audio_interval": [
              101.4,
              107.10000000000001
            ],
            "target_interval": [
              103.4,
              105.10000000000001
            ],
            "text": "说明和数据表,",
            "words": [
              {
                "word": "说",
                "start": 103.42,
                "end": 103.86,
                "probability": 0.99951171875
              },
              {
                "word": "明",
                "start": 103.86,
                "end": 104.18,
                "probability": 0.998046875
              },
              {
                "word": "和",
                "start": 104.18,
                "end": 104.5,
                "probability": 0.986328125
              },
              {
                "word": "数",
                "start": 104.5,
                "end": 104.68,
                "probability": 0.47998046875
              },
              {
                "word": "据",
                "start": 104.68,
                "end": 104.82000000000001,
                "probability": 0.7392578125
              },
              {
                "word": "表,",
                "start": 104.82000000000001,
                "end": 105.0,
                "probability": 1.0
              }
            ],
            "raw_result": {
              "text": "可以按内容类型,写好背景,说明和数据表,甚至直接点一下。",
              "segments": [
                {
                  "id": 0,
                  "seek": 0,
                  "start": 0.0,
                  "end": 5.58,
                  "text": "可以按内容类型,写好背景,说明和数据表,甚至直接点一下。",
                  "tokens": [
                    50365,
                    6723,
                    26613,
                    34742,
                    25750,
                    22113,
                    119,
                    39823,
                    11,
                    5676,
                    247,
                    2131,
                    46329,
                    50218,
                    11,
                    8090,
                    11100,
                    12565,
                    33188,
                    26075,
                    106,
                    17571,
                    11,
                    17556,
                    20844,
                    43297,
                    12579,
                    8861,
                    1543,
                    50650
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.1863205663619503,
                  "compression_ratio": 0.8764044943820225,
                  "no_speech_prob": 0.8773801326751709,
                  "words": [
                    {
                      "word": "可以",
                      "start": 0.0,
                      "end": 0.1,
                      "probability": 0.444580078125
                    },
                    {
                      "word": "按",
                      "start": 0.1,
                      "end": 0.16,
                      "probability": 0.99560546875
                    },
                    {
                      "word": "内",
                      "start": 0.16,
                      "end": 0.32,
                      "probability": 0.9375
                    },
                    {
                      "word": "容",
                      "start": 0.32,
                      "end": 0.48,
                      "probability": 1.0
                    },
                    {
                      "word": "类",
                      "start": 0.48,
                      "end": 0.66,
                      "probability": 0.9990234375
                    },
                    {
                      "word": "型,",
                      "start": 0.66,
                      "end": 0.82,
                      "probability": 0.99853515625
                    },
                    {
                      "word": "写",
                      "start": 0.9,
                      "end": 1.48,
                      "probability": 0.9990234375
                    },
                    {
                      "word": "好",
                      "start": 1.48,
                      "end": 1.56,
                      "probability": 1.0
                    },
                    {
                      "word": "背",
                      "start": 1.56,
                      "end": 1.68,
                      "probability": 0.99951171875
                    },
                    {
                      "word": "景,",
                      "start": 1.68,
                      "end": 1.96,
                      "probability": 1.0
                    },
                    {
                      "word": "说",
                      "start": 2.02,
                      "end": 2.46,
                      "probability": 0.99951171875
                    },
                    {
                      "word": "明",
                      "start": 2.46,
                      "end": 2.78,
                      "probability": 0.998046875
                    },
                    {
                      "word": "和",
                      "start": 2.78,
                      "end": 3.1,
                      "probability": 0.986328125
                    },
                    {
                      "word": "数",
                      "start": 3.1,
                      "end": 3.28,
                      "probability": 0.47998046875
                    },
                    {
                      "word": "据",
                      "start": 3.28,
                      "end": 3.42,
                      "probability": 0.7392578125
                    },
                    {
                      "word": "表,",
                      "start": 3.42,
                      "end": 3.6,
                      "probability": 1.0
                    },
                    {
                      "word": "甚",
                      "start": 3.72,
                      "end": 4.28,
                      "probability": 0.99951171875
                    },
                    {
                      "word": "至",
                      "start": 4.28,
                      "end": 4.54,
                      "probability": 1.0
                    },
                    {
                      "word": "直接",
                      "start": 4.54,
                      "end": 5.1,
                      "probability": 0.978515625
                    },
                    {
                      "word": "点",
                      "start": 5.1,
                      "end": 5.4,
                      "probability": 0.99951171875
                    },
                    {
                      "word": "一下。",
                      "start": 5.4,
                      "end": 5.58,
                      "probability": 0.84423828125
                    }
                  ]
                }
              ],
              "language": "zh",
              "duration": 5.7
            },
            "status": "COMPLETE",
            "requested_language": "zh"
          },
          "suppression_proposed": null
        }
      ],
      "language": "zh",
      "packet_ids": [
        3
      ],
      "classification": "BEST_SUPPORTED_WITH_REVIEW",
      "independent_ASR": {
        "engine": "whisper-large-v3",
        "model_path": "/Users/sagawa/AI_WORK/cache/huggingface/hub/models--mlx-community--whisper-large-v3",
        "language": "zh",
        "audio_interval": [
          101.4,
          107.10000000000001
        ],
        "target_interval": [
          103.4,
          105.10000000000001
        ],
        "text": "说明和数据表,",
        "words": [
          {
            "word": "说",
            "start": 103.42,
            "end": 103.86,
            "probability": 0.99951171875
          },
          {
            "word": "明",
            "start": 103.86,
            "end": 104.18,
            "probability": 0.998046875
          },
          {
            "word": "和",
            "start": 104.18,
            "end": 104.5,
            "probability": 0.986328125
          },
          {
            "word": "数",
            "start": 104.5,
            "end": 104.68,
            "probability": 0.47998046875
          },
          {
            "word": "据",
            "start": 104.68,
            "end": 104.82000000000001,
            "probability": 0.7392578125
          },
          {
            "word": "表,",
            "start": 104.82000000000001,
            "end": 105.0,
            "probability": 1.0
          }
        ],
        "raw_result": {
          "text": "可以按内容类型,写好背景,说明和数据表,甚至直接点一下。",
          "segments": [
            {
              "id": 0,
              "seek": 0,
              "start": 0.0,
              "end": 5.58,
              "text": "可以按内容类型,写好背景,说明和数据表,甚至直接点一下。",
              "tokens": [
                50365,
                6723,
                26613,
                34742,
                25750,
                22113,
                119,
                39823,
                11,
                5676,
                247,
                2131,
                46329,
                50218,
                11,
                8090,
                11100,
                12565,
                33188,
                26075,
                106,
                17571,
                11,
                17556,
                20844,
                43297,
                12579,
                8861,
                1543,
                50650
              ],
              "temperature": 0.0,
              "avg_logprob": -0.1863205663619503,
              "compression_ratio": 0.8764044943820225,
              "no_speech_prob": 0.8773801326751709,
              "words": [
                {
                  "word": "可以",
                  "start": 0.0,
                  "end": 0.1,
                  "probability": 0.444580078125
                },
                {
                  "word": "按",
                  "start": 0.1,
                  "end": 0.16,
                  "probability": 0.99560546875
                },
                {
                  "word": "内",
                  "start": 0.16,
                  "end": 0.32,
                  "probability": 0.9375
                },
                {
                  "word": "容",
                  "start": 0.32,
                  "end": 0.48,
                  "probability": 1.0
                },
                {
                  "word": "类",
                  "start": 0.48,
                  "end": 0.66,
                  "probability": 0.9990234375
                },
                {
                  "word": "型,",
                  "start": 0.66,
                  "end": 0.82,
                  "probability": 0.99853515625
                },
                {
                  "word": "写",
                  "start": 0.9,
                  "end": 1.48,
                  "probability": 0.9990234375
                },
                {
                  "word": "好",
                  "start": 1.48,
                  "end": 1.56,
                  "probability": 1.0
                },
                {
                  "word": "背",
                  "start": 1.56,
                  "end": 1.68,
                  "probability": 0.99951171875
                },
                {
                  "word": "景,",
                  "start": 1.68,
                  "end": 1.96,
                  "probability": 1.0
                },
                {
                  "word": "说",
                  "start": 2.02,
                  "end": 2.46,
                  "probability": 0.99951171875
                },
                {
                  "word": "明",
                  "start": 2.46,
                  "end": 2.78,
                  "probability": 0.998046875
                },
                {
                  "word": "和",
                  "start": 2.78,
                  "end": 3.1,
                  "probability": 0.986328125
                },
                {
                  "word": "数",
                  "start": 3.1,
                  "end": 3.28,
                  "probability": 0.47998046875
                },
                {
                  "word": "据",
                  "start": 3.28,
                  "end": 3.42,
                  "probability": 0.7392578125
                },
                {
                  "word": "表,",
                  "start": 3.42,
                  "end": 3.6,
                  "probability": 1.0
                },
                {
                  "word": "甚",
                  "start": 3.72,
                  "end": 4.28,
                  "probability": 0.99951171875
                },
                {
                  "word": "至",
                  "start": 4.28,
                  "end": 4.54,
                  "probability": 1.0
                },
                {
                  "word": "直接",
                  "start": 4.54,
                  "end": 5.1,
                  "probability": 0.978515625
                },
                {
                  "word": "点",
                  "start": 5.1,
                  "end": 5.4,
                  "probability": 0.99951171875
                },
                {
                  "word": "一下。",
                  "start": 5.4,
                  "end": 5.58,
                  "probability": 0.84423828125
                }
              ]
            }
          ],
          "language": "zh",
          "duration": 5.7
        },
        "status": "COMPLETE",
        "requested_language": "zh"
      },
      "raw_word_timestamps": [
        {
          "word": "说",
          "start": 103.4,
          "end": 103.94000000000001,
          "probability": 0.78369140625
        },
        {
          "word": "明",
          "start": 103.94000000000001,
          "end": 104.18,
          "probability": 0.99951171875
        },
        {
          "word": "和",
          "start": 104.18,
          "end": 104.60000000000001,
          "probability": 0.99951171875
        },
        {
          "word": "数",
          "start": 104.60000000000001,
          "end": 104.74000000000001,
          "probability": 0.438232421875
        },
        {
          "word": "语",
          "start": 104.74000000000001,
          "end": 104.88000000000001,
          "probability": 0.986328125
        },
        {
          "word": "表",
          "start": 104.88000000000001,
          "end": 105.10000000000001,
          "probability": 0.99951171875
        }
      ],
      "word_timestamps": [],
      "source_text": "数语表",
      "review_status": "BEST_SUPPORTED_WITH_REVIEW",
      "canonical_timing_evidence": "CUE_ANCHORED",
      "index": 52
    },
    {
      "start": 105.5,
      "end": 109.26,
      "text": "甚至直接点一下让大模型帮你自动生成术语",
      "raw_asr_text": "甚至直接点一下让大模型帮你自动生成术语",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 4,
      "words": [
        {
          "word": "甚",
          "start": 105.4,
          "end": 105.76,
          "probability": 0.1585693359375
        },
        {
          "word": "至",
          "start": 105.76,
          "end": 106.04,
          "probability": 0.99951171875
        },
        {
          "word": "直接",
          "start": 106.04,
          "end": 106.64,
          "probability": 0.9052734375
        },
        {
          "word": "点",
          "start": 106.64,
          "end": 106.86,
          "probability": 0.9990234375
        },
        {
          "word": "一下",
          "start": 106.86,
          "end": 107.22,
          "probability": 0.99853515625
        },
        {
          "word": "让",
          "start": 107.22,
          "end": 107.56,
          "probability": 0.984375
        },
        {
          "word": "大",
          "start": 107.56,
          "end": 107.74000000000001,
          "probability": 0.99755859375
        },
        {
          "word": "模",
          "start": 107.74000000000001,
          "end": 107.88000000000001,
          "probability": 0.9990234375
        },
        {
          "word": "型",
          "start": 107.88000000000001,
          "end": 108.06,
          "probability": 0.99951171875
        },
        {
          "word": "帮",
          "start": 108.06,
          "end": 108.24000000000001,
          "probability": 0.99853515625
        },
        {
          "word": "你",
          "start": 108.24000000000001,
          "end": 108.38000000000001,
          "probability": 0.99755859375
        },
        {
          "word": "自",
          "start": 108.38000000000001,
          "end": 108.52,
          "probability": 0.99755859375
        },
        {
          "word": "动",
          "start": 108.52,
          "end": 108.64,
          "probability": 0.998046875
        },
        {
          "word": "生",
          "start": 108.64,
          "end": 108.78,
          "probability": 0.33984375
        },
        {
          "word": "成",
          "start": 108.78,
          "end": 108.98,
          "probability": 0.99462890625
        },
        {
          "word": "术",
          "start": 108.98,
          "end": 109.12,
          "probability": 0.97412109375
        },
        {
          "word": "语",
          "start": 109.12,
          "end": 109.26,
          "probability": 0.9990234375
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.1897983819666043,
        "compression_ratio": 1.2267657992565055,
        "no_speech_prob": 6.547795639022524e-11
      },
      "lid_status": "UNCERTAIN",
      "alternatives": [
        {
          "language": "zh",
          "text": "甚至直接点一下让大模型帮你自动生成术语翻译时它还会自动结合上文语境不会翻得前言不搭后语翻译接入的是你自己的大模型API像Deepstick同意签问或者任何建容OpenAI的接口都能用按你自己的用量计备啊会自动释出关闭模型思考过程的参数来降低延迟到底"
        }
      ],
      "id": 53,
      "review_annotation": "",
      "row_id": "53",
      "bounded_row_id": 53,
      "parent_row_ids": [
        53
      ],
      "parent_row_id": 53,
      "packet_id": 4,
      "LID_evidence": [
        {
          "region_id": 18,
          "probabilities": {
            "zh": 0.997512956460317
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        18
      ],
      "language_confidence": 0.997512956460317,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 104.7,
      "packet_end": 134.10000000000002,
      "raw_text": "甚至直接点一下让大模型帮你自动生成术语",
      "selected_text": "甚至直接点一下让大模型帮你自动生成术语",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            53
          ],
          "packet_id": 4,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        4
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 53
    },
    {
      "start": 109.64,
      "end": 112.34,
      "text": "翻译时它还会自动结合上文语境",
      "raw_asr_text": "翻译时它还会自动结合上文语境",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 4,
      "words": [
        {
          "word": "翻",
          "start": 109.64,
          "end": 110.0,
          "probability": 0.986328125
        },
        {
          "word": "译",
          "start": 110.0,
          "end": 110.18,
          "probability": 0.9990234375
        },
        {
          "word": "时",
          "start": 110.18,
          "end": 110.4,
          "probability": 0.9970703125
        },
        {
          "word": "它",
          "start": 110.4,
          "end": 110.84,
          "probability": 0.93701171875
        },
        {
          "word": "还",
          "start": 110.84,
          "end": 110.96000000000001,
          "probability": 0.99853515625
        },
        {
          "word": "会",
          "start": 110.96000000000001,
          "end": 111.10000000000001,
          "probability": 0.99951171875
        },
        {
          "word": "自",
          "start": 111.10000000000001,
          "end": 111.28,
          "probability": 0.99951171875
        },
        {
          "word": "动",
          "start": 111.28,
          "end": 111.42,
          "probability": 0.99951171875
        },
        {
          "word": "结",
          "start": 111.42,
          "end": 111.58,
          "probability": 0.9990234375
        },
        {
          "word": "合",
          "start": 111.58,
          "end": 111.72,
          "probability": 0.998046875
        },
        {
          "word": "上",
          "start": 111.72,
          "end": 111.86,
          "probability": 0.99951171875
        },
        {
          "word": "文",
          "start": 111.86,
          "end": 112.02000000000001,
          "probability": 0.96728515625
        },
        {
          "word": "语",
          "start": 112.02000000000001,
          "end": 112.14,
          "probability": 0.76171875
        },
        {
          "word": "境",
          "start": 112.14,
          "end": 112.34,
          "probability": 0.455078125
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.1897983819666043,
        "compression_ratio": 1.2267657992565055,
        "no_speech_prob": 6.547795639022524e-11
      },
      "lid_status": "UNCERTAIN",
      "alternatives": [
        {
          "language": "zh",
          "text": "甚至直接点一下让大模型帮你自动生成术语翻译时它还会自动结合上文语境不会翻得前言不搭后语翻译接入的是你自己的大模型API像Deepstick同意签问或者任何建容OpenAI的接口都能用按你自己的用量计备啊会自动释出关闭模型思考过程的参数来降低延迟到底"
        }
      ],
      "id": 54,
      "review_annotation": "",
      "row_id": "54",
      "bounded_row_id": 54,
      "parent_row_ids": [
        54
      ],
      "parent_row_id": 54,
      "packet_id": 4,
      "LID_evidence": [
        {
          "region_id": 19,
          "probabilities": {
            "zh": 0.9634838955742973
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        19
      ],
      "language_confidence": 0.9634838955742973,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 104.7,
      "packet_end": 134.10000000000002,
      "raw_text": "翻译时它还会自动结合上文语境",
      "selected_text": "翻译时它还会自动结合上文语境",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            54
          ],
          "packet_id": 4,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        4
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 54
    },
    {
      "start": 112.34,
      "end": 114.62,
      "text": "不会翻得前言不搭后语",
      "raw_asr_text": "不会翻得前言不搭后语",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 4,
      "words": [
        {
          "word": "不会",
          "start": 112.34,
          "end": 113.06,
          "probability": 0.99365234375
        },
        {
          "word": "翻",
          "start": 113.06,
          "end": 113.26,
          "probability": 0.9814453125
        },
        {
          "word": "得",
          "start": 113.26,
          "end": 113.46000000000001,
          "probability": 0.6572265625
        },
        {
          "word": "前",
          "start": 113.46000000000001,
          "end": 113.84,
          "probability": 0.998046875
        },
        {
          "word": "言",
          "start": 113.84,
          "end": 113.98,
          "probability": 0.5185546875
        },
        {
          "word": "不",
          "start": 113.98,
          "end": 114.16,
          "probability": 0.99560546875
        },
        {
          "word": "搭",
          "start": 114.16,
          "end": 114.3,
          "probability": 0.724609375
        },
        {
          "word": "后",
          "start": 114.3,
          "end": 114.48,
          "probability": 0.99365234375
        },
        {
          "word": "语",
          "start": 114.48,
          "end": 114.62,
          "probability": 0.9892578125
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.1897983819666043,
        "compression_ratio": 1.2267657992565055,
        "no_speech_prob": 6.547795639022524e-11
      },
      "lid_status": "UNCERTAIN",
      "alternatives": [
        {
          "language": "zh",
          "text": "甚至直接点一下让大模型帮你自动生成术语翻译时它还会自动结合上文语境不会翻得前言不搭后语翻译接入的是你自己的大模型API像Deepstick同意签问或者任何建容OpenAI的接口都能用按你自己的用量计备啊会自动释出关闭模型思考过程的参数来降低延迟到底"
        }
      ],
      "id": 55,
      "review_annotation": "",
      "row_id": "55",
      "bounded_row_id": 55,
      "parent_row_ids": [
        55
      ],
      "parent_row_id": 55,
      "packet_id": 4,
      "LID_evidence": [
        {
          "region_id": 19,
          "probabilities": {
            "zh": 0.9634838955742973
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        19
      ],
      "language_confidence": 0.9634838955742973,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 104.7,
      "packet_end": 134.10000000000002,
      "raw_text": "不会翻得前言不搭后语",
      "selected_text": "不会翻得前言不搭后语",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            55
          ],
          "packet_id": 4,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        4
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 55
    },
    {
      "start": 115.24000000000001,
      "end": 119.0,
      "text": "翻译接入的是你自己的大模型API",
      "raw_asr_text": "翻译接入的是你自己的大模型API",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 4,
      "words": [
        {
          "word": "翻",
          "start": 115.24000000000001,
          "end": 115.60000000000001,
          "probability": 0.99951171875
        },
        {
          "word": "译",
          "start": 115.60000000000001,
          "end": 115.8,
          "probability": 1.0
        },
        {
          "word": "接",
          "start": 115.8,
          "end": 116.0,
          "probability": 0.5869140625
        },
        {
          "word": "入",
          "start": 116.0,
          "end": 116.16,
          "probability": 0.99609375
        },
        {
          "word": "的是",
          "start": 116.16,
          "end": 116.44,
          "probability": 0.98583984375
        },
        {
          "word": "你",
          "start": 116.44,
          "end": 116.84,
          "probability": 0.9990234375
        },
        {
          "word": "自己",
          "start": 116.84,
          "end": 117.12,
          "probability": 0.99951171875
        },
        {
          "word": "的",
          "start": 117.12,
          "end": 117.28,
          "probability": 1.0
        },
        {
          "word": "大",
          "start": 117.28,
          "end": 117.42,
          "probability": 1.0
        },
        {
          "word": "模",
          "start": 117.42,
          "end": 117.58,
          "probability": 1.0
        },
        {
          "word": "型",
          "start": 117.58,
          "end": 117.78,
          "probability": 0.99951171875
        },
        {
          "word": "AP",
          "start": 117.78,
          "end": 118.08,
          "probability": 0.94677734375
        },
        {
          "word": "I",
          "start": 118.08,
          "end": 119.0,
          "probability": 1.0
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.1897983819666043,
        "compression_ratio": 1.2267657992565055,
        "no_speech_prob": 6.547795639022524e-11
      },
      "lid_status": "UNCERTAIN",
      "alternatives": [
        {
          "language": "zh",
          "text": "甚至直接点一下让大模型帮你自动生成术语翻译时它还会自动结合上文语境不会翻得前言不搭后语翻译接入的是你自己的大模型API像Deepstick同意签问或者任何建容OpenAI的接口都能用按你自己的用量计备啊会自动释出关闭模型思考过程的参数来降低延迟到底"
        }
      ],
      "id": 56,
      "review_annotation": "",
      "row_id": "56",
      "bounded_row_id": 56,
      "parent_row_ids": [
        56
      ],
      "parent_row_id": 56,
      "packet_id": 4,
      "LID_evidence": [
        {
          "region_id": 20,
          "probabilities": {
            "zh": 0.7845521216619421
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        20
      ],
      "language_confidence": 0.7845521216619421,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 104.7,
      "packet_end": 134.10000000000002,
      "raw_text": "翻译接入的是你自己的大模型API",
      "selected_text": "翻译接入的是你自己的大模型API",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            56
          ],
          "packet_id": 4,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        4
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 56
    },
    {
      "start": 119.0,
      "end": 120.08,
      "text": "像Deepstick",
      "raw_asr_text": "像Deepstick",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 4,
      "words": [
        {
          "word": "像",
          "start": 119.0,
          "end": 119.5,
          "probability": 0.99267578125
        },
        {
          "word": "De",
          "start": 119.5,
          "end": 119.68,
          "probability": 0.78515625
        },
        {
          "word": "ep",
          "start": 119.68,
          "end": 119.78,
          "probability": 0.994140625
        },
        {
          "word": "stick",
          "start": 119.78,
          "end": 120.08,
          "probability": 0.61474609375
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.1897983819666043,
        "compression_ratio": 1.2267657992565055,
        "no_speech_prob": 6.547795639022524e-11
      },
      "lid_status": "UNCERTAIN",
      "alternatives": [
        {
          "language": "zh",
          "text": "甚至直接点一下让大模型帮你自动生成术语翻译时它还会自动结合上文语境不会翻得前言不搭后语翻译接入的是你自己的大模型API像Deepstick同意签问或者任何建容OpenAI的接口都能用按你自己的用量计备啊会自动释出关闭模型思考过程的参数来降低延迟到底"
        }
      ],
      "id": 57,
      "review_annotation": "待確認",
      "row_id": "57",
      "bounded_row_id": 57,
      "parent_row_ids": [
        57
      ],
      "parent_row_id": 57,
      "packet_id": 4,
      "LID_evidence": [
        {
          "region_id": 21,
          "probabilities": {
            "zh": 0.7078524202418824
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        21
      ],
      "language_confidence": 0.7078524202418824,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 104.7,
      "packet_end": 134.10000000000002,
      "raw_text": "像Deepstick",
      "selected_text": "像Deepstick",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            57
          ],
          "packet_id": 4,
          "engine": "mlx-whisper"
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "57",
          "before": "像Deepstick",
          "after": "像Deepstick",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "Deepstick疑似DeepSeek误识",
          "context_evidence": "上下文为AI模型，'DeepSeek'为常见大模型",
          "independent_ASR": {
            "engine": "whisper-large-v3",
            "model_path": "/Users/sagawa/AI_WORK/cache/huggingface/hub/models--mlx-community--whisper-large-v3",
            "language": "zh",
            "audio_interval": [
              117.0,
              122.08
            ],
            "target_interval": [
              119.0,
              120.08
            ],
            "text": "像Dipstick",
            "words": [
              {
                "word": "像",
                "start": 118.92,
                "end": 119.34,
                "probability": 0.6943359375
              },
              {
                "word": "D",
                "start": 119.34,
                "end": 119.58,
                "probability": 0.394775390625
              },
              {
                "word": "ip",
                "start": 119.58,
                "end": 119.7,
                "probability": 0.9580078125
              },
              {
                "word": "stick",
                "start": 119.7,
                "end": 119.98,
                "probability": 0.76708984375
              }
            ],
            "raw_result": {
              "text": "自己的大模型API像Dipstick同意签问或者任何",
              "segments": [
                {
                  "id": 0,
                  "seek": 0,
                  "start": 0.0,
                  "end": 1.3,
                  "text": "自己的大模型API",
                  "tokens": [
                    50365,
                    17645,
                    1546,
                    3582,
                    41908,
                    39823,
                    4715,
                    40,
                    50433
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.29771603856767925,
                  "compression_ratio": 0.835820895522388,
                  "no_speech_prob": 0.8034570217132568,
                  "words": [
                    {
                      "word": "自己",
                      "start": 0.0,
                      "end": 0.12,
                      "probability": 0.284423828125
                    },
                    {
                      "word": "的",
                      "start": 0.12,
                      "end": 0.24,
                      "probability": 0.99365234375
                    },
                    {
                      "word": "大",
                      "start": 0.24,
                      "end": 0.34,
                      "probability": 0.99609375
                    },
                    {
                      "word": "模",
                      "start": 0.34,
                      "end": 0.5,
                      "probability": 0.998046875
                    },
                    {
                      "word": "型",
                      "start": 0.5,
                      "end": 0.74,
                      "probability": 0.998046875
                    },
                    {
                      "word": "AP",
                      "start": 0.74,
                      "end": 0.98,
                      "probability": 0.7900390625
                    },
                    {
                      "word": "I",
                      "start": 0.98,
                      "end": 1.3,
                      "probability": 0.9931640625
                    }
                  ]
                },
                {
                  "id": 1,
                  "seek": 0,
                  "start": 1.9200000000000004,
                  "end": 2.98,
                  "text": "像Dipstick",
                  "tokens": [
                    50433,
                    12760,
                    35,
                    647,
                    11881,
                    50521
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.29771603856767925,
                  "compression_ratio": 0.835820895522388,
                  "no_speech_prob": 0.8034570217132568,
                  "words": [
                    {
                      "word": "像",
                      "start": 1.9200000000000004,
                      "end": 2.34,
                      "probability": 0.6943359375
                    },
                    {
                      "word": "D",
                      "start": 2.34,
                      "end": 2.58,
                      "probability": 0.394775390625
                    },
                    {
                      "word": "ip",
                      "start": 2.58,
                      "end": 2.7,
                      "probability": 0.9580078125
                    },
                    {
                      "word": "stick",
                      "start": 2.7,
                      "end": 2.98,
                      "probability": 0.76708984375
                    }
                  ]
                },
                {
                  "id": 2,
                  "seek": 0,
                  "start": 2.98,
                  "end": 4.3,
                  "text": "同意签问",
                  "tokens": [
                    50521,
                    13089,
                    9042,
                    7973,
                    122,
                    22064,
                    50586
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.29771603856767925,
                  "compression_ratio": 0.835820895522388,
                  "no_speech_prob": 0.8034570217132568,
                  "words": [
                    {
                      "word": "同",
                      "start": 2.98,
                      "end": 3.72,
                      "probability": 0.32568359375
                    },
                    {
                      "word": "意",
                      "start": 3.72,
                      "end": 3.94,
                      "probability": 0.65625
                    },
                    {
                      "word": "签",
                      "start": 3.94,
                      "end": 4.14,
                      "probability": 0.8994140625
                    },
                    {
                      "word": "问",
                      "start": 4.14,
                      "end": 4.3,
                      "probability": 0.580078125
                    }
                  ]
                },
                {
                  "id": 3,
                  "seek": 0,
                  "start": 4.3,
                  "end": 5.06,
                  "text": "或者任何",
                  "tokens": [
                    50586,
                    31148,
                    26443,
                    11906,
                    50619
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.29771603856767925,
                  "compression_ratio": 0.835820895522388,
                  "no_speech_prob": 0.8034570217132568,
                  "words": [
                    {
                      "word": "或者",
                      "start": 4.3,
                      "end": 4.74,
                      "probability": 0.96435546875
                    },
                    {
                      "word": "任",
                      "start": 4.74,
                      "end": 4.92,
                      "probability": 0.99609375
                    },
                    {
                      "word": "何",
                      "start": 4.92,
                      "end": 5.06,
                      "probability": 0.99560546875
                    }
                  ]
                }
              ],
              "language": "zh",
              "duration": 5.08
            },
            "status": "COMPLETE",
            "requested_language": "zh"
          },
          "suppression_proposed": null
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "57",
          "before": "像Deepstick",
          "after": "像Deepstick",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "Name error: 'Deepstick' vs 'DeepSeek'",
          "context_evidence": "Context: '像DeepSeek'",
          "independent_ASR": {
            "engine": "whisper-large-v3",
            "model_path": "/Users/sagawa/AI_WORK/cache/huggingface/hub/models--mlx-community--whisper-large-v3",
            "language": "zh",
            "audio_interval": [
              117.0,
              122.08
            ],
            "target_interval": [
              119.0,
              120.08
            ],
            "text": "像Dipstick",
            "words": [
              {
                "word": "像",
                "start": 118.92,
                "end": 119.34,
                "probability": 0.6943359375
              },
              {
                "word": "D",
                "start": 119.34,
                "end": 119.58,
                "probability": 0.394775390625
              },
              {
                "word": "ip",
                "start": 119.58,
                "end": 119.7,
                "probability": 0.9580078125
              },
              {
                "word": "stick",
                "start": 119.7,
                "end": 119.98,
                "probability": 0.76708984375
              }
            ],
            "raw_result": {
              "text": "自己的大模型API像Dipstick同意签问或者任何",
              "segments": [
                {
                  "id": 0,
                  "seek": 0,
                  "start": 0.0,
                  "end": 1.3,
                  "text": "自己的大模型API",
                  "tokens": [
                    50365,
                    17645,
                    1546,
                    3582,
                    41908,
                    39823,
                    4715,
                    40,
                    50433
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.29771603856767925,
                  "compression_ratio": 0.835820895522388,
                  "no_speech_prob": 0.8034570217132568,
                  "words": [
                    {
                      "word": "自己",
                      "start": 0.0,
                      "end": 0.12,
                      "probability": 0.284423828125
                    },
                    {
                      "word": "的",
                      "start": 0.12,
                      "end": 0.24,
                      "probability": 0.99365234375
                    },
                    {
                      "word": "大",
                      "start": 0.24,
                      "end": 0.34,
                      "probability": 0.99609375
                    },
                    {
                      "word": "模",
                      "start": 0.34,
                      "end": 0.5,
                      "probability": 0.998046875
                    },
                    {
                      "word": "型",
                      "start": 0.5,
                      "end": 0.74,
                      "probability": 0.998046875
                    },
                    {
                      "word": "AP",
                      "start": 0.74,
                      "end": 0.98,
                      "probability": 0.7900390625
                    },
                    {
                      "word": "I",
                      "start": 0.98,
                      "end": 1.3,
                      "probability": 0.9931640625
                    }
                  ]
                },
                {
                  "id": 1,
                  "seek": 0,
                  "start": 1.9200000000000004,
                  "end": 2.98,
                  "text": "像Dipstick",
                  "tokens": [
                    50433,
                    12760,
                    35,
                    647,
                    11881,
                    50521
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.29771603856767925,
                  "compression_ratio": 0.835820895522388,
                  "no_speech_prob": 0.8034570217132568,
                  "words": [
                    {
                      "word": "像",
                      "start": 1.9200000000000004,
                      "end": 2.34,
                      "probability": 0.6943359375
                    },
                    {
                      "word": "D",
                      "start": 2.34,
                      "end": 2.58,
                      "probability": 0.394775390625
                    },
                    {
                      "word": "ip",
                      "start": 2.58,
                      "end": 2.7,
                      "probability": 0.9580078125
                    },
                    {
                      "word": "stick",
                      "start": 2.7,
                      "end": 2.98,
                      "probability": 0.76708984375
                    }
                  ]
                },
                {
                  "id": 2,
                  "seek": 0,
                  "start": 2.98,
                  "end": 4.3,
                  "text": "同意签问",
                  "tokens": [
                    50521,
                    13089,
                    9042,
                    7973,
                    122,
                    22064,
                    50586
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.29771603856767925,
                  "compression_ratio": 0.835820895522388,
                  "no_speech_prob": 0.8034570217132568,
                  "words": [
                    {
                      "word": "同",
                      "start": 2.98,
                      "end": 3.72,
                      "probability": 0.32568359375
                    },
                    {
                      "word": "意",
                      "start": 3.72,
                      "end": 3.94,
                      "probability": 0.65625
                    },
                    {
                      "word": "签",
                      "start": 3.94,
                      "end": 4.14,
                      "probability": 0.8994140625
                    },
                    {
                      "word": "问",
                      "start": 4.14,
                      "end": 4.3,
                      "probability": 0.580078125
                    }
                  ]
                },
                {
                  "id": 3,
                  "seek": 0,
                  "start": 4.3,
                  "end": 5.06,
                  "text": "或者任何",
                  "tokens": [
                    50586,
                    31148,
                    26443,
                    11906,
                    50619
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.29771603856767925,
                  "compression_ratio": 0.835820895522388,
                  "no_speech_prob": 0.8034570217132568,
                  "words": [
                    {
                      "word": "或者",
                      "start": 4.3,
                      "end": 4.74,
                      "probability": 0.96435546875
                    },
                    {
                      "word": "任",
                      "start": 4.74,
                      "end": 4.92,
                      "probability": 0.99609375
                    },
                    {
                      "word": "何",
                      "start": 4.92,
                      "end": 5.06,
                      "probability": 0.99560546875
                    }
                  ]
                }
              ],
              "language": "zh",
              "duration": 5.08
            },
            "status": "COMPLETE",
            "requested_language": "zh"
          },
          "suppression_proposed": null
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "57",
          "before": "像Deepstick",
          "after": "像Deepstick",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "Name error: 'Deepstick' vs 'DeepSeek'",
          "context_evidence": "Context: '像DeepSeek'",
          "independent_ASR": {
            "engine": "whisper-large-v3",
            "model_path": "/Users/sagawa/AI_WORK/cache/huggingface/hub/models--mlx-community--whisper-large-v3",
            "language": "zh",
            "audio_interval": [
              117.0,
              122.08
            ],
            "target_interval": [
              119.0,
              120.08
            ],
            "text": "像Dipstick",
            "words": [
              {
                "word": "像",
                "start": 118.92,
                "end": 119.34,
                "probability": 0.6943359375
              },
              {
                "word": "D",
                "start": 119.34,
                "end": 119.58,
                "probability": 0.394775390625
              },
              {
                "word": "ip",
                "start": 119.58,
                "end": 119.7,
                "probability": 0.9580078125
              },
              {
                "word": "stick",
                "start": 119.7,
                "end": 119.98,
                "probability": 0.76708984375
              }
            ],
            "raw_result": {
              "text": "自己的大模型API像Dipstick同意签问或者任何",
              "segments": [
                {
                  "id": 0,
                  "seek": 0,
                  "start": 0.0,
                  "end": 1.3,
                  "text": "自己的大模型API",
                  "tokens": [
                    50365,
                    17645,
                    1546,
                    3582,
                    41908,
                    39823,
                    4715,
                    40,
                    50433
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.29771603856767925,
                  "compression_ratio": 0.835820895522388,
                  "no_speech_prob": 0.8034570217132568,
                  "words": [
                    {
                      "word": "自己",
                      "start": 0.0,
                      "end": 0.12,
                      "probability": 0.284423828125
                    },
                    {
                      "word": "的",
                      "start": 0.12,
                      "end": 0.24,
                      "probability": 0.99365234375
                    },
                    {
                      "word": "大",
                      "start": 0.24,
                      "end": 0.34,
                      "probability": 0.99609375
                    },
                    {
                      "word": "模",
                      "start": 0.34,
                      "end": 0.5,
                      "probability": 0.998046875
                    },
                    {
                      "word": "型",
                      "start": 0.5,
                      "end": 0.74,
                      "probability": 0.998046875
                    },
                    {
                      "word": "AP",
                      "start": 0.74,
                      "end": 0.98,
                      "probability": 0.7900390625
                    },
                    {
                      "word": "I",
                      "start": 0.98,
                      "end": 1.3,
                      "probability": 0.9931640625
                    }
                  ]
                },
                {
                  "id": 1,
                  "seek": 0,
                  "start": 1.9200000000000004,
                  "end": 2.98,
                  "text": "像Dipstick",
                  "tokens": [
                    50433,
                    12760,
                    35,
                    647,
                    11881,
                    50521
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.29771603856767925,
                  "compression_ratio": 0.835820895522388,
                  "no_speech_prob": 0.8034570217132568,
                  "words": [
                    {
                      "word": "像",
                      "start": 1.9200000000000004,
                      "end": 2.34,
                      "probability": 0.6943359375
                    },
                    {
                      "word": "D",
                      "start": 2.34,
                      "end": 2.58,
                      "probability": 0.394775390625
                    },
                    {
                      "word": "ip",
                      "start": 2.58,
                      "end": 2.7,
                      "probability": 0.9580078125
                    },
                    {
                      "word": "stick",
                      "start": 2.7,
                      "end": 2.98,
                      "probability": 0.76708984375
                    }
                  ]
                },
                {
                  "id": 2,
                  "seek": 0,
                  "start": 2.98,
                  "end": 4.3,
                  "text": "同意签问",
                  "tokens": [
                    50521,
                    13089,
                    9042,
                    7973,
                    122,
                    22064,
                    50586
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.29771603856767925,
                  "compression_ratio": 0.835820895522388,
                  "no_speech_prob": 0.8034570217132568,
                  "words": [
                    {
                      "word": "同",
                      "start": 2.98,
                      "end": 3.72,
                      "probability": 0.32568359375
                    },
                    {
                      "word": "意",
                      "start": 3.72,
                      "end": 3.94,
                      "probability": 0.65625
                    },
                    {
                      "word": "签",
                      "start": 3.94,
                      "end": 4.14,
                      "probability": 0.8994140625
                    },
                    {
                      "word": "问",
                      "start": 4.14,
                      "end": 4.3,
                      "probability": 0.580078125
                    }
                  ]
                },
                {
                  "id": 3,
                  "seek": 0,
                  "start": 4.3,
                  "end": 5.06,
                  "text": "或者任何",
                  "tokens": [
                    50586,
                    31148,
                    26443,
                    11906,
                    50619
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.29771603856767925,
                  "compression_ratio": 0.835820895522388,
                  "no_speech_prob": 0.8034570217132568,
                  "words": [
                    {
                      "word": "或者",
                      "start": 4.3,
                      "end": 4.74,
                      "probability": 0.96435546875
                    },
                    {
                      "word": "任",
                      "start": 4.74,
                      "end": 4.92,
                      "probability": 0.99609375
                    },
                    {
                      "word": "何",
                      "start": 4.92,
                      "end": 5.06,
                      "probability": 0.99560546875
                    }
                  ]
                }
              ],
              "language": "zh",
              "duration": 5.08
            },
            "status": "COMPLETE",
            "requested_language": "zh"
          },
          "suppression_proposed": null
        }
      ],
      "language": "zh",
      "packet_ids": [
        4
      ],
      "classification": "BEST_SUPPORTED_WITH_REVIEW",
      "independent_ASR": {
        "engine": "whisper-large-v3",
        "model_path": "/Users/sagawa/AI_WORK/cache/huggingface/hub/models--mlx-community--whisper-large-v3",
        "language": "zh",
        "audio_interval": [
          117.0,
          122.08
        ],
        "target_interval": [
          119.0,
          120.08
        ],
        "text": "像Dipstick",
        "words": [
          {
            "word": "像",
            "start": 118.92,
            "end": 119.34,
            "probability": 0.6943359375
          },
          {
            "word": "D",
            "start": 119.34,
            "end": 119.58,
            "probability": 0.394775390625
          },
          {
            "word": "ip",
            "start": 119.58,
            "end": 119.7,
            "probability": 0.9580078125
          },
          {
            "word": "stick",
            "start": 119.7,
            "end": 119.98,
            "probability": 0.76708984375
          }
        ],
        "raw_result": {
          "text": "自己的大模型API像Dipstick同意签问或者任何",
          "segments": [
            {
              "id": 0,
              "seek": 0,
              "start": 0.0,
              "end": 1.3,
              "text": "自己的大模型API",
              "tokens": [
                50365,
                17645,
                1546,
                3582,
                41908,
                39823,
                4715,
                40,
                50433
              ],
              "temperature": 0.0,
              "avg_logprob": -0.29771603856767925,
              "compression_ratio": 0.835820895522388,
              "no_speech_prob": 0.8034570217132568,
              "words": [
                {
                  "word": "自己",
                  "start": 0.0,
                  "end": 0.12,
                  "probability": 0.284423828125
                },
                {
                  "word": "的",
                  "start": 0.12,
                  "end": 0.24,
                  "probability": 0.99365234375
                },
                {
                  "word": "大",
                  "start": 0.24,
                  "end": 0.34,
                  "probability": 0.99609375
                },
                {
                  "word": "模",
                  "start": 0.34,
                  "end": 0.5,
                  "probability": 0.998046875
                },
                {
                  "word": "型",
                  "start": 0.5,
                  "end": 0.74,
                  "probability": 0.998046875
                },
                {
                  "word": "AP",
                  "start": 0.74,
                  "end": 0.98,
                  "probability": 0.7900390625
                },
                {
                  "word": "I",
                  "start": 0.98,
                  "end": 1.3,
                  "probability": 0.9931640625
                }
              ]
            },
            {
              "id": 1,
              "seek": 0,
              "start": 1.9200000000000004,
              "end": 2.98,
              "text": "像Dipstick",
              "tokens": [
                50433,
                12760,
                35,
                647,
                11881,
                50521
              ],
              "temperature": 0.0,
              "avg_logprob": -0.29771603856767925,
              "compression_ratio": 0.835820895522388,
              "no_speech_prob": 0.8034570217132568,
              "words": [
                {
                  "word": "像",
                  "start": 1.9200000000000004,
                  "end": 2.34,
                  "probability": 0.6943359375
                },
                {
                  "word": "D",
                  "start": 2.34,
                  "end": 2.58,
                  "probability": 0.394775390625
                },
                {
                  "word": "ip",
                  "start": 2.58,
                  "end": 2.7,
                  "probability": 0.9580078125
                },
                {
                  "word": "stick",
                  "start": 2.7,
                  "end": 2.98,
                  "probability": 0.76708984375
                }
              ]
            },
            {
              "id": 2,
              "seek": 0,
              "start": 2.98,
              "end": 4.3,
              "text": "同意签问",
              "tokens": [
                50521,
                13089,
                9042,
                7973,
                122,
                22064,
                50586
              ],
              "temperature": 0.0,
              "avg_logprob": -0.29771603856767925,
              "compression_ratio": 0.835820895522388,
              "no_speech_prob": 0.8034570217132568,
              "words": [
                {
                  "word": "同",
                  "start": 2.98,
                  "end": 3.72,
                  "probability": 0.32568359375
                },
                {
                  "word": "意",
                  "start": 3.72,
                  "end": 3.94,
                  "probability": 0.65625
                },
                {
                  "word": "签",
                  "start": 3.94,
                  "end": 4.14,
                  "probability": 0.8994140625
                },
                {
                  "word": "问",
                  "start": 4.14,
                  "end": 4.3,
                  "probability": 0.580078125
                }
              ]
            },
            {
              "id": 3,
              "seek": 0,
              "start": 4.3,
              "end": 5.06,
              "text": "或者任何",
              "tokens": [
                50586,
                31148,
                26443,
                11906,
                50619
              ],
              "temperature": 0.0,
              "avg_logprob": -0.29771603856767925,
              "compression_ratio": 0.835820895522388,
              "no_speech_prob": 0.8034570217132568,
              "words": [
                {
                  "word": "或者",
                  "start": 4.3,
                  "end": 4.74,
                  "probability": 0.96435546875
                },
                {
                  "word": "任",
                  "start": 4.74,
                  "end": 4.92,
                  "probability": 0.99609375
                },
                {
                  "word": "何",
                  "start": 4.92,
                  "end": 5.06,
                  "probability": 0.99560546875
                }
              ]
            }
          ],
          "language": "zh",
          "duration": 5.08
        },
        "status": "COMPLETE",
        "requested_language": "zh"
      },
      "source_text": "像Deepstick",
      "review_status": "BEST_SUPPORTED_WITH_REVIEW",
      "canonical_timing_evidence": "CUE_ANCHORED",
      "index": 57
    },
    {
      "start": 120.08,
      "end": 124.24000000000001,
      "text": "建容",
      "raw_asr_text": "同意签问或者任何建容OpenAI的接口都能用",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 4,
      "words": [],
      "asr_metrics": {
        "avg_logprob": -0.1897983819666043,
        "compression_ratio": 1.2267657992565055,
        "no_speech_prob": 6.547795639022524e-11
      },
      "lid_status": "UNCERTAIN",
      "alternatives": [
        {
          "language": "zh",
          "text": "甚至直接点一下让大模型帮你自动生成术语翻译时它还会自动结合上文语境不会翻得前言不搭后语翻译接入的是你自己的大模型API像Deepstick同意签问或者任何建容OpenAI的接口都能用按你自己的用量计备啊会自动释出关闭模型思考过程的参数来降低延迟到底"
        }
      ],
      "id": 58,
      "review_annotation": "待確認",
      "row_id": "58",
      "bounded_row_id": 58,
      "parent_row_ids": [
        58
      ],
      "parent_row_id": 58,
      "packet_id": 4,
      "LID_evidence": [
        {
          "region_id": 21,
          "probabilities": {
            "zh": 0.7078524202418824
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        21
      ],
      "language_confidence": 0.7078524202418824,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 104.7,
      "packet_end": 134.10000000000002,
      "raw_text": "同意签问或者任何建容OpenAI的接口都能用",
      "selected_text": "建容",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            58
          ],
          "packet_id": 4,
          "engine": "mlx-whisper"
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "58",
          "before": "同意签问或者任何建容OpenAI的接口都能用",
          "after": "同意签问",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "同意签问疑似通义千问误识",
          "context_evidence": "上下文为AI模型，'通义千问'为常见大模型",
          "independent_ASR": {
            "engine": "whisper-large-v3",
            "model_path": "/Users/sagawa/AI_WORK/cache/huggingface/hub/models--mlx-community--whisper-large-v3",
            "language": "zh",
            "audio_interval": [
              118.08,
              126.24000000000001
            ],
            "target_interval": [
              120.08,
              124.24000000000001
            ],
            "text": "同一千万或者任何兼容OpenAI的接口都能用",
            "words": [
              {
                "word": "同",
                "start": 119.96,
                "end": 120.72,
                "probability": 0.147705078125
              },
              {
                "word": "一",
                "start": 120.72,
                "end": 120.89999999999999,
                "probability": 0.5234375
              },
              {
                "word": "千",
                "start": 120.89999999999999,
                "end": 121.12,
                "probability": 0.215576171875
              },
              {
                "word": "万",
                "start": 121.12,
                "end": 121.28,
                "probability": 0.82421875
              },
              {
                "word": "或者",
                "start": 121.28,
                "end": 121.72,
                "probability": 0.98583984375
              },
              {
                "word": "任",
                "start": 121.72,
                "end": 121.92,
                "probability": 0.99951171875
              },
              {
                "word": "何",
                "start": 121.92,
                "end": 122.14,
                "probability": 0.99951171875
              },
              {
                "word": "兼",
                "start": 122.14,
                "end": 122.32,
                "probability": 0.9072265625
              },
              {
                "word": "容",
                "start": 122.32,
                "end": 122.5,
                "probability": 1.0
              },
              {
                "word": "Open",
                "start": 122.5,
                "end": 122.82,
                "probability": 0.91064453125
              },
              {
                "word": "AI",
                "start": 122.82,
                "end": 123.12,
                "probability": 0.9931640625
              },
              {
                "word": "的",
                "start": 123.12,
                "end": 123.34,
                "probability": 0.98779296875
              },
              {
                "word": "接",
                "start": 123.34,
                "end": 123.48,
                "probability": 0.87744140625
              },
              {
                "word": "口",
                "start": 123.48,
                "end": 123.64,
                "probability": 0.99951171875
              },
              {
                "word": "都",
                "start": 123.64,
                "end": 123.75999999999999,
                "probability": 0.99658203125
              },
              {
                "word": "能",
                "start": 123.75999999999999,
                "end": 123.92,
                "probability": 0.9990234375
              },
              {
                "word": "用",
                "start": 123.92,
                "end": 124.16,
                "probability": 0.98828125
              }
            ],
            "raw_result": {
              "text": "像Dipstick同一千万或者任何兼容OpenAI的接口都能用按你自己的用量去做",
              "segments": [
                {
                  "id": 0,
                  "seek": 0,
                  "start": 0.8200000000000023,
                  "end": 1.88,
                  "text": "像Dipstick",
                  "tokens": [
                    50365,
                    12760,
                    35,
                    647,
                    11881,
                    50467
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.34069514918971705,
                  "compression_ratio": 0.8932038834951457,
                  "no_speech_prob": 0.852008044719696,
                  "words": [
                    {
                      "word": "像",
                      "start": 0.8200000000000023,
                      "end": 1.26,
                      "probability": 0.1815185546875
                    },
                    {
                      "word": "D",
                      "start": 1.26,
                      "end": 1.5,
                      "probability": 0.576171875
                    },
                    {
                      "word": "ip",
                      "start": 1.5,
                      "end": 1.62,
                      "probability": 0.9765625
                    },
                    {
                      "word": "stick",
                      "start": 1.62,
                      "end": 1.88,
                      "probability": 0.837890625
                    }
                  ]
                },
                {
                  "id": 1,
                  "seek": 0,
                  "start": 1.88,
                  "end": 6.08,
                  "text": "同一千万或者任何兼容OpenAI的接口都能用",
                  "tokens": [
                    50467,
                    13089,
                    2257,
                    20787,
                    23570,
                    31148,
                    26443,
                    11906,
                    2347,
                    120,
                    25750,
                    45569,
                    48698,
                    1546,
                    14468,
                    18144,
                    7182,
                    8225,
                    9254,
                    50674
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.34069514918971705,
                  "compression_ratio": 0.8932038834951457,
                  "no_speech_prob": 0.852008044719696,
                  "words": [
                    {
                      "word": "同",
                      "start": 1.88,
                      "end": 2.64,
                      "probability": 0.147705078125
                    },
                    {
                      "word": "一",
                      "start": 2.64,
                      "end": 2.82,
                      "probability": 0.5234375
                    },
                    {
                      "word": "千",
                      "start": 2.82,
                      "end": 3.04,
                      "probability": 0.215576171875
                    },
                    {
                      "word": "万",
                      "start": 3.04,
                      "end": 3.2,
                      "probability": 0.82421875
                    },
                    {
                      "word": "或者",
                      "start": 3.2,
                      "end": 3.64,
                      "probability": 0.98583984375
                    },
                    {
                      "word": "任",
                      "start": 3.64,
                      "end": 3.84,
                      "probability": 0.99951171875
                    },
                    {
                      "word": "何",
                      "start": 3.84,
                      "end": 4.06,
                      "probability": 0.99951171875
                    },
                    {
                      "word": "兼",
                      "start": 4.06,
                      "end": 4.24,
                      "probability": 0.9072265625
                    },
                    {
                      "word": "容",
                      "start": 4.24,
                      "end": 4.42,
                      "probability": 1.0
                    },
                    {
                      "word": "Open",
                      "start": 4.42,
                      "end": 4.74,
                      "probability": 0.91064453125
                    },
                    {
                      "word": "AI",
                      "start": 4.74,
                      "end": 5.04,
                      "probability": 0.9931640625
                    },
                    {
                      "word": "的",
                      "start": 5.04,
                      "end": 5.26,
                      "probability": 0.98779296875
                    },
                    {
                      "word": "接",
                      "start": 5.26,
                      "end": 5.4,
                      "probability": 0.87744140625
                    },
                    {
                      "word": "口",
                      "start": 5.4,
                      "end": 5.56,
                      "probability": 0.99951171875
                    },
                    {
                      "word": "都",
                      "start": 5.56,
                      "end": 5.68,
                      "probability": 0.99658203125
                    },
                    {
                      "word": "能",
                      "start": 5.68,
                      "end": 5.84,
                      "probability": 0.9990234375
                    },
                    {
                      "word": "用",
                      "start": 5.84,
                      "end": 6.08,
                      "probability": 0.98828125
                    }
                  ]
                },
                {
                  "id": 2,
                  "seek": 0,
                  "start": 6.600000000000002,
                  "end": 8.14,
                  "text": "按你自己的用量去做",
                  "tokens": [
                    50674,
                    26613,
                    2166,
                    17645,
                    1546,
                    9254,
                    26748,
                    6734,
                    10907,
                    50773
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.34069514918971705,
                  "compression_ratio": 0.8932038834951457,
                  "no_speech_prob": 0.852008044719696,
                  "words": [
                    {
                      "word": "按",
                      "start": 6.600000000000002,
                      "end": 7.04,
                      "probability": 0.99072265625
                    },
                    {
                      "word": "你",
                      "start": 7.04,
                      "end": 7.24,
                      "probability": 0.98974609375
                    },
                    {
                      "word": "自己",
                      "start": 7.24,
                      "end": 7.5,
                      "probability": 0.99853515625
                    },
                    {
                      "word": "的",
                      "start": 7.5,
                      "end": 7.74,
                      "probability": 0.9970703125
                    },
                    {
                      "word": "用",
                      "start": 7.74,
                      "end": 7.82,
                      "probability": 0.96044921875
                    },
                    {
                      "word": "量",
                      "start": 7.82,
                      "end": 8.04,
                      "probability": 0.98046875
                    },
                    {
                      "word": "去",
                      "start": 8.04,
                      "end": 8.08,
                      "probability": 0.7265625
                    },
                    {
                      "word": "做",
                      "start": 8.08,
                      "end": 8.14,
                      "probability": 0.08990478515625
                    }
                  ]
                }
              ],
              "language": "zh",
              "duration": 8.16
            },
            "status": "COMPLETE",
            "requested_language": "zh"
          },
          "suppression_proposed": null
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "58",
          "before": "同意签问",
          "after": "建容",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "建容疑似兼容误识",
          "context_evidence": "上下文为API接口，'兼容'更符合逻辑",
          "independent_ASR": {
            "engine": "whisper-large-v3",
            "model_path": "/Users/sagawa/AI_WORK/cache/huggingface/hub/models--mlx-community--whisper-large-v3",
            "language": "zh",
            "audio_interval": [
              118.08,
              126.24000000000001
            ],
            "target_interval": [
              120.08,
              124.24000000000001
            ],
            "text": "同一千万或者任何兼容OpenAI的接口都能用",
            "words": [
              {
                "word": "同",
                "start": 119.96,
                "end": 120.72,
                "probability": 0.147705078125
              },
              {
                "word": "一",
                "start": 120.72,
                "end": 120.89999999999999,
                "probability": 0.5234375
              },
              {
                "word": "千",
                "start": 120.89999999999999,
                "end": 121.12,
                "probability": 0.215576171875
              },
              {
                "word": "万",
                "start": 121.12,
                "end": 121.28,
                "probability": 0.82421875
              },
              {
                "word": "或者",
                "start": 121.28,
                "end": 121.72,
                "probability": 0.98583984375
              },
              {
                "word": "任",
                "start": 121.72,
                "end": 121.92,
                "probability": 0.99951171875
              },
              {
                "word": "何",
                "start": 121.92,
                "end": 122.14,
                "probability": 0.99951171875
              },
              {
                "word": "兼",
                "start": 122.14,
                "end": 122.32,
                "probability": 0.9072265625
              },
              {
                "word": "容",
                "start": 122.32,
                "end": 122.5,
                "probability": 1.0
              },
              {
                "word": "Open",
                "start": 122.5,
                "end": 122.82,
                "probability": 0.91064453125
              },
              {
                "word": "AI",
                "start": 122.82,
                "end": 123.12,
                "probability": 0.9931640625
              },
              {
                "word": "的",
                "start": 123.12,
                "end": 123.34,
                "probability": 0.98779296875
              },
              {
                "word": "接",
                "start": 123.34,
                "end": 123.48,
                "probability": 0.87744140625
              },
              {
                "word": "口",
                "start": 123.48,
                "end": 123.64,
                "probability": 0.99951171875
              },
              {
                "word": "都",
                "start": 123.64,
                "end": 123.75999999999999,
                "probability": 0.99658203125
              },
              {
                "word": "能",
                "start": 123.75999999999999,
                "end": 123.92,
                "probability": 0.9990234375
              },
              {
                "word": "用",
                "start": 123.92,
                "end": 124.16,
                "probability": 0.98828125
              }
            ],
            "raw_result": {
              "text": "像Dipstick同一千万或者任何兼容OpenAI的接口都能用按你自己的用量去做",
              "segments": [
                {
                  "id": 0,
                  "seek": 0,
                  "start": 0.8200000000000023,
                  "end": 1.88,
                  "text": "像Dipstick",
                  "tokens": [
                    50365,
                    12760,
                    35,
                    647,
                    11881,
                    50467
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.34069514918971705,
                  "compression_ratio": 0.8932038834951457,
                  "no_speech_prob": 0.852008044719696,
                  "words": [
                    {
                      "word": "像",
                      "start": 0.8200000000000023,
                      "end": 1.26,
                      "probability": 0.1815185546875
                    },
                    {
                      "word": "D",
                      "start": 1.26,
                      "end": 1.5,
                      "probability": 0.576171875
                    },
                    {
                      "word": "ip",
                      "start": 1.5,
                      "end": 1.62,
                      "probability": 0.9765625
                    },
                    {
                      "word": "stick",
                      "start": 1.62,
                      "end": 1.88,
                      "probability": 0.837890625
                    }
                  ]
                },
                {
                  "id": 1,
                  "seek": 0,
                  "start": 1.88,
                  "end": 6.08,
                  "text": "同一千万或者任何兼容OpenAI的接口都能用",
                  "tokens": [
                    50467,
                    13089,
                    2257,
                    20787,
                    23570,
                    31148,
                    26443,
                    11906,
                    2347,
                    120,
                    25750,
                    45569,
                    48698,
                    1546,
                    14468,
                    18144,
                    7182,
                    8225,
                    9254,
                    50674
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.34069514918971705,
                  "compression_ratio": 0.8932038834951457,
                  "no_speech_prob": 0.852008044719696,
                  "words": [
                    {
                      "word": "同",
                      "start": 1.88,
                      "end": 2.64,
                      "probability": 0.147705078125
                    },
                    {
                      "word": "一",
                      "start": 2.64,
                      "end": 2.82,
                      "probability": 0.5234375
                    },
                    {
                      "word": "千",
                      "start": 2.82,
                      "end": 3.04,
                      "probability": 0.215576171875
                    },
                    {
                      "word": "万",
                      "start": 3.04,
                      "end": 3.2,
                      "probability": 0.82421875
                    },
                    {
                      "word": "或者",
                      "start": 3.2,
                      "end": 3.64,
                      "probability": 0.98583984375
                    },
                    {
                      "word": "任",
                      "start": 3.64,
                      "end": 3.84,
                      "probability": 0.99951171875
                    },
                    {
                      "word": "何",
                      "start": 3.84,
                      "end": 4.06,
                      "probability": 0.99951171875
                    },
                    {
                      "word": "兼",
                      "start": 4.06,
                      "end": 4.24,
                      "probability": 0.9072265625
                    },
                    {
                      "word": "容",
                      "start": 4.24,
                      "end": 4.42,
                      "probability": 1.0
                    },
                    {
                      "word": "Open",
                      "start": 4.42,
                      "end": 4.74,
                      "probability": 0.91064453125
                    },
                    {
                      "word": "AI",
                      "start": 4.74,
                      "end": 5.04,
                      "probability": 0.9931640625
                    },
                    {
                      "word": "的",
                      "start": 5.04,
                      "end": 5.26,
                      "probability": 0.98779296875
                    },
                    {
                      "word": "接",
                      "start": 5.26,
                      "end": 5.4,
                      "probability": 0.87744140625
                    },
                    {
                      "word": "口",
                      "start": 5.4,
                      "end": 5.56,
                      "probability": 0.99951171875
                    },
                    {
                      "word": "都",
                      "start": 5.56,
                      "end": 5.68,
                      "probability": 0.99658203125
                    },
                    {
                      "word": "能",
                      "start": 5.68,
                      "end": 5.84,
                      "probability": 0.9990234375
                    },
                    {
                      "word": "用",
                      "start": 5.84,
                      "end": 6.08,
                      "probability": 0.98828125
                    }
                  ]
                },
                {
                  "id": 2,
                  "seek": 0,
                  "start": 6.600000000000002,
                  "end": 8.14,
                  "text": "按你自己的用量去做",
                  "tokens": [
                    50674,
                    26613,
                    2166,
                    17645,
                    1546,
                    9254,
                    26748,
                    6734,
                    10907,
                    50773
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.34069514918971705,
                  "compression_ratio": 0.8932038834951457,
                  "no_speech_prob": 0.852008044719696,
                  "words": [
                    {
                      "word": "按",
                      "start": 6.600000000000002,
                      "end": 7.04,
                      "probability": 0.99072265625
                    },
                    {
                      "word": "你",
                      "start": 7.04,
                      "end": 7.24,
                      "probability": 0.98974609375
                    },
                    {
                      "word": "自己",
                      "start": 7.24,
                      "end": 7.5,
                      "probability": 0.99853515625
                    },
                    {
                      "word": "的",
                      "start": 7.5,
                      "end": 7.74,
                      "probability": 0.9970703125
                    },
                    {
                      "word": "用",
                      "start": 7.74,
                      "end": 7.82,
                      "probability": 0.96044921875
                    },
                    {
                      "word": "量",
                      "start": 7.82,
                      "end": 8.04,
                      "probability": 0.98046875
                    },
                    {
                      "word": "去",
                      "start": 8.04,
                      "end": 8.08,
                      "probability": 0.7265625
                    },
                    {
                      "word": "做",
                      "start": 8.08,
                      "end": 8.14,
                      "probability": 0.08990478515625
                    }
                  ]
                }
              ],
              "language": "zh",
              "duration": 8.16
            },
            "status": "COMPLETE",
            "requested_language": "zh"
          },
          "suppression_proposed": null
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "58",
          "before": "建容",
          "after": "建容",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "Nonsense: '建容' vs '兼容' (compatible)",
          "context_evidence": "Context: '兼容OpenAI的接口'",
          "independent_ASR": {
            "engine": "whisper-large-v3",
            "model_path": "/Users/sagawa/AI_WORK/cache/huggingface/hub/models--mlx-community--whisper-large-v3",
            "language": "zh",
            "audio_interval": [
              118.08,
              126.24000000000001
            ],
            "target_interval": [
              120.08,
              124.24000000000001
            ],
            "text": "同一千万或者任何兼容OpenAI的接口都能用",
            "words": [
              {
                "word": "同",
                "start": 119.96,
                "end": 120.72,
                "probability": 0.147705078125
              },
              {
                "word": "一",
                "start": 120.72,
                "end": 120.89999999999999,
                "probability": 0.5234375
              },
              {
                "word": "千",
                "start": 120.89999999999999,
                "end": 121.12,
                "probability": 0.215576171875
              },
              {
                "word": "万",
                "start": 121.12,
                "end": 121.28,
                "probability": 0.82421875
              },
              {
                "word": "或者",
                "start": 121.28,
                "end": 121.72,
                "probability": 0.98583984375
              },
              {
                "word": "任",
                "start": 121.72,
                "end": 121.92,
                "probability": 0.99951171875
              },
              {
                "word": "何",
                "start": 121.92,
                "end": 122.14,
                "probability": 0.99951171875
              },
              {
                "word": "兼",
                "start": 122.14,
                "end": 122.32,
                "probability": 0.9072265625
              },
              {
                "word": "容",
                "start": 122.32,
                "end": 122.5,
                "probability": 1.0
              },
              {
                "word": "Open",
                "start": 122.5,
                "end": 122.82,
                "probability": 0.91064453125
              },
              {
                "word": "AI",
                "start": 122.82,
                "end": 123.12,
                "probability": 0.9931640625
              },
              {
                "word": "的",
                "start": 123.12,
                "end": 123.34,
                "probability": 0.98779296875
              },
              {
                "word": "接",
                "start": 123.34,
                "end": 123.48,
                "probability": 0.87744140625
              },
              {
                "word": "口",
                "start": 123.48,
                "end": 123.64,
                "probability": 0.99951171875
              },
              {
                "word": "都",
                "start": 123.64,
                "end": 123.75999999999999,
                "probability": 0.99658203125
              },
              {
                "word": "能",
                "start": 123.75999999999999,
                "end": 123.92,
                "probability": 0.9990234375
              },
              {
                "word": "用",
                "start": 123.92,
                "end": 124.16,
                "probability": 0.98828125
              }
            ],
            "raw_result": {
              "text": "像Dipstick同一千万或者任何兼容OpenAI的接口都能用按你自己的用量去做",
              "segments": [
                {
                  "id": 0,
                  "seek": 0,
                  "start": 0.8200000000000023,
                  "end": 1.88,
                  "text": "像Dipstick",
                  "tokens": [
                    50365,
                    12760,
                    35,
                    647,
                    11881,
                    50467
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.34069514918971705,
                  "compression_ratio": 0.8932038834951457,
                  "no_speech_prob": 0.852008044719696,
                  "words": [
                    {
                      "word": "像",
                      "start": 0.8200000000000023,
                      "end": 1.26,
                      "probability": 0.1815185546875
                    },
                    {
                      "word": "D",
                      "start": 1.26,
                      "end": 1.5,
                      "probability": 0.576171875
                    },
                    {
                      "word": "ip",
                      "start": 1.5,
                      "end": 1.62,
                      "probability": 0.9765625
                    },
                    {
                      "word": "stick",
                      "start": 1.62,
                      "end": 1.88,
                      "probability": 0.837890625
                    }
                  ]
                },
                {
                  "id": 1,
                  "seek": 0,
                  "start": 1.88,
                  "end": 6.08,
                  "text": "同一千万或者任何兼容OpenAI的接口都能用",
                  "tokens": [
                    50467,
                    13089,
                    2257,
                    20787,
                    23570,
                    31148,
                    26443,
                    11906,
                    2347,
                    120,
                    25750,
                    45569,
                    48698,
                    1546,
                    14468,
                    18144,
                    7182,
                    8225,
                    9254,
                    50674
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.34069514918971705,
                  "compression_ratio": 0.8932038834951457,
                  "no_speech_prob": 0.852008044719696,
                  "words": [
                    {
                      "word": "同",
                      "start": 1.88,
                      "end": 2.64,
                      "probability": 0.147705078125
                    },
                    {
                      "word": "一",
                      "start": 2.64,
                      "end": 2.82,
                      "probability": 0.5234375
                    },
                    {
                      "word": "千",
                      "start": 2.82,
                      "end": 3.04,
                      "probability": 0.215576171875
                    },
                    {
                      "word": "万",
                      "start": 3.04,
                      "end": 3.2,
                      "probability": 0.82421875
                    },
                    {
                      "word": "或者",
                      "start": 3.2,
                      "end": 3.64,
                      "probability": 0.98583984375
                    },
                    {
                      "word": "任",
                      "start": 3.64,
                      "end": 3.84,
                      "probability": 0.99951171875
                    },
                    {
                      "word": "何",
                      "start": 3.84,
                      "end": 4.06,
                      "probability": 0.99951171875
                    },
                    {
                      "word": "兼",
                      "start": 4.06,
                      "end": 4.24,
                      "probability": 0.9072265625
                    },
                    {
                      "word": "容",
                      "start": 4.24,
                      "end": 4.42,
                      "probability": 1.0
                    },
                    {
                      "word": "Open",
                      "start": 4.42,
                      "end": 4.74,
                      "probability": 0.91064453125
                    },
                    {
                      "word": "AI",
                      "start": 4.74,
                      "end": 5.04,
                      "probability": 0.9931640625
                    },
                    {
                      "word": "的",
                      "start": 5.04,
                      "end": 5.26,
                      "probability": 0.98779296875
                    },
                    {
                      "word": "接",
                      "start": 5.26,
                      "end": 5.4,
                      "probability": 0.87744140625
                    },
                    {
                      "word": "口",
                      "start": 5.4,
                      "end": 5.56,
                      "probability": 0.99951171875
                    },
                    {
                      "word": "都",
                      "start": 5.56,
                      "end": 5.68,
                      "probability": 0.99658203125
                    },
                    {
                      "word": "能",
                      "start": 5.68,
                      "end": 5.84,
                      "probability": 0.9990234375
                    },
                    {
                      "word": "用",
                      "start": 5.84,
                      "end": 6.08,
                      "probability": 0.98828125
                    }
                  ]
                },
                {
                  "id": 2,
                  "seek": 0,
                  "start": 6.600000000000002,
                  "end": 8.14,
                  "text": "按你自己的用量去做",
                  "tokens": [
                    50674,
                    26613,
                    2166,
                    17645,
                    1546,
                    9254,
                    26748,
                    6734,
                    10907,
                    50773
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.34069514918971705,
                  "compression_ratio": 0.8932038834951457,
                  "no_speech_prob": 0.852008044719696,
                  "words": [
                    {
                      "word": "按",
                      "start": 6.600000000000002,
                      "end": 7.04,
                      "probability": 0.99072265625
                    },
                    {
                      "word": "你",
                      "start": 7.04,
                      "end": 7.24,
                      "probability": 0.98974609375
                    },
                    {
                      "word": "自己",
                      "start": 7.24,
                      "end": 7.5,
                      "probability": 0.99853515625
                    },
                    {
                      "word": "的",
                      "start": 7.5,
                      "end": 7.74,
                      "probability": 0.9970703125
                    },
                    {
                      "word": "用",
                      "start": 7.74,
                      "end": 7.82,
                      "probability": 0.96044921875
                    },
                    {
                      "word": "量",
                      "start": 7.82,
                      "end": 8.04,
                      "probability": 0.98046875
                    },
                    {
                      "word": "去",
                      "start": 8.04,
                      "end": 8.08,
                      "probability": 0.7265625
                    },
                    {
                      "word": "做",
                      "start": 8.08,
                      "end": 8.14,
                      "probability": 0.08990478515625
                    }
                  ]
                }
              ],
              "language": "zh",
              "duration": 8.16
            },
            "status": "COMPLETE",
            "requested_language": "zh"
          },
          "suppression_proposed": null
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "58",
          "before": "建容",
          "after": "建容",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "Nonsense: '建容' vs '兼容' (compatible)",
          "context_evidence": "Context: '兼容OpenAI的接口'",
          "independent_ASR": {
            "engine": "whisper-large-v3",
            "model_path": "/Users/sagawa/AI_WORK/cache/huggingface/hub/models--mlx-community--whisper-large-v3",
            "language": "zh",
            "audio_interval": [
              118.08,
              126.24000000000001
            ],
            "target_interval": [
              120.08,
              124.24000000000001
            ],
            "text": "同一千万或者任何兼容OpenAI的接口都能用",
            "words": [
              {
                "word": "同",
                "start": 119.96,
                "end": 120.72,
                "probability": 0.147705078125
              },
              {
                "word": "一",
                "start": 120.72,
                "end": 120.89999999999999,
                "probability": 0.5234375
              },
              {
                "word": "千",
                "start": 120.89999999999999,
                "end": 121.12,
                "probability": 0.215576171875
              },
              {
                "word": "万",
                "start": 121.12,
                "end": 121.28,
                "probability": 0.82421875
              },
              {
                "word": "或者",
                "start": 121.28,
                "end": 121.72,
                "probability": 0.98583984375
              },
              {
                "word": "任",
                "start": 121.72,
                "end": 121.92,
                "probability": 0.99951171875
              },
              {
                "word": "何",
                "start": 121.92,
                "end": 122.14,
                "probability": 0.99951171875
              },
              {
                "word": "兼",
                "start": 122.14,
                "end": 122.32,
                "probability": 0.9072265625
              },
              {
                "word": "容",
                "start": 122.32,
                "end": 122.5,
                "probability": 1.0
              },
              {
                "word": "Open",
                "start": 122.5,
                "end": 122.82,
                "probability": 0.91064453125
              },
              {
                "word": "AI",
                "start": 122.82,
                "end": 123.12,
                "probability": 0.9931640625
              },
              {
                "word": "的",
                "start": 123.12,
                "end": 123.34,
                "probability": 0.98779296875
              },
              {
                "word": "接",
                "start": 123.34,
                "end": 123.48,
                "probability": 0.87744140625
              },
              {
                "word": "口",
                "start": 123.48,
                "end": 123.64,
                "probability": 0.99951171875
              },
              {
                "word": "都",
                "start": 123.64,
                "end": 123.75999999999999,
                "probability": 0.99658203125
              },
              {
                "word": "能",
                "start": 123.75999999999999,
                "end": 123.92,
                "probability": 0.9990234375
              },
              {
                "word": "用",
                "start": 123.92,
                "end": 124.16,
                "probability": 0.98828125
              }
            ],
            "raw_result": {
              "text": "像Dipstick同一千万或者任何兼容OpenAI的接口都能用按你自己的用量去做",
              "segments": [
                {
                  "id": 0,
                  "seek": 0,
                  "start": 0.8200000000000023,
                  "end": 1.88,
                  "text": "像Dipstick",
                  "tokens": [
                    50365,
                    12760,
                    35,
                    647,
                    11881,
                    50467
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.34069514918971705,
                  "compression_ratio": 0.8932038834951457,
                  "no_speech_prob": 0.852008044719696,
                  "words": [
                    {
                      "word": "像",
                      "start": 0.8200000000000023,
                      "end": 1.26,
                      "probability": 0.1815185546875
                    },
                    {
                      "word": "D",
                      "start": 1.26,
                      "end": 1.5,
                      "probability": 0.576171875
                    },
                    {
                      "word": "ip",
                      "start": 1.5,
                      "end": 1.62,
                      "probability": 0.9765625
                    },
                    {
                      "word": "stick",
                      "start": 1.62,
                      "end": 1.88,
                      "probability": 0.837890625
                    }
                  ]
                },
                {
                  "id": 1,
                  "seek": 0,
                  "start": 1.88,
                  "end": 6.08,
                  "text": "同一千万或者任何兼容OpenAI的接口都能用",
                  "tokens": [
                    50467,
                    13089,
                    2257,
                    20787,
                    23570,
                    31148,
                    26443,
                    11906,
                    2347,
                    120,
                    25750,
                    45569,
                    48698,
                    1546,
                    14468,
                    18144,
                    7182,
                    8225,
                    9254,
                    50674
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.34069514918971705,
                  "compression_ratio": 0.8932038834951457,
                  "no_speech_prob": 0.852008044719696,
                  "words": [
                    {
                      "word": "同",
                      "start": 1.88,
                      "end": 2.64,
                      "probability": 0.147705078125
                    },
                    {
                      "word": "一",
                      "start": 2.64,
                      "end": 2.82,
                      "probability": 0.5234375
                    },
                    {
                      "word": "千",
                      "start": 2.82,
                      "end": 3.04,
                      "probability": 0.215576171875
                    },
                    {
                      "word": "万",
                      "start": 3.04,
                      "end": 3.2,
                      "probability": 0.82421875
                    },
                    {
                      "word": "或者",
                      "start": 3.2,
                      "end": 3.64,
                      "probability": 0.98583984375
                    },
                    {
                      "word": "任",
                      "start": 3.64,
                      "end": 3.84,
                      "probability": 0.99951171875
                    },
                    {
                      "word": "何",
                      "start": 3.84,
                      "end": 4.06,
                      "probability": 0.99951171875
                    },
                    {
                      "word": "兼",
                      "start": 4.06,
                      "end": 4.24,
                      "probability": 0.9072265625
                    },
                    {
                      "word": "容",
                      "start": 4.24,
                      "end": 4.42,
                      "probability": 1.0
                    },
                    {
                      "word": "Open",
                      "start": 4.42,
                      "end": 4.74,
                      "probability": 0.91064453125
                    },
                    {
                      "word": "AI",
                      "start": 4.74,
                      "end": 5.04,
                      "probability": 0.9931640625
                    },
                    {
                      "word": "的",
                      "start": 5.04,
                      "end": 5.26,
                      "probability": 0.98779296875
                    },
                    {
                      "word": "接",
                      "start": 5.26,
                      "end": 5.4,
                      "probability": 0.87744140625
                    },
                    {
                      "word": "口",
                      "start": 5.4,
                      "end": 5.56,
                      "probability": 0.99951171875
                    },
                    {
                      "word": "都",
                      "start": 5.56,
                      "end": 5.68,
                      "probability": 0.99658203125
                    },
                    {
                      "word": "能",
                      "start": 5.68,
                      "end": 5.84,
                      "probability": 0.9990234375
                    },
                    {
                      "word": "用",
                      "start": 5.84,
                      "end": 6.08,
                      "probability": 0.98828125
                    }
                  ]
                },
                {
                  "id": 2,
                  "seek": 0,
                  "start": 6.600000000000002,
                  "end": 8.14,
                  "text": "按你自己的用量去做",
                  "tokens": [
                    50674,
                    26613,
                    2166,
                    17645,
                    1546,
                    9254,
                    26748,
                    6734,
                    10907,
                    50773
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.34069514918971705,
                  "compression_ratio": 0.8932038834951457,
                  "no_speech_prob": 0.852008044719696,
                  "words": [
                    {
                      "word": "按",
                      "start": 6.600000000000002,
                      "end": 7.04,
                      "probability": 0.99072265625
                    },
                    {
                      "word": "你",
                      "start": 7.04,
                      "end": 7.24,
                      "probability": 0.98974609375
                    },
                    {
                      "word": "自己",
                      "start": 7.24,
                      "end": 7.5,
                      "probability": 0.99853515625
                    },
                    {
                      "word": "的",
                      "start": 7.5,
                      "end": 7.74,
                      "probability": 0.9970703125
                    },
                    {
                      "word": "用",
                      "start": 7.74,
                      "end": 7.82,
                      "probability": 0.96044921875
                    },
                    {
                      "word": "量",
                      "start": 7.82,
                      "end": 8.04,
                      "probability": 0.98046875
                    },
                    {
                      "word": "去",
                      "start": 8.04,
                      "end": 8.08,
                      "probability": 0.7265625
                    },
                    {
                      "word": "做",
                      "start": 8.08,
                      "end": 8.14,
                      "probability": 0.08990478515625
                    }
                  ]
                }
              ],
              "language": "zh",
              "duration": 8.16
            },
            "status": "COMPLETE",
            "requested_language": "zh"
          },
          "suppression_proposed": null
        }
      ],
      "language": "zh",
      "packet_ids": [
        4
      ],
      "classification": "BEST_SUPPORTED_WITH_REVIEW",
      "independent_ASR": {
        "engine": "whisper-large-v3",
        "model_path": "/Users/sagawa/AI_WORK/cache/huggingface/hub/models--mlx-community--whisper-large-v3",
        "language": "zh",
        "audio_interval": [
          118.08,
          126.24000000000001
        ],
        "target_interval": [
          120.08,
          124.24000000000001
        ],
        "text": "同一千万或者任何兼容OpenAI的接口都能用",
        "words": [
          {
            "word": "同",
            "start": 119.96,
            "end": 120.72,
            "probability": 0.147705078125
          },
          {
            "word": "一",
            "start": 120.72,
            "end": 120.89999999999999,
            "probability": 0.5234375
          },
          {
            "word": "千",
            "start": 120.89999999999999,
            "end": 121.12,
            "probability": 0.215576171875
          },
          {
            "word": "万",
            "start": 121.12,
            "end": 121.28,
            "probability": 0.82421875
          },
          {
            "word": "或者",
            "start": 121.28,
            "end": 121.72,
            "probability": 0.98583984375
          },
          {
            "word": "任",
            "start": 121.72,
            "end": 121.92,
            "probability": 0.99951171875
          },
          {
            "word": "何",
            "start": 121.92,
            "end": 122.14,
            "probability": 0.99951171875
          },
          {
            "word": "兼",
            "start": 122.14,
            "end": 122.32,
            "probability": 0.9072265625
          },
          {
            "word": "容",
            "start": 122.32,
            "end": 122.5,
            "probability": 1.0
          },
          {
            "word": "Open",
            "start": 122.5,
            "end": 122.82,
            "probability": 0.91064453125
          },
          {
            "word": "AI",
            "start": 122.82,
            "end": 123.12,
            "probability": 0.9931640625
          },
          {
            "word": "的",
            "start": 123.12,
            "end": 123.34,
            "probability": 0.98779296875
          },
          {
            "word": "接",
            "start": 123.34,
            "end": 123.48,
            "probability": 0.87744140625
          },
          {
            "word": "口",
            "start": 123.48,
            "end": 123.64,
            "probability": 0.99951171875
          },
          {
            "word": "都",
            "start": 123.64,
            "end": 123.75999999999999,
            "probability": 0.99658203125
          },
          {
            "word": "能",
            "start": 123.75999999999999,
            "end": 123.92,
            "probability": 0.9990234375
          },
          {
            "word": "用",
            "start": 123.92,
            "end": 124.16,
            "probability": 0.98828125
          }
        ],
        "raw_result": {
          "text": "像Dipstick同一千万或者任何兼容OpenAI的接口都能用按你自己的用量去做",
          "segments": [
            {
              "id": 0,
              "seek": 0,
              "start": 0.8200000000000023,
              "end": 1.88,
              "text": "像Dipstick",
              "tokens": [
                50365,
                12760,
                35,
                647,
                11881,
                50467
              ],
              "temperature": 0.0,
              "avg_logprob": -0.34069514918971705,
              "compression_ratio": 0.8932038834951457,
              "no_speech_prob": 0.852008044719696,
              "words": [
                {
                  "word": "像",
                  "start": 0.8200000000000023,
                  "end": 1.26,
                  "probability": 0.1815185546875
                },
                {
                  "word": "D",
                  "start": 1.26,
                  "end": 1.5,
                  "probability": 0.576171875
                },
                {
                  "word": "ip",
                  "start": 1.5,
                  "end": 1.62,
                  "probability": 0.9765625
                },
                {
                  "word": "stick",
                  "start": 1.62,
                  "end": 1.88,
                  "probability": 0.837890625
                }
              ]
            },
            {
              "id": 1,
              "seek": 0,
              "start": 1.88,
              "end": 6.08,
              "text": "同一千万或者任何兼容OpenAI的接口都能用",
              "tokens": [
                50467,
                13089,
                2257,
                20787,
                23570,
                31148,
                26443,
                11906,
                2347,
                120,
                25750,
                45569,
                48698,
                1546,
                14468,
                18144,
                7182,
                8225,
                9254,
                50674
              ],
              "temperature": 0.0,
              "avg_logprob": -0.34069514918971705,
              "compression_ratio": 0.8932038834951457,
              "no_speech_prob": 0.852008044719696,
              "words": [
                {
                  "word": "同",
                  "start": 1.88,
                  "end": 2.64,
                  "probability": 0.147705078125
                },
                {
                  "word": "一",
                  "start": 2.64,
                  "end": 2.82,
                  "probability": 0.5234375
                },
                {
                  "word": "千",
                  "start": 2.82,
                  "end": 3.04,
                  "probability": 0.215576171875
                },
                {
                  "word": "万",
                  "start": 3.04,
                  "end": 3.2,
                  "probability": 0.82421875
                },
                {
                  "word": "或者",
                  "start": 3.2,
                  "end": 3.64,
                  "probability": 0.98583984375
                },
                {
                  "word": "任",
                  "start": 3.64,
                  "end": 3.84,
                  "probability": 0.99951171875
                },
                {
                  "word": "何",
                  "start": 3.84,
                  "end": 4.06,
                  "probability": 0.99951171875
                },
                {
                  "word": "兼",
                  "start": 4.06,
                  "end": 4.24,
                  "probability": 0.9072265625
                },
                {
                  "word": "容",
                  "start": 4.24,
                  "end": 4.42,
                  "probability": 1.0
                },
                {
                  "word": "Open",
                  "start": 4.42,
                  "end": 4.74,
                  "probability": 0.91064453125
                },
                {
                  "word": "AI",
                  "start": 4.74,
                  "end": 5.04,
                  "probability": 0.9931640625
                },
                {
                  "word": "的",
                  "start": 5.04,
                  "end": 5.26,
                  "probability": 0.98779296875
                },
                {
                  "word": "接",
                  "start": 5.26,
                  "end": 5.4,
                  "probability": 0.87744140625
                },
                {
                  "word": "口",
                  "start": 5.4,
                  "end": 5.56,
                  "probability": 0.99951171875
                },
                {
                  "word": "都",
                  "start": 5.56,
                  "end": 5.68,
                  "probability": 0.99658203125
                },
                {
                  "word": "能",
                  "start": 5.68,
                  "end": 5.84,
                  "probability": 0.9990234375
                },
                {
                  "word": "用",
                  "start": 5.84,
                  "end": 6.08,
                  "probability": 0.98828125
                }
              ]
            },
            {
              "id": 2,
              "seek": 0,
              "start": 6.600000000000002,
              "end": 8.14,
              "text": "按你自己的用量去做",
              "tokens": [
                50674,
                26613,
                2166,
                17645,
                1546,
                9254,
                26748,
                6734,
                10907,
                50773
              ],
              "temperature": 0.0,
              "avg_logprob": -0.34069514918971705,
              "compression_ratio": 0.8932038834951457,
              "no_speech_prob": 0.852008044719696,
              "words": [
                {
                  "word": "按",
                  "start": 6.600000000000002,
                  "end": 7.04,
                  "probability": 0.99072265625
                },
                {
                  "word": "你",
                  "start": 7.04,
                  "end": 7.24,
                  "probability": 0.98974609375
                },
                {
                  "word": "自己",
                  "start": 7.24,
                  "end": 7.5,
                  "probability": 0.99853515625
                },
                {
                  "word": "的",
                  "start": 7.5,
                  "end": 7.74,
                  "probability": 0.9970703125
                },
                {
                  "word": "用",
                  "start": 7.74,
                  "end": 7.82,
                  "probability": 0.96044921875
                },
                {
                  "word": "量",
                  "start": 7.82,
                  "end": 8.04,
                  "probability": 0.98046875
                },
                {
                  "word": "去",
                  "start": 8.04,
                  "end": 8.08,
                  "probability": 0.7265625
                },
                {
                  "word": "做",
                  "start": 8.08,
                  "end": 8.14,
                  "probability": 0.08990478515625
                }
              ]
            }
          ],
          "language": "zh",
          "duration": 8.16
        },
        "status": "COMPLETE",
        "requested_language": "zh"
      },
      "raw_word_timestamps": [
        {
          "word": "同",
          "start": 120.08,
          "end": 120.80000000000001,
          "probability": 0.830078125
        },
        {
          "word": "意",
          "start": 120.80000000000001,
          "end": 120.98,
          "probability": 0.82275390625
        },
        {
          "word": "签",
          "start": 120.98,
          "end": 121.16,
          "probability": 0.96533203125
        },
        {
          "word": "问",
          "start": 121.16,
          "end": 121.38,
          "probability": 0.80615234375
        },
        {
          "word": "或者",
          "start": 121.38,
          "end": 121.80000000000001,
          "probability": 0.99658203125
        },
        {
          "word": "任",
          "start": 121.80000000000001,
          "end": 122.0,
          "probability": 0.99951171875
        },
        {
          "word": "何",
          "start": 122.0,
          "end": 122.18,
          "probability": 1.0
        },
        {
          "word": "建",
          "start": 122.18,
          "end": 122.36,
          "probability": 0.1092529296875
        },
        {
          "word": "容",
          "start": 122.36,
          "end": 122.56,
          "probability": 0.6923828125
        },
        {
          "word": "Open",
          "start": 122.56,
          "end": 122.94,
          "probability": 0.96533203125
        },
        {
          "word": "AI",
          "start": 122.94,
          "end": 123.26,
          "probability": 0.9931640625
        },
        {
          "word": "的",
          "start": 123.26,
          "end": 123.4,
          "probability": 0.998046875
        },
        {
          "word": "接",
          "start": 123.4,
          "end": 123.52000000000001,
          "probability": 0.8330078125
        },
        {
          "word": "口",
          "start": 123.52000000000001,
          "end": 123.68,
          "probability": 0.9990234375
        },
        {
          "word": "都",
          "start": 123.68,
          "end": 123.84,
          "probability": 0.99951171875
        },
        {
          "word": "能",
          "start": 123.84,
          "end": 123.98,
          "probability": 0.99951171875
        },
        {
          "word": "用",
          "start": 123.98,
          "end": 124.24000000000001,
          "probability": 0.99951171875
        }
      ],
      "word_timestamps": [],
      "source_text": "建容",
      "review_status": "BEST_SUPPORTED_WITH_REVIEW",
      "canonical_timing_evidence": "CUE_ANCHORED",
      "index": 58
    },
    {
      "start": 124.84,
      "end": 126.58,
      "text": "计备",
      "raw_asr_text": "按你自己的用量计备",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 4,
      "words": [],
      "asr_metrics": {
        "avg_logprob": -0.1897983819666043,
        "compression_ratio": 1.2267657992565055,
        "no_speech_prob": 6.547795639022524e-11
      },
      "lid_status": "UNCERTAIN",
      "alternatives": [
        {
          "language": "zh",
          "text": "甚至直接点一下让大模型帮你自动生成术语翻译时它还会自动结合上文语境不会翻得前言不搭后语翻译接入的是你自己的大模型API像Deepstick同意签问或者任何建容OpenAI的接口都能用按你自己的用量计备啊会自动释出关闭模型思考过程的参数来降低延迟到底"
        }
      ],
      "id": 59,
      "review_annotation": "待確認",
      "row_id": "59",
      "bounded_row_id": 59,
      "parent_row_ids": [
        59
      ],
      "parent_row_id": 59,
      "packet_id": 4,
      "LID_evidence": [
        {
          "region_id": 22,
          "probabilities": {
            "zh": 0.871153898537159
          },
          "status": "UNCERTAIN"
        }
      ],
      "VAD_region_ids": [
        22
      ],
      "language_confidence": 0.871153898537159,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 104.7,
      "packet_end": 134.10000000000002,
      "raw_text": "按你自己的用量计备",
      "selected_text": "计备",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            59
          ],
          "packet_id": 4,
          "engine": "mlx-whisper"
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "59",
          "before": "按你自己的用量计备",
          "after": "计备",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "计备疑似计费误识",
          "context_evidence": "上下文为用量，'计费'更符合逻辑",
          "independent_ASR": {
            "engine": "whisper-large-v3",
            "model_path": "/Users/sagawa/AI_WORK/cache/huggingface/hub/models--mlx-community--whisper-large-v3",
            "language": "zh",
            "audio_interval": [
              122.84,
              128.57999999999998
            ],
            "target_interval": [
              124.84,
              126.58
            ],
            "text": "按你自己的用量计备",
            "words": [
              {
                "word": "按",
                "start": 124.72,
                "end": 125.12,
                "probability": 0.9404296875
              },
              {
                "word": "你",
                "start": 125.12,
                "end": 125.32000000000001,
                "probability": 0.94775390625
              },
              {
                "word": "自己",
                "start": 125.32000000000001,
                "end": 125.58,
                "probability": 0.99365234375
              },
              {
                "word": "的",
                "start": 125.58,
                "end": 125.82000000000001,
                "probability": 0.98388671875
              },
              {
                "word": "用",
                "start": 125.82000000000001,
                "end": 125.9,
                "probability": 0.98193359375
              },
              {
                "word": "量",
                "start": 125.9,
                "end": 126.10000000000001,
                "probability": 0.88720703125
              },
              {
                "word": "计",
                "start": 126.10000000000001,
                "end": 126.32000000000001,
                "probability": 0.88671875
              },
              {
                "word": "备",
                "start": 126.32000000000001,
                "end": 126.52000000000001,
                "probability": 0.6865234375
              }
            ],
            "raw_result": {
              "text": "OpenAI的接口都能用按你自己的用量计备啊",
              "segments": [
                {
                  "id": 0,
                  "seek": 0,
                  "start": 0.0,
                  "end": 1.32,
                  "text": "OpenAI的接口都能用",
                  "tokens": [
                    50365,
                    45569,
                    48698,
                    1546,
                    14468,
                    18144,
                    7182,
                    8225,
                    9254,
                    50437
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.3802910584669847,
                  "compression_ratio": 0.8571428571428571,
                  "no_speech_prob": 0.7416872382164001,
                  "words": [
                    {
                      "word": "Open",
                      "start": 0.0,
                      "end": 0.14,
                      "probability": 0.0011539459228515625
                    },
                    {
                      "word": "AI",
                      "start": 0.14,
                      "end": 0.28,
                      "probability": 0.958984375
                    },
                    {
                      "word": "的",
                      "start": 0.28,
                      "end": 0.5,
                      "probability": 0.78564453125
                    },
                    {
                      "word": "接",
                      "start": 0.5,
                      "end": 0.64,
                      "probability": 0.86181640625
                    },
                    {
                      "word": "口",
                      "start": 0.64,
                      "end": 0.8,
                      "probability": 0.99951171875
                    },
                    {
                      "word": "都",
                      "start": 0.8,
                      "end": 0.94,
                      "probability": 0.994140625
                    },
                    {
                      "word": "能",
                      "start": 0.94,
                      "end": 1.08,
                      "probability": 0.998046875
                    },
                    {
                      "word": "用",
                      "start": 1.08,
                      "end": 1.32,
                      "probability": 0.99365234375
                    }
                  ]
                },
                {
                  "id": 1,
                  "seek": 0,
                  "start": 1.8799999999999994,
                  "end": 3.68,
                  "text": "按你自己的用量计备",
                  "tokens": [
                    50437,
                    26613,
                    2166,
                    17645,
                    1546,
                    9254,
                    26748,
                    7422,
                    94,
                    1787,
                    229,
                    50553
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.3802910584669847,
                  "compression_ratio": 0.8571428571428571,
                  "no_speech_prob": 0.7416872382164001,
                  "words": [
                    {
                      "word": "按",
                      "start": 1.8799999999999994,
                      "end": 2.28,
                      "probability": 0.9404296875
                    },
                    {
                      "word": "你",
                      "start": 2.28,
                      "end": 2.48,
                      "probability": 0.94775390625
                    },
                    {
                      "word": "自己",
                      "start": 2.48,
                      "end": 2.74,
                      "probability": 0.99365234375
                    },
                    {
                      "word": "的",
                      "start": 2.74,
                      "end": 2.98,
                      "probability": 0.98388671875
                    },
                    {
                      "word": "用",
                      "start": 2.98,
                      "end": 3.06,
                      "probability": 0.98193359375
                    },
                    {
                      "word": "量",
                      "start": 3.06,
                      "end": 3.26,
                      "probability": 0.88720703125
                    },
                    {
                      "word": "计",
                      "start": 3.26,
                      "end": 3.48,
                      "probability": 0.88671875
                    },
                    {
                      "word": "备",
                      "start": 3.48,
                      "end": 3.68,
                      "probability": 0.6865234375
                    }
                  ]
                },
                {
                  "id": 2,
                  "seek": 0,
                  "start": 3.68,
                  "end": 4.46,
                  "text": "啊",
                  "tokens": [
                    50553,
                    4905,
                    50595
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.3802910584669847,
                  "compression_ratio": 0.8571428571428571,
                  "no_speech_prob": 0.7416872382164001,
                  "words": [
                    {
                      "word": "啊",
                      "start": 3.68,
                      "end": 4.46,
                      "probability": 0.75732421875
                    }
                  ]
                }
              ],
              "language": "zh",
              "duration": 5.74
            },
            "status": "COMPLETE",
            "requested_language": "zh"
          },
          "suppression_proposed": null
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "59",
          "before": "计备",
          "after": "计备",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "Nonsense: '计备' vs '计费' (billing)",
          "context_evidence": "Context: '按你自己的用量计费'",
          "independent_ASR": {
            "engine": "whisper-large-v3",
            "model_path": "/Users/sagawa/AI_WORK/cache/huggingface/hub/models--mlx-community--whisper-large-v3",
            "language": "zh",
            "audio_interval": [
              122.84,
              128.57999999999998
            ],
            "target_interval": [
              124.84,
              126.58
            ],
            "text": "按你自己的用量计备",
            "words": [
              {
                "word": "按",
                "start": 124.72,
                "end": 125.12,
                "probability": 0.9404296875
              },
              {
                "word": "你",
                "start": 125.12,
                "end": 125.32000000000001,
                "probability": 0.94775390625
              },
              {
                "word": "自己",
                "start": 125.32000000000001,
                "end": 125.58,
                "probability": 0.99365234375
              },
              {
                "word": "的",
                "start": 125.58,
                "end": 125.82000000000001,
                "probability": 0.98388671875
              },
              {
                "word": "用",
                "start": 125.82000000000001,
                "end": 125.9,
                "probability": 0.98193359375
              },
              {
                "word": "量",
                "start": 125.9,
                "end": 126.10000000000001,
                "probability": 0.88720703125
              },
              {
                "word": "计",
                "start": 126.10000000000001,
                "end": 126.32000000000001,
                "probability": 0.88671875
              },
              {
                "word": "备",
                "start": 126.32000000000001,
                "end": 126.52000000000001,
                "probability": 0.6865234375
              }
            ],
            "raw_result": {
              "text": "OpenAI的接口都能用按你自己的用量计备啊",
              "segments": [
                {
                  "id": 0,
                  "seek": 0,
                  "start": 0.0,
                  "end": 1.32,
                  "text": "OpenAI的接口都能用",
                  "tokens": [
                    50365,
                    45569,
                    48698,
                    1546,
                    14468,
                    18144,
                    7182,
                    8225,
                    9254,
                    50437
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.3802910584669847,
                  "compression_ratio": 0.8571428571428571,
                  "no_speech_prob": 0.7416872382164001,
                  "words": [
                    {
                      "word": "Open",
                      "start": 0.0,
                      "end": 0.14,
                      "probability": 0.0011539459228515625
                    },
                    {
                      "word": "AI",
                      "start": 0.14,
                      "end": 0.28,
                      "probability": 0.958984375
                    },
                    {
                      "word": "的",
                      "start": 0.28,
                      "end": 0.5,
                      "probability": 0.78564453125
                    },
                    {
                      "word": "接",
                      "start": 0.5,
                      "end": 0.64,
                      "probability": 0.86181640625
                    },
                    {
                      "word": "口",
                      "start": 0.64,
                      "end": 0.8,
                      "probability": 0.99951171875
                    },
                    {
                      "word": "都",
                      "start": 0.8,
                      "end": 0.94,
                      "probability": 0.994140625
                    },
                    {
                      "word": "能",
                      "start": 0.94,
                      "end": 1.08,
                      "probability": 0.998046875
                    },
                    {
                      "word": "用",
                      "start": 1.08,
                      "end": 1.32,
                      "probability": 0.99365234375
                    }
                  ]
                },
                {
                  "id": 1,
                  "seek": 0,
                  "start": 1.8799999999999994,
                  "end": 3.68,
                  "text": "按你自己的用量计备",
                  "tokens": [
                    50437,
                    26613,
                    2166,
                    17645,
                    1546,
                    9254,
                    26748,
                    7422,
                    94,
                    1787,
                    229,
                    50553
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.3802910584669847,
                  "compression_ratio": 0.8571428571428571,
                  "no_speech_prob": 0.7416872382164001,
                  "words": [
                    {
                      "word": "按",
                      "start": 1.8799999999999994,
                      "end": 2.28,
                      "probability": 0.9404296875
                    },
                    {
                      "word": "你",
                      "start": 2.28,
                      "end": 2.48,
                      "probability": 0.94775390625
                    },
                    {
                      "word": "自己",
                      "start": 2.48,
                      "end": 2.74,
                      "probability": 0.99365234375
                    },
                    {
                      "word": "的",
                      "start": 2.74,
                      "end": 2.98,
                      "probability": 0.98388671875
                    },
                    {
                      "word": "用",
                      "start": 2.98,
                      "end": 3.06,
                      "probability": 0.98193359375
                    },
                    {
                      "word": "量",
                      "start": 3.06,
                      "end": 3.26,
                      "probability": 0.88720703125
                    },
                    {
                      "word": "计",
                      "start": 3.26,
                      "end": 3.48,
                      "probability": 0.88671875
                    },
                    {
                      "word": "备",
                      "start": 3.48,
                      "end": 3.68,
                      "probability": 0.6865234375
                    }
                  ]
                },
                {
                  "id": 2,
                  "seek": 0,
                  "start": 3.68,
                  "end": 4.46,
                  "text": "啊",
                  "tokens": [
                    50553,
                    4905,
                    50595
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.3802910584669847,
                  "compression_ratio": 0.8571428571428571,
                  "no_speech_prob": 0.7416872382164001,
                  "words": [
                    {
                      "word": "啊",
                      "start": 3.68,
                      "end": 4.46,
                      "probability": 0.75732421875
                    }
                  ]
                }
              ],
              "language": "zh",
              "duration": 5.74
            },
            "status": "COMPLETE",
            "requested_language": "zh"
          },
          "suppression_proposed": null
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "59",
          "before": "计备",
          "after": "计备",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "Nonsense: '计备' vs '计费' (billing)",
          "context_evidence": "Context: '按你自己的用量计费'",
          "independent_ASR": {
            "engine": "whisper-large-v3",
            "model_path": "/Users/sagawa/AI_WORK/cache/huggingface/hub/models--mlx-community--whisper-large-v3",
            "language": "zh",
            "audio_interval": [
              122.84,
              128.57999999999998
            ],
            "target_interval": [
              124.84,
              126.58
            ],
            "text": "按你自己的用量计备",
            "words": [
              {
                "word": "按",
                "start": 124.72,
                "end": 125.12,
                "probability": 0.9404296875
              },
              {
                "word": "你",
                "start": 125.12,
                "end": 125.32000000000001,
                "probability": 0.94775390625
              },
              {
                "word": "自己",
                "start": 125.32000000000001,
                "end": 125.58,
                "probability": 0.99365234375
              },
              {
                "word": "的",
                "start": 125.58,
                "end": 125.82000000000001,
                "probability": 0.98388671875
              },
              {
                "word": "用",
                "start": 125.82000000000001,
                "end": 125.9,
                "probability": 0.98193359375
              },
              {
                "word": "量",
                "start": 125.9,
                "end": 126.10000000000001,
                "probability": 0.88720703125
              },
              {
                "word": "计",
                "start": 126.10000000000001,
                "end": 126.32000000000001,
                "probability": 0.88671875
              },
              {
                "word": "备",
                "start": 126.32000000000001,
                "end": 126.52000000000001,
                "probability": 0.6865234375
              }
            ],
            "raw_result": {
              "text": "OpenAI的接口都能用按你自己的用量计备啊",
              "segments": [
                {
                  "id": 0,
                  "seek": 0,
                  "start": 0.0,
                  "end": 1.32,
                  "text": "OpenAI的接口都能用",
                  "tokens": [
                    50365,
                    45569,
                    48698,
                    1546,
                    14468,
                    18144,
                    7182,
                    8225,
                    9254,
                    50437
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.3802910584669847,
                  "compression_ratio": 0.8571428571428571,
                  "no_speech_prob": 0.7416872382164001,
                  "words": [
                    {
                      "word": "Open",
                      "start": 0.0,
                      "end": 0.14,
                      "probability": 0.0011539459228515625
                    },
                    {
                      "word": "AI",
                      "start": 0.14,
                      "end": 0.28,
                      "probability": 0.958984375
                    },
                    {
                      "word": "的",
                      "start": 0.28,
                      "end": 0.5,
                      "probability": 0.78564453125
                    },
                    {
                      "word": "接",
                      "start": 0.5,
                      "end": 0.64,
                      "probability": 0.86181640625
                    },
                    {
                      "word": "口",
                      "start": 0.64,
                      "end": 0.8,
                      "probability": 0.99951171875
                    },
                    {
                      "word": "都",
                      "start": 0.8,
                      "end": 0.94,
                      "probability": 0.994140625
                    },
                    {
                      "word": "能",
                      "start": 0.94,
                      "end": 1.08,
                      "probability": 0.998046875
                    },
                    {
                      "word": "用",
                      "start": 1.08,
                      "end": 1.32,
                      "probability": 0.99365234375
                    }
                  ]
                },
                {
                  "id": 1,
                  "seek": 0,
                  "start": 1.8799999999999994,
                  "end": 3.68,
                  "text": "按你自己的用量计备",
                  "tokens": [
                    50437,
                    26613,
                    2166,
                    17645,
                    1546,
                    9254,
                    26748,
                    7422,
                    94,
                    1787,
                    229,
                    50553
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.3802910584669847,
                  "compression_ratio": 0.8571428571428571,
                  "no_speech_prob": 0.7416872382164001,
                  "words": [
                    {
                      "word": "按",
                      "start": 1.8799999999999994,
                      "end": 2.28,
                      "probability": 0.9404296875
                    },
                    {
                      "word": "你",
                      "start": 2.28,
                      "end": 2.48,
                      "probability": 0.94775390625
                    },
                    {
                      "word": "自己",
                      "start": 2.48,
                      "end": 2.74,
                      "probability": 0.99365234375
                    },
                    {
                      "word": "的",
                      "start": 2.74,
                      "end": 2.98,
                      "probability": 0.98388671875
                    },
                    {
                      "word": "用",
                      "start": 2.98,
                      "end": 3.06,
                      "probability": 0.98193359375
                    },
                    {
                      "word": "量",
                      "start": 3.06,
                      "end": 3.26,
                      "probability": 0.88720703125
                    },
                    {
                      "word": "计",
                      "start": 3.26,
                      "end": 3.48,
                      "probability": 0.88671875
                    },
                    {
                      "word": "备",
                      "start": 3.48,
                      "end": 3.68,
                      "probability": 0.6865234375
                    }
                  ]
                },
                {
                  "id": 2,
                  "seek": 0,
                  "start": 3.68,
                  "end": 4.46,
                  "text": "啊",
                  "tokens": [
                    50553,
                    4905,
                    50595
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.3802910584669847,
                  "compression_ratio": 0.8571428571428571,
                  "no_speech_prob": 0.7416872382164001,
                  "words": [
                    {
                      "word": "啊",
                      "start": 3.68,
                      "end": 4.46,
                      "probability": 0.75732421875
                    }
                  ]
                }
              ],
              "language": "zh",
              "duration": 5.74
            },
            "status": "COMPLETE",
            "requested_language": "zh"
          },
          "suppression_proposed": null
        }
      ],
      "language": "zh",
      "packet_ids": [
        4
      ],
      "classification": "BEST_SUPPORTED_WITH_REVIEW",
      "independent_ASR": {
        "engine": "whisper-large-v3",
        "model_path": "/Users/sagawa/AI_WORK/cache/huggingface/hub/models--mlx-community--whisper-large-v3",
        "language": "zh",
        "audio_interval": [
          122.84,
          128.57999999999998
        ],
        "target_interval": [
          124.84,
          126.58
        ],
        "text": "按你自己的用量计备",
        "words": [
          {
            "word": "按",
            "start": 124.72,
            "end": 125.12,
            "probability": 0.9404296875
          },
          {
            "word": "你",
            "start": 125.12,
            "end": 125.32000000000001,
            "probability": 0.94775390625
          },
          {
            "word": "自己",
            "start": 125.32000000000001,
            "end": 125.58,
            "probability": 0.99365234375
          },
          {
            "word": "的",
            "start": 125.58,
            "end": 125.82000000000001,
            "probability": 0.98388671875
          },
          {
            "word": "用",
            "start": 125.82000000000001,
            "end": 125.9,
            "probability": 0.98193359375
          },
          {
            "word": "量",
            "start": 125.9,
            "end": 126.10000000000001,
            "probability": 0.88720703125
          },
          {
            "word": "计",
            "start": 126.10000000000001,
            "end": 126.32000000000001,
            "probability": 0.88671875
          },
          {
            "word": "备",
            "start": 126.32000000000001,
            "end": 126.52000000000001,
            "probability": 0.6865234375
          }
        ],
        "raw_result": {
          "text": "OpenAI的接口都能用按你自己的用量计备啊",
          "segments": [
            {
              "id": 0,
              "seek": 0,
              "start": 0.0,
              "end": 1.32,
              "text": "OpenAI的接口都能用",
              "tokens": [
                50365,
                45569,
                48698,
                1546,
                14468,
                18144,
                7182,
                8225,
                9254,
                50437
              ],
              "temperature": 0.0,
              "avg_logprob": -0.3802910584669847,
              "compression_ratio": 0.8571428571428571,
              "no_speech_prob": 0.7416872382164001,
              "words": [
                {
                  "word": "Open",
                  "start": 0.0,
                  "end": 0.14,
                  "probability": 0.0011539459228515625
                },
                {
                  "word": "AI",
                  "start": 0.14,
                  "end": 0.28,
                  "probability": 0.958984375
                },
                {
                  "word": "的",
                  "start": 0.28,
                  "end": 0.5,
                  "probability": 0.78564453125
                },
                {
                  "word": "接",
                  "start": 0.5,
                  "end": 0.64,
                  "probability": 0.86181640625
                },
                {
                  "word": "口",
                  "start": 0.64,
                  "end": 0.8,
                  "probability": 0.99951171875
                },
                {
                  "word": "都",
                  "start": 0.8,
                  "end": 0.94,
                  "probability": 0.994140625
                },
                {
                  "word": "能",
                  "start": 0.94,
                  "end": 1.08,
                  "probability": 0.998046875
                },
                {
                  "word": "用",
                  "start": 1.08,
                  "end": 1.32,
                  "probability": 0.99365234375
                }
              ]
            },
            {
              "id": 1,
              "seek": 0,
              "start": 1.8799999999999994,
              "end": 3.68,
              "text": "按你自己的用量计备",
              "tokens": [
                50437,
                26613,
                2166,
                17645,
                1546,
                9254,
                26748,
                7422,
                94,
                1787,
                229,
                50553
              ],
              "temperature": 0.0,
              "avg_logprob": -0.3802910584669847,
              "compression_ratio": 0.8571428571428571,
              "no_speech_prob": 0.7416872382164001,
              "words": [
                {
                  "word": "按",
                  "start": 1.8799999999999994,
                  "end": 2.28,
                  "probability": 0.9404296875
                },
                {
                  "word": "你",
                  "start": 2.28,
                  "end": 2.48,
                  "probability": 0.94775390625
                },
                {
                  "word": "自己",
                  "start": 2.48,
                  "end": 2.74,
                  "probability": 0.99365234375
                },
                {
                  "word": "的",
                  "start": 2.74,
                  "end": 2.98,
                  "probability": 0.98388671875
                },
                {
                  "word": "用",
                  "start": 2.98,
                  "end": 3.06,
                  "probability": 0.98193359375
                },
                {
                  "word": "量",
                  "start": 3.06,
                  "end": 3.26,
                  "probability": 0.88720703125
                },
                {
                  "word": "计",
                  "start": 3.26,
                  "end": 3.48,
                  "probability": 0.88671875
                },
                {
                  "word": "备",
                  "start": 3.48,
                  "end": 3.68,
                  "probability": 0.6865234375
                }
              ]
            },
            {
              "id": 2,
              "seek": 0,
              "start": 3.68,
              "end": 4.46,
              "text": "啊",
              "tokens": [
                50553,
                4905,
                50595
              ],
              "temperature": 0.0,
              "avg_logprob": -0.3802910584669847,
              "compression_ratio": 0.8571428571428571,
              "no_speech_prob": 0.7416872382164001,
              "words": [
                {
                  "word": "啊",
                  "start": 3.68,
                  "end": 4.46,
                  "probability": 0.75732421875
                }
              ]
            }
          ],
          "language": "zh",
          "duration": 5.74
        },
        "status": "COMPLETE",
        "requested_language": "zh"
      },
      "raw_word_timestamps": [
        {
          "word": "按",
          "start": 124.84,
          "end": 125.2,
          "probability": 0.9990234375
        },
        {
          "word": "你",
          "start": 125.2,
          "end": 125.38,
          "probability": 0.998046875
        },
        {
          "word": "自己",
          "start": 125.38,
          "end": 125.7,
          "probability": 0.99951171875
        },
        {
          "word": "的",
          "start": 125.7,
          "end": 125.84,
          "probability": 1.0
        },
        {
          "word": "用",
          "start": 125.84,
          "end": 125.98,
          "probability": 0.98046875
        },
        {
          "word": "量",
          "start": 125.98,
          "end": 126.16,
          "probability": 0.85888671875
        },
        {
          "word": "计",
          "start": 126.16,
          "end": 126.34,
          "probability": 0.97265625
        },
        {
          "word": "备",
          "start": 126.34,
          "end": 126.58,
          "probability": 0.765625
        }
      ],
      "word_timestamps": [],
      "source_text": "计备",
      "review_status": "BEST_SUPPORTED_WITH_REVIEW",
      "canonical_timing_evidence": "CUE_ANCHORED",
      "index": 59
    },
    {
      "start": 127.04,
      "end": 127.4,
      "text": "啊",
      "raw_asr_text": "啊",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 4,
      "words": [
        {
          "word": "啊",
          "start": 127.04,
          "end": 127.4,
          "probability": 0.25439453125
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.1897983819666043,
        "compression_ratio": 1.2267657992565055,
        "no_speech_prob": 6.547795639022524e-11
      },
      "lid_status": "UNCERTAIN",
      "alternatives": [
        {
          "language": "zh",
          "text": "甚至直接点一下让大模型帮你自动生成术语翻译时它还会自动结合上文语境不会翻得前言不搭后语翻译接入的是你自己的大模型API像Deepstick同意签问或者任何建容OpenAI的接口都能用按你自己的用量计备啊会自动释出关闭模型思考过程的参数来降低延迟到底"
        }
      ],
      "id": 60,
      "review_annotation": "",
      "row_id": "60",
      "bounded_row_id": 60,
      "parent_row_ids": [
        60
      ],
      "parent_row_id": 60,
      "packet_id": 4,
      "LID_evidence": [],
      "VAD_region_ids": [],
      "language_confidence": 0,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 104.7,
      "packet_end": 134.10000000000002,
      "raw_text": "啊",
      "selected_text": "啊",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            60
          ],
          "packet_id": 4,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        4
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 60
    },
    {
      "start": 128.16,
      "end": 129.32,
      "text": "会自动释出",
      "raw_asr_text": "会自动释出",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 4,
      "words": [
        {
          "word": "会",
          "start": 128.16,
          "end": 128.52,
          "probability": 0.9765625
        },
        {
          "word": "自",
          "start": 128.52,
          "end": 128.7,
          "probability": 1.0
        },
        {
          "word": "动",
          "start": 128.7,
          "end": 128.84,
          "probability": 0.99951171875
        },
        {
          "word": "释",
          "start": 128.84,
          "end": 129.02,
          "probability": 0.9501953125
        },
        {
          "word": "出",
          "start": 129.02,
          "end": 129.32,
          "probability": 0.9990234375
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.1897983819666043,
        "compression_ratio": 1.2267657992565055,
        "no_speech_prob": 6.547795639022524e-11
      },
      "lid_status": "UNCERTAIN",
      "alternatives": [
        {
          "language": "zh",
          "text": "甚至直接点一下让大模型帮你自动生成术语翻译时它还会自动结合上文语境不会翻得前言不搭后语翻译接入的是你自己的大模型API像Deepstick同意签问或者任何建容OpenAI的接口都能用按你自己的用量计备啊会自动释出关闭模型思考过程的参数来降低延迟到底"
        }
      ],
      "id": 61,
      "review_annotation": "",
      "row_id": "61",
      "bounded_row_id": 61,
      "parent_row_ids": [
        61
      ],
      "parent_row_id": 61,
      "packet_id": 4,
      "LID_evidence": [
        {
          "region_id": 23,
          "probabilities": {
            "zh": 0.9974001348018646
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        23
      ],
      "language_confidence": 0.9974001348018646,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 104.7,
      "packet_end": 134.10000000000002,
      "raw_text": "会自动释出",
      "selected_text": "会自动释出",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            61
          ],
          "packet_id": 4,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        4
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 61
    },
    {
      "start": 129.32,
      "end": 132.92000000000002,
      "text": "关闭模型思考过程的参数来降低延迟",
      "raw_asr_text": "关闭模型思考过程的参数来降低延迟",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 4,
      "words": [
        {
          "word": "关",
          "start": 129.32,
          "end": 129.82,
          "probability": 0.22021484375
        },
        {
          "word": "闭",
          "start": 129.82,
          "end": 129.94,
          "probability": 0.943359375
        },
        {
          "word": "模",
          "start": 129.94,
          "end": 130.1,
          "probability": 0.99951171875
        },
        {
          "word": "型",
          "start": 130.1,
          "end": 130.28,
          "probability": 1.0
        },
        {
          "word": "思",
          "start": 130.28,
          "end": 130.46,
          "probability": 0.99951171875
        },
        {
          "word": "考",
          "start": 130.46,
          "end": 130.62,
          "probability": 1.0
        },
        {
          "word": "过",
          "start": 130.62,
          "end": 130.8,
          "probability": 0.99853515625
        },
        {
          "word": "程",
          "start": 130.8,
          "end": 131.0,
          "probability": 1.0
        },
        {
          "word": "的",
          "start": 131.0,
          "end": 131.1,
          "probability": 0.99951171875
        },
        {
          "word": "参",
          "start": 131.1,
          "end": 131.32,
          "probability": 0.99951171875
        },
        {
          "word": "数",
          "start": 131.32,
          "end": 131.52,
          "probability": 0.99462890625
        },
        {
          "word": "来",
          "start": 131.52,
          "end": 131.98000000000002,
          "probability": 0.99951171875
        },
        {
          "word": "降",
          "start": 131.98000000000002,
          "end": 132.18,
          "probability": 0.998046875
        },
        {
          "word": "低",
          "start": 132.18,
          "end": 132.4,
          "probability": 0.9990234375
        },
        {
          "word": "延",
          "start": 132.4,
          "end": 132.62,
          "probability": 0.99951171875
        },
        {
          "word": "迟",
          "start": 132.62,
          "end": 132.92000000000002,
          "probability": 0.99609375
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.1897983819666043,
        "compression_ratio": 1.2267657992565055,
        "no_speech_prob": 6.547795639022524e-11
      },
      "lid_status": "UNCERTAIN",
      "alternatives": [
        {
          "language": "zh",
          "text": "甚至直接点一下让大模型帮你自动生成术语翻译时它还会自动结合上文语境不会翻得前言不搭后语翻译接入的是你自己的大模型API像Deepstick同意签问或者任何建容OpenAI的接口都能用按你自己的用量计备啊会自动释出关闭模型思考过程的参数来降低延迟到底"
        }
      ],
      "id": 62,
      "review_annotation": "",
      "row_id": "62",
      "bounded_row_id": 62,
      "parent_row_ids": [
        62
      ],
      "parent_row_id": 62,
      "packet_id": 4,
      "LID_evidence": [
        {
          "region_id": 23,
          "probabilities": {
            "zh": 0.9974001348018646
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        23
      ],
      "language_confidence": 0.9974001348018646,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 104.7,
      "packet_end": 134.10000000000002,
      "raw_text": "关闭模型思考过程的参数来降低延迟",
      "selected_text": "关闭模型思考过程的参数来降低延迟",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            62
          ],
          "packet_id": 4,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        4
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 62
    },
    {
      "start": 133.5,
      "end": 137.23999999999998,
      "text": "资金一请求头",
      "raw_asr_text": "高级设置里还能填user agent和资金一请求头",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 5,
      "words": [],
      "asr_metrics": {
        "avg_logprob": -0.14073291240922556,
        "compression_ratio": 1.1727574750830565,
        "no_speech_prob": 2.7202724448005533e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "高级设置里还能填user agent和资金一请求头不需要Key的服务把Key留空就行另外设置里还提供了一个可选的Javi在线辅助功能打开之后可以帮你做智能断句和一文检查所以平时用起来其实就三步填好你的大模型API并测试连接允许权限然后点一下开始字幕就可以安心看直播了"
        }
      ],
      "id": 63,
      "review_annotation": "待確認",
      "row_id": "63",
      "bounded_row_id": 63,
      "parent_row_ids": [
        63
      ],
      "parent_row_id": 63,
      "packet_id": 5,
      "LID_evidence": [
        {
          "region_id": 24,
          "probabilities": {
            "zh": 0.9884023517370224
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        24
      ],
      "language_confidence": 0.9884023517370224,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 132.7,
      "packet_end": 161.3,
      "raw_text": "高级设置里还能填user agent和资金一请求头",
      "selected_text": "资金一请求头",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            63
          ],
          "packet_id": 5,
          "engine": "mlx-whisper"
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "63",
          "before": "高级设置里还能填user agent和资金一请求头",
          "after": "资金一请求头",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "资金一疑似自定义请求头误识",
          "context_evidence": "上下文为API配置，'自定义'更符合逻辑",
          "independent_ASR": {
            "engine": "whisper-large-v3",
            "model_path": "/Users/sagawa/AI_WORK/cache/huggingface/hub/models--mlx-community--whisper-large-v3",
            "language": "zh",
            "audio_interval": [
              131.5,
              139.23999999999998
            ],
            "target_interval": [
              133.5,
              137.23999999999998
            ],
            "text": "高级设置里还能添UserAgent和资金一起求头",
            "words": [
              {
                "word": "高",
                "start": 133.3,
                "end": 133.74,
                "probability": 0.97119140625
              },
              {
                "word": "级",
                "start": 133.74,
                "end": 134.02,
                "probability": 0.9990234375
              },
              {
                "word": "设",
                "start": 134.02,
                "end": 134.18,
                "probability": 0.9990234375
              },
              {
                "word": "置",
                "start": 134.18,
                "end": 134.26,
                "probability": 0.99853515625
              },
              {
                "word": "里",
                "start": 134.26,
                "end": 134.44,
                "probability": 0.998046875
              },
              {
                "word": "还",
                "start": 134.44,
                "end": 134.76,
                "probability": 0.994140625
              },
              {
                "word": "能",
                "start": 134.76,
                "end": 134.9,
                "probability": 0.99951171875
              },
              {
                "word": "添",
                "start": 134.9,
                "end": 135.06,
                "probability": 0.9375
              },
              {
                "word": "U",
                "start": 135.06,
                "end": 135.16,
                "probability": 0.673828125
              },
              {
                "word": "ser",
                "start": 135.16,
                "end": 135.4,
                "probability": 0.99560546875
              },
              {
                "word": " Agent",
                "start": 135.4,
                "end": 135.74,
                "probability": 0.89697265625
              },
              {
                "word": "和",
                "start": 135.74,
                "end": 136.24,
                "probability": 0.98681640625
              },
              {
                "word": "资",
                "start": 136.24,
                "end": 136.5,
                "probability": 0.93603515625
              },
              {
                "word": "金",
                "start": 136.5,
                "end": 136.58,
                "probability": 0.78173828125
              },
              {
                "word": "一起",
                "start": 136.58,
                "end": 136.72,
                "probability": 0.79345703125
              },
              {
                "word": "求",
                "start": 136.72,
                "end": 136.98,
                "probability": 0.87109375
              },
              {
                "word": "头",
                "start": 136.98,
                "end": 137.2,
                "probability": 0.68115234375
              }
            ],
            "raw_result": {
              "text": "来降低延迟高级设置里还能添User Agent和资金一起求头不需要Key的服务",
              "segments": [
                {
                  "id": 0,
                  "seek": 0,
                  "start": 0.0,
                  "end": 1.32,
                  "text": "来降低延迟",
                  "tokens": [
                    50365,
                    6912,
                    47421,
                    41377,
                    17527,
                    114,
                    3316,
                    253,
                    50436
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.1633597640104072,
                  "compression_ratio": 0.8921568627450981,
                  "no_speech_prob": 0.8626763820648193,
                  "words": [
                    {
                      "word": "来",
                      "start": 0.0,
                      "end": 0.44,
                      "probability": 0.9052734375
                    },
                    {
                      "word": "降",
                      "start": 0.44,
                      "end": 0.64,
                      "probability": 0.9970703125
                    },
                    {
                      "word": "低",
                      "start": 0.64,
                      "end": 0.86,
                      "probability": 0.9990234375
                    },
                    {
                      "word": "延",
                      "start": 0.86,
                      "end": 1.08,
                      "probability": 0.99853515625
                    },
                    {
                      "word": "迟",
                      "start": 1.08,
                      "end": 1.32,
                      "probability": 0.998046875
                    }
                  ]
                },
                {
                  "id": 1,
                  "seek": 0,
                  "start": 1.8000000000000003,
                  "end": 5.7,
                  "text": "高级设置里还能添User Agent和资金一起求头",
                  "tokens": [
                    50436,
                    12979,
                    16853,
                    100,
                    7422,
                    122,
                    34719,
                    15759,
                    14852,
                    8225,
                    14176,
                    119,
                    52,
                    12484,
                    27174,
                    12565,
                    5266,
                    226,
                    19117,
                    29567,
                    32718,
                    39862,
                    50657
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.1633597640104072,
                  "compression_ratio": 0.8921568627450981,
                  "no_speech_prob": 0.8626763820648193,
                  "words": [
                    {
                      "word": "高",
                      "start": 1.8000000000000003,
                      "end": 2.24,
                      "probability": 0.97119140625
                    },
                    {
                      "word": "级",
                      "start": 2.24,
                      "end": 2.52,
                      "probability": 0.9990234375
                    },
                    {
                      "word": "设",
                      "start": 2.52,
                      "end": 2.68,
                      "probability": 0.9990234375
                    },
                    {
                      "word": "置",
                      "start": 2.68,
                      "end": 2.76,
                      "probability": 0.99853515625
                    },
                    {
                      "word": "里",
                      "start": 2.76,
                      "end": 2.94,
                      "probability": 0.998046875
                    },
                    {
                      "word": "还",
                      "start": 2.94,
                      "end": 3.26,
                      "probability": 0.994140625
                    },
                    {
                      "word": "能",
                      "start": 3.26,
                      "end": 3.4,
                      "probability": 0.99951171875
                    },
                    {
                      "word": "添",
                      "start": 3.4,
                      "end": 3.56,
                      "probability": 0.9375
                    },
                    {
                      "word": "U",
                      "start": 3.56,
                      "end": 3.66,
                      "probability": 0.673828125
                    },
                    {
                      "word": "ser",
                      "start": 3.66,
                      "end": 3.9,
                      "probability": 0.99560546875
                    },
                    {
                      "word": " Agent",
                      "start": 3.9,
                      "end": 4.24,
                      "probability": 0.89697265625
                    },
                    {
                      "word": "和",
                      "start": 4.24,
                      "end": 4.74,
                      "probability": 0.98681640625
                    },
                    {
                      "word": "资",
                      "start": 4.74,
                      "end": 5.0,
                      "probability": 0.93603515625
                    },
                    {
                      "word": "金",
                      "start": 5.0,
                      "end": 5.08,
                      "probability": 0.78173828125
                    },
                    {
                      "word": "一起",
                      "start": 5.08,
                      "end": 5.22,
                      "probability": 0.79345703125
                    },
                    {
                      "word": "求",
                      "start": 5.22,
                      "end": 5.48,
                      "probability": 0.87109375
                    },
                    {
                      "word": "头",
                      "start": 5.48,
                      "end": 5.7,
                      "probability": 0.68115234375
                    }
                  ]
                },
                {
                  "id": 2,
                  "seek": 0,
                  "start": 5.7,
                  "end": 7.46,
                  "text": "不需要Key的服务",
                  "tokens": [
                    50657,
                    1960,
                    35748,
                    42,
                    2030,
                    1546,
                    27408,
                    5087,
                    94,
                    50740
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.1633597640104072,
                  "compression_ratio": 0.8921568627450981,
                  "no_speech_prob": 0.8626763820648193,
                  "words": [
                    {
                      "word": "不",
                      "start": 5.7,
                      "end": 6.48,
                      "probability": 0.9951171875
                    },
                    {
                      "word": "需要",
                      "start": 6.48,
                      "end": 6.76,
                      "probability": 0.99951171875
                    },
                    {
                      "word": "K",
                      "start": 6.76,
                      "end": 6.9,
                      "probability": 0.80810546875
                    },
                    {
                      "word": "ey",
                      "start": 6.9,
                      "end": 7.02,
                      "probability": 0.49072265625
                    },
                    {
                      "word": "的",
                      "start": 7.02,
                      "end": 7.08,
                      "probability": 0.9951171875
                    },
                    {
                      "word": "服",
                      "start": 7.08,
                      "end": 7.22,
                      "probability": 0.99951171875
                    },
                    {
                      "word": "务",
                      "start": 7.22,
                      "end": 7.46,
                      "probability": 1.0
                    }
                  ]
                }
              ],
              "language": "zh",
              "duration": 7.74
            },
            "status": "COMPLETE",
            "requested_language": "zh"
          },
          "suppression_proposed": null
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "63",
          "before": "资金一请求头",
          "after": "资金一请求头",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "Nonsense: '资金一请求头' vs 'UserAgent和请求头'",
          "context_evidence": "ASR: '高级设置里还能添UserAgent和资金一起求头'",
          "independent_ASR": {
            "engine": "whisper-large-v3",
            "model_path": "/Users/sagawa/AI_WORK/cache/huggingface/hub/models--mlx-community--whisper-large-v3",
            "language": "zh",
            "audio_interval": [
              131.5,
              139.23999999999998
            ],
            "target_interval": [
              133.5,
              137.23999999999998
            ],
            "text": "高级设置里还能添UserAgent和资金一起求头",
            "words": [
              {
                "word": "高",
                "start": 133.3,
                "end": 133.74,
                "probability": 0.97119140625
              },
              {
                "word": "级",
                "start": 133.74,
                "end": 134.02,
                "probability": 0.9990234375
              },
              {
                "word": "设",
                "start": 134.02,
                "end": 134.18,
                "probability": 0.9990234375
              },
              {
                "word": "置",
                "start": 134.18,
                "end": 134.26,
                "probability": 0.99853515625
              },
              {
                "word": "里",
                "start": 134.26,
                "end": 134.44,
                "probability": 0.998046875
              },
              {
                "word": "还",
                "start": 134.44,
                "end": 134.76,
                "probability": 0.994140625
              },
              {
                "word": "能",
                "start": 134.76,
                "end": 134.9,
                "probability": 0.99951171875
              },
              {
                "word": "添",
                "start": 134.9,
                "end": 135.06,
                "probability": 0.9375
              },
              {
                "word": "U",
                "start": 135.06,
                "end": 135.16,
                "probability": 0.673828125
              },
              {
                "word": "ser",
                "start": 135.16,
                "end": 135.4,
                "probability": 0.99560546875
              },
              {
                "word": " Agent",
                "start": 135.4,
                "end": 135.74,
                "probability": 0.89697265625
              },
              {
                "word": "和",
                "start": 135.74,
                "end": 136.24,
                "probability": 0.98681640625
              },
              {
                "word": "资",
                "start": 136.24,
                "end": 136.5,
                "probability": 0.93603515625
              },
              {
                "word": "金",
                "start": 136.5,
                "end": 136.58,
                "probability": 0.78173828125
              },
              {
                "word": "一起",
                "start": 136.58,
                "end": 136.72,
                "probability": 0.79345703125
              },
              {
                "word": "求",
                "start": 136.72,
                "end": 136.98,
                "probability": 0.87109375
              },
              {
                "word": "头",
                "start": 136.98,
                "end": 137.2,
                "probability": 0.68115234375
              }
            ],
            "raw_result": {
              "text": "来降低延迟高级设置里还能添User Agent和资金一起求头不需要Key的服务",
              "segments": [
                {
                  "id": 0,
                  "seek": 0,
                  "start": 0.0,
                  "end": 1.32,
                  "text": "来降低延迟",
                  "tokens": [
                    50365,
                    6912,
                    47421,
                    41377,
                    17527,
                    114,
                    3316,
                    253,
                    50436
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.1633597640104072,
                  "compression_ratio": 0.8921568627450981,
                  "no_speech_prob": 0.8626763820648193,
                  "words": [
                    {
                      "word": "来",
                      "start": 0.0,
                      "end": 0.44,
                      "probability": 0.9052734375
                    },
                    {
                      "word": "降",
                      "start": 0.44,
                      "end": 0.64,
                      "probability": 0.9970703125
                    },
                    {
                      "word": "低",
                      "start": 0.64,
                      "end": 0.86,
                      "probability": 0.9990234375
                    },
                    {
                      "word": "延",
                      "start": 0.86,
                      "end": 1.08,
                      "probability": 0.99853515625
                    },
                    {
                      "word": "迟",
                      "start": 1.08,
                      "end": 1.32,
                      "probability": 0.998046875
                    }
                  ]
                },
                {
                  "id": 1,
                  "seek": 0,
                  "start": 1.8000000000000003,
                  "end": 5.7,
                  "text": "高级设置里还能添User Agent和资金一起求头",
                  "tokens": [
                    50436,
                    12979,
                    16853,
                    100,
                    7422,
                    122,
                    34719,
                    15759,
                    14852,
                    8225,
                    14176,
                    119,
                    52,
                    12484,
                    27174,
                    12565,
                    5266,
                    226,
                    19117,
                    29567,
                    32718,
                    39862,
                    50657
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.1633597640104072,
                  "compression_ratio": 0.8921568627450981,
                  "no_speech_prob": 0.8626763820648193,
                  "words": [
                    {
                      "word": "高",
                      "start": 1.8000000000000003,
                      "end": 2.24,
                      "probability": 0.97119140625
                    },
                    {
                      "word": "级",
                      "start": 2.24,
                      "end": 2.52,
                      "probability": 0.9990234375
                    },
                    {
                      "word": "设",
                      "start": 2.52,
                      "end": 2.68,
                      "probability": 0.9990234375
                    },
                    {
                      "word": "置",
                      "start": 2.68,
                      "end": 2.76,
                      "probability": 0.99853515625
                    },
                    {
                      "word": "里",
                      "start": 2.76,
                      "end": 2.94,
                      "probability": 0.998046875
                    },
                    {
                      "word": "还",
                      "start": 2.94,
                      "end": 3.26,
                      "probability": 0.994140625
                    },
                    {
                      "word": "能",
                      "start": 3.26,
                      "end": 3.4,
                      "probability": 0.99951171875
                    },
                    {
                      "word": "添",
                      "start": 3.4,
                      "end": 3.56,
                      "probability": 0.9375
                    },
                    {
                      "word": "U",
                      "start": 3.56,
                      "end": 3.66,
                      "probability": 0.673828125
                    },
                    {
                      "word": "ser",
                      "start": 3.66,
                      "end": 3.9,
                      "probability": 0.99560546875
                    },
                    {
                      "word": " Agent",
                      "start": 3.9,
                      "end": 4.24,
                      "probability": 0.89697265625
                    },
                    {
                      "word": "和",
                      "start": 4.24,
                      "end": 4.74,
                      "probability": 0.98681640625
                    },
                    {
                      "word": "资",
                      "start": 4.74,
                      "end": 5.0,
                      "probability": 0.93603515625
                    },
                    {
                      "word": "金",
                      "start": 5.0,
                      "end": 5.08,
                      "probability": 0.78173828125
                    },
                    {
                      "word": "一起",
                      "start": 5.08,
                      "end": 5.22,
                      "probability": 0.79345703125
                    },
                    {
                      "word": "求",
                      "start": 5.22,
                      "end": 5.48,
                      "probability": 0.87109375
                    },
                    {
                      "word": "头",
                      "start": 5.48,
                      "end": 5.7,
                      "probability": 0.68115234375
                    }
                  ]
                },
                {
                  "id": 2,
                  "seek": 0,
                  "start": 5.7,
                  "end": 7.46,
                  "text": "不需要Key的服务",
                  "tokens": [
                    50657,
                    1960,
                    35748,
                    42,
                    2030,
                    1546,
                    27408,
                    5087,
                    94,
                    50740
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.1633597640104072,
                  "compression_ratio": 0.8921568627450981,
                  "no_speech_prob": 0.8626763820648193,
                  "words": [
                    {
                      "word": "不",
                      "start": 5.7,
                      "end": 6.48,
                      "probability": 0.9951171875
                    },
                    {
                      "word": "需要",
                      "start": 6.48,
                      "end": 6.76,
                      "probability": 0.99951171875
                    },
                    {
                      "word": "K",
                      "start": 6.76,
                      "end": 6.9,
                      "probability": 0.80810546875
                    },
                    {
                      "word": "ey",
                      "start": 6.9,
                      "end": 7.02,
                      "probability": 0.49072265625
                    },
                    {
                      "word": "的",
                      "start": 7.02,
                      "end": 7.08,
                      "probability": 0.9951171875
                    },
                    {
                      "word": "服",
                      "start": 7.08,
                      "end": 7.22,
                      "probability": 0.99951171875
                    },
                    {
                      "word": "务",
                      "start": 7.22,
                      "end": 7.46,
                      "probability": 1.0
                    }
                  ]
                }
              ],
              "language": "zh",
              "duration": 7.74
            },
            "status": "COMPLETE",
            "requested_language": "zh"
          },
          "suppression_proposed": null
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "63",
          "before": "资金一请求头",
          "after": "资金一请求头",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "Nonsense: '资金一请求头' vs 'UserAgent和请求头'",
          "context_evidence": "ASR: '高级设置里还能添UserAgent和资金一起求头'",
          "independent_ASR": {
            "engine": "whisper-large-v3",
            "model_path": "/Users/sagawa/AI_WORK/cache/huggingface/hub/models--mlx-community--whisper-large-v3",
            "language": "zh",
            "audio_interval": [
              131.5,
              139.23999999999998
            ],
            "target_interval": [
              133.5,
              137.23999999999998
            ],
            "text": "高级设置里还能添UserAgent和资金一起求头",
            "words": [
              {
                "word": "高",
                "start": 133.3,
                "end": 133.74,
                "probability": 0.97119140625
              },
              {
                "word": "级",
                "start": 133.74,
                "end": 134.02,
                "probability": 0.9990234375
              },
              {
                "word": "设",
                "start": 134.02,
                "end": 134.18,
                "probability": 0.9990234375
              },
              {
                "word": "置",
                "start": 134.18,
                "end": 134.26,
                "probability": 0.99853515625
              },
              {
                "word": "里",
                "start": 134.26,
                "end": 134.44,
                "probability": 0.998046875
              },
              {
                "word": "还",
                "start": 134.44,
                "end": 134.76,
                "probability": 0.994140625
              },
              {
                "word": "能",
                "start": 134.76,
                "end": 134.9,
                "probability": 0.99951171875
              },
              {
                "word": "添",
                "start": 134.9,
                "end": 135.06,
                "probability": 0.9375
              },
              {
                "word": "U",
                "start": 135.06,
                "end": 135.16,
                "probability": 0.673828125
              },
              {
                "word": "ser",
                "start": 135.16,
                "end": 135.4,
                "probability": 0.99560546875
              },
              {
                "word": " Agent",
                "start": 135.4,
                "end": 135.74,
                "probability": 0.89697265625
              },
              {
                "word": "和",
                "start": 135.74,
                "end": 136.24,
                "probability": 0.98681640625
              },
              {
                "word": "资",
                "start": 136.24,
                "end": 136.5,
                "probability": 0.93603515625
              },
              {
                "word": "金",
                "start": 136.5,
                "end": 136.58,
                "probability": 0.78173828125
              },
              {
                "word": "一起",
                "start": 136.58,
                "end": 136.72,
                "probability": 0.79345703125
              },
              {
                "word": "求",
                "start": 136.72,
                "end": 136.98,
                "probability": 0.87109375
              },
              {
                "word": "头",
                "start": 136.98,
                "end": 137.2,
                "probability": 0.68115234375
              }
            ],
            "raw_result": {
              "text": "来降低延迟高级设置里还能添User Agent和资金一起求头不需要Key的服务",
              "segments": [
                {
                  "id": 0,
                  "seek": 0,
                  "start": 0.0,
                  "end": 1.32,
                  "text": "来降低延迟",
                  "tokens": [
                    50365,
                    6912,
                    47421,
                    41377,
                    17527,
                    114,
                    3316,
                    253,
                    50436
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.1633597640104072,
                  "compression_ratio": 0.8921568627450981,
                  "no_speech_prob": 0.8626763820648193,
                  "words": [
                    {
                      "word": "来",
                      "start": 0.0,
                      "end": 0.44,
                      "probability": 0.9052734375
                    },
                    {
                      "word": "降",
                      "start": 0.44,
                      "end": 0.64,
                      "probability": 0.9970703125
                    },
                    {
                      "word": "低",
                      "start": 0.64,
                      "end": 0.86,
                      "probability": 0.9990234375
                    },
                    {
                      "word": "延",
                      "start": 0.86,
                      "end": 1.08,
                      "probability": 0.99853515625
                    },
                    {
                      "word": "迟",
                      "start": 1.08,
                      "end": 1.32,
                      "probability": 0.998046875
                    }
                  ]
                },
                {
                  "id": 1,
                  "seek": 0,
                  "start": 1.8000000000000003,
                  "end": 5.7,
                  "text": "高级设置里还能添User Agent和资金一起求头",
                  "tokens": [
                    50436,
                    12979,
                    16853,
                    100,
                    7422,
                    122,
                    34719,
                    15759,
                    14852,
                    8225,
                    14176,
                    119,
                    52,
                    12484,
                    27174,
                    12565,
                    5266,
                    226,
                    19117,
                    29567,
                    32718,
                    39862,
                    50657
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.1633597640104072,
                  "compression_ratio": 0.8921568627450981,
                  "no_speech_prob": 0.8626763820648193,
                  "words": [
                    {
                      "word": "高",
                      "start": 1.8000000000000003,
                      "end": 2.24,
                      "probability": 0.97119140625
                    },
                    {
                      "word": "级",
                      "start": 2.24,
                      "end": 2.52,
                      "probability": 0.9990234375
                    },
                    {
                      "word": "设",
                      "start": 2.52,
                      "end": 2.68,
                      "probability": 0.9990234375
                    },
                    {
                      "word": "置",
                      "start": 2.68,
                      "end": 2.76,
                      "probability": 0.99853515625
                    },
                    {
                      "word": "里",
                      "start": 2.76,
                      "end": 2.94,
                      "probability": 0.998046875
                    },
                    {
                      "word": "还",
                      "start": 2.94,
                      "end": 3.26,
                      "probability": 0.994140625
                    },
                    {
                      "word": "能",
                      "start": 3.26,
                      "end": 3.4,
                      "probability": 0.99951171875
                    },
                    {
                      "word": "添",
                      "start": 3.4,
                      "end": 3.56,
                      "probability": 0.9375
                    },
                    {
                      "word": "U",
                      "start": 3.56,
                      "end": 3.66,
                      "probability": 0.673828125
                    },
                    {
                      "word": "ser",
                      "start": 3.66,
                      "end": 3.9,
                      "probability": 0.99560546875
                    },
                    {
                      "word": " Agent",
                      "start": 3.9,
                      "end": 4.24,
                      "probability": 0.89697265625
                    },
                    {
                      "word": "和",
                      "start": 4.24,
                      "end": 4.74,
                      "probability": 0.98681640625
                    },
                    {
                      "word": "资",
                      "start": 4.74,
                      "end": 5.0,
                      "probability": 0.93603515625
                    },
                    {
                      "word": "金",
                      "start": 5.0,
                      "end": 5.08,
                      "probability": 0.78173828125
                    },
                    {
                      "word": "一起",
                      "start": 5.08,
                      "end": 5.22,
                      "probability": 0.79345703125
                    },
                    {
                      "word": "求",
                      "start": 5.22,
                      "end": 5.48,
                      "probability": 0.87109375
                    },
                    {
                      "word": "头",
                      "start": 5.48,
                      "end": 5.7,
                      "probability": 0.68115234375
                    }
                  ]
                },
                {
                  "id": 2,
                  "seek": 0,
                  "start": 5.7,
                  "end": 7.46,
                  "text": "不需要Key的服务",
                  "tokens": [
                    50657,
                    1960,
                    35748,
                    42,
                    2030,
                    1546,
                    27408,
                    5087,
                    94,
                    50740
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.1633597640104072,
                  "compression_ratio": 0.8921568627450981,
                  "no_speech_prob": 0.8626763820648193,
                  "words": [
                    {
                      "word": "不",
                      "start": 5.7,
                      "end": 6.48,
                      "probability": 0.9951171875
                    },
                    {
                      "word": "需要",
                      "start": 6.48,
                      "end": 6.76,
                      "probability": 0.99951171875
                    },
                    {
                      "word": "K",
                      "start": 6.76,
                      "end": 6.9,
                      "probability": 0.80810546875
                    },
                    {
                      "word": "ey",
                      "start": 6.9,
                      "end": 7.02,
                      "probability": 0.49072265625
                    },
                    {
                      "word": "的",
                      "start": 7.02,
                      "end": 7.08,
                      "probability": 0.9951171875
                    },
                    {
                      "word": "服",
                      "start": 7.08,
                      "end": 7.22,
                      "probability": 0.99951171875
                    },
                    {
                      "word": "务",
                      "start": 7.22,
                      "end": 7.46,
                      "probability": 1.0
                    }
                  ]
                }
              ],
              "language": "zh",
              "duration": 7.74
            },
            "status": "COMPLETE",
            "requested_language": "zh"
          },
          "suppression_proposed": null
        }
      ],
      "language": "zh",
      "packet_ids": [
        5
      ],
      "classification": "BEST_SUPPORTED_WITH_REVIEW",
      "independent_ASR": {
        "engine": "whisper-large-v3",
        "model_path": "/Users/sagawa/AI_WORK/cache/huggingface/hub/models--mlx-community--whisper-large-v3",
        "language": "zh",
        "audio_interval": [
          131.5,
          139.23999999999998
        ],
        "target_interval": [
          133.5,
          137.23999999999998
        ],
        "text": "高级设置里还能添UserAgent和资金一起求头",
        "words": [
          {
            "word": "高",
            "start": 133.3,
            "end": 133.74,
            "probability": 0.97119140625
          },
          {
            "word": "级",
            "start": 133.74,
            "end": 134.02,
            "probability": 0.9990234375
          },
          {
            "word": "设",
            "start": 134.02,
            "end": 134.18,
            "probability": 0.9990234375
          },
          {
            "word": "置",
            "start": 134.18,
            "end": 134.26,
            "probability": 0.99853515625
          },
          {
            "word": "里",
            "start": 134.26,
            "end": 134.44,
            "probability": 0.998046875
          },
          {
            "word": "还",
            "start": 134.44,
            "end": 134.76,
            "probability": 0.994140625
          },
          {
            "word": "能",
            "start": 134.76,
            "end": 134.9,
            "probability": 0.99951171875
          },
          {
            "word": "添",
            "start": 134.9,
            "end": 135.06,
            "probability": 0.9375
          },
          {
            "word": "U",
            "start": 135.06,
            "end": 135.16,
            "probability": 0.673828125
          },
          {
            "word": "ser",
            "start": 135.16,
            "end": 135.4,
            "probability": 0.99560546875
          },
          {
            "word": " Agent",
            "start": 135.4,
            "end": 135.74,
            "probability": 0.89697265625
          },
          {
            "word": "和",
            "start": 135.74,
            "end": 136.24,
            "probability": 0.98681640625
          },
          {
            "word": "资",
            "start": 136.24,
            "end": 136.5,
            "probability": 0.93603515625
          },
          {
            "word": "金",
            "start": 136.5,
            "end": 136.58,
            "probability": 0.78173828125
          },
          {
            "word": "一起",
            "start": 136.58,
            "end": 136.72,
            "probability": 0.79345703125
          },
          {
            "word": "求",
            "start": 136.72,
            "end": 136.98,
            "probability": 0.87109375
          },
          {
            "word": "头",
            "start": 136.98,
            "end": 137.2,
            "probability": 0.68115234375
          }
        ],
        "raw_result": {
          "text": "来降低延迟高级设置里还能添User Agent和资金一起求头不需要Key的服务",
          "segments": [
            {
              "id": 0,
              "seek": 0,
              "start": 0.0,
              "end": 1.32,
              "text": "来降低延迟",
              "tokens": [
                50365,
                6912,
                47421,
                41377,
                17527,
                114,
                3316,
                253,
                50436
              ],
              "temperature": 0.0,
              "avg_logprob": -0.1633597640104072,
              "compression_ratio": 0.8921568627450981,
              "no_speech_prob": 0.8626763820648193,
              "words": [
                {
                  "word": "来",
                  "start": 0.0,
                  "end": 0.44,
                  "probability": 0.9052734375
                },
                {
                  "word": "降",
                  "start": 0.44,
                  "end": 0.64,
                  "probability": 0.9970703125
                },
                {
                  "word": "低",
                  "start": 0.64,
                  "end": 0.86,
                  "probability": 0.9990234375
                },
                {
                  "word": "延",
                  "start": 0.86,
                  "end": 1.08,
                  "probability": 0.99853515625
                },
                {
                  "word": "迟",
                  "start": 1.08,
                  "end": 1.32,
                  "probability": 0.998046875
                }
              ]
            },
            {
              "id": 1,
              "seek": 0,
              "start": 1.8000000000000003,
              "end": 5.7,
              "text": "高级设置里还能添User Agent和资金一起求头",
              "tokens": [
                50436,
                12979,
                16853,
                100,
                7422,
                122,
                34719,
                15759,
                14852,
                8225,
                14176,
                119,
                52,
                12484,
                27174,
                12565,
                5266,
                226,
                19117,
                29567,
                32718,
                39862,
                50657
              ],
              "temperature": 0.0,
              "avg_logprob": -0.1633597640104072,
              "compression_ratio": 0.8921568627450981,
              "no_speech_prob": 0.8626763820648193,
              "words": [
                {
                  "word": "高",
                  "start": 1.8000000000000003,
                  "end": 2.24,
                  "probability": 0.97119140625
                },
                {
                  "word": "级",
                  "start": 2.24,
                  "end": 2.52,
                  "probability": 0.9990234375
                },
                {
                  "word": "设",
                  "start": 2.52,
                  "end": 2.68,
                  "probability": 0.9990234375
                },
                {
                  "word": "置",
                  "start": 2.68,
                  "end": 2.76,
                  "probability": 0.99853515625
                },
                {
                  "word": "里",
                  "start": 2.76,
                  "end": 2.94,
                  "probability": 0.998046875
                },
                {
                  "word": "还",
                  "start": 2.94,
                  "end": 3.26,
                  "probability": 0.994140625
                },
                {
                  "word": "能",
                  "start": 3.26,
                  "end": 3.4,
                  "probability": 0.99951171875
                },
                {
                  "word": "添",
                  "start": 3.4,
                  "end": 3.56,
                  "probability": 0.9375
                },
                {
                  "word": "U",
                  "start": 3.56,
                  "end": 3.66,
                  "probability": 0.673828125
                },
                {
                  "word": "ser",
                  "start": 3.66,
                  "end": 3.9,
                  "probability": 0.99560546875
                },
                {
                  "word": " Agent",
                  "start": 3.9,
                  "end": 4.24,
                  "probability": 0.89697265625
                },
                {
                  "word": "和",
                  "start": 4.24,
                  "end": 4.74,
                  "probability": 0.98681640625
                },
                {
                  "word": "资",
                  "start": 4.74,
                  "end": 5.0,
                  "probability": 0.93603515625
                },
                {
                  "word": "金",
                  "start": 5.0,
                  "end": 5.08,
                  "probability": 0.78173828125
                },
                {
                  "word": "一起",
                  "start": 5.08,
                  "end": 5.22,
                  "probability": 0.79345703125
                },
                {
                  "word": "求",
                  "start": 5.22,
                  "end": 5.48,
                  "probability": 0.87109375
                },
                {
                  "word": "头",
                  "start": 5.48,
                  "end": 5.7,
                  "probability": 0.68115234375
                }
              ]
            },
            {
              "id": 2,
              "seek": 0,
              "start": 5.7,
              "end": 7.46,
              "text": "不需要Key的服务",
              "tokens": [
                50657,
                1960,
                35748,
                42,
                2030,
                1546,
                27408,
                5087,
                94,
                50740
              ],
              "temperature": 0.0,
              "avg_logprob": -0.1633597640104072,
              "compression_ratio": 0.8921568627450981,
              "no_speech_prob": 0.8626763820648193,
              "words": [
                {
                  "word": "不",
                  "start": 5.7,
                  "end": 6.48,
                  "probability": 0.9951171875
                },
                {
                  "word": "需要",
                  "start": 6.48,
                  "end": 6.76,
                  "probability": 0.99951171875
                },
                {
                  "word": "K",
                  "start": 6.76,
                  "end": 6.9,
                  "probability": 0.80810546875
                },
                {
                  "word": "ey",
                  "start": 6.9,
                  "end": 7.02,
                  "probability": 0.49072265625
                },
                {
                  "word": "的",
                  "start": 7.02,
                  "end": 7.08,
                  "probability": 0.9951171875
                },
                {
                  "word": "服",
                  "start": 7.08,
                  "end": 7.22,
                  "probability": 0.99951171875
                },
                {
                  "word": "务",
                  "start": 7.22,
                  "end": 7.46,
                  "probability": 1.0
                }
              ]
            }
          ],
          "language": "zh",
          "duration": 7.74
        },
        "status": "COMPLETE",
        "requested_language": "zh"
      },
      "raw_word_timestamps": [
        {
          "word": "高",
          "start": 133.48,
          "end": 133.82,
          "probability": 0.8857421875
        },
        {
          "word": "级",
          "start": 133.82,
          "end": 134.01999999999998,
          "probability": 0.998046875
        },
        {
          "word": "设",
          "start": 134.01999999999998,
          "end": 134.17999999999998,
          "probability": 0.9951171875
        },
        {
          "word": "置",
          "start": 134.17999999999998,
          "end": 134.32,
          "probability": 0.99853515625
        },
        {
          "word": "里",
          "start": 134.32,
          "end": 134.48,
          "probability": 0.998046875
        },
        {
          "word": "还",
          "start": 134.48,
          "end": 134.79999999999998,
          "probability": 0.9912109375
        },
        {
          "word": "能",
          "start": 134.79999999999998,
          "end": 134.94,
          "probability": 0.9990234375
        },
        {
          "word": "填",
          "start": 134.94,
          "end": 135.07999999999998,
          "probability": 0.890625
        },
        {
          "word": "user",
          "start": 135.07999999999998,
          "end": 135.45999999999998,
          "probability": 0.52294921875
        },
        {
          "word": " agent",
          "start": 135.45999999999998,
          "end": 135.92,
          "probability": 0.91357421875
        },
        {
          "word": "和",
          "start": 135.92,
          "end": 136.33999999999997,
          "probability": 0.98486328125
        },
        {
          "word": "资",
          "start": 136.33999999999997,
          "end": 136.48,
          "probability": 0.787109375
        },
        {
          "word": "金",
          "start": 136.48,
          "end": 136.6,
          "probability": 0.81201171875
        },
        {
          "word": "一",
          "start": 136.6,
          "end": 136.73999999999998,
          "probability": 0.72607421875
        },
        {
          "word": "请",
          "start": 136.73999999999998,
          "end": 136.88,
          "probability": 0.85205078125
        },
        {
          "word": "求",
          "start": 136.88,
          "end": 137.04,
          "probability": 0.98583984375
        },
        {
          "word": "头",
          "start": 137.04,
          "end": 137.23999999999998,
          "probability": 0.39404296875
        }
      ],
      "word_timestamps": [],
      "source_text": "资金一请求头",
      "review_status": "BEST_SUPPORTED_WITH_REVIEW",
      "canonical_timing_evidence": "CUE_ANCHORED",
      "index": 63
    },
    {
      "start": 137.73999999999998,
      "end": 140.26,
      "text": "不需要Key的服务把Key留空就行",
      "raw_asr_text": "不需要Key的服务把Key留空就行",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 5,
      "words": [
        {
          "word": "不",
          "start": 137.73999999999998,
          "end": 138.07999999999998,
          "probability": 0.9619140625
        },
        {
          "word": "需要",
          "start": 138.07999999999998,
          "end": 138.32,
          "probability": 0.9990234375
        },
        {
          "word": "K",
          "start": 138.32,
          "end": 138.5,
          "probability": 0.509765625
        },
        {
          "word": "ey",
          "start": 138.5,
          "end": 138.54,
          "probability": 0.5029296875
        },
        {
          "word": "的",
          "start": 138.54,
          "end": 138.67999999999998,
          "probability": 0.9921875
        },
        {
          "word": "服",
          "start": 138.67999999999998,
          "end": 138.79999999999998,
          "probability": 1.0
        },
        {
          "word": "务",
          "start": 138.79999999999998,
          "end": 139.0,
          "probability": 1.0
        },
        {
          "word": "把",
          "start": 139.0,
          "end": 139.32,
          "probability": 0.97900390625
        },
        {
          "word": "K",
          "start": 139.32,
          "end": 139.45999999999998,
          "probability": 0.96728515625
        },
        {
          "word": "ey",
          "start": 139.45999999999998,
          "end": 139.54,
          "probability": 1.0
        },
        {
          "word": "留",
          "start": 139.54,
          "end": 139.66,
          "probability": 0.95068359375
        },
        {
          "word": "空",
          "start": 139.66,
          "end": 139.83999999999997,
          "probability": 0.97021484375
        },
        {
          "word": "就",
          "start": 139.83999999999997,
          "end": 140.0,
          "probability": 0.99951171875
        },
        {
          "word": "行",
          "start": 140.0,
          "end": 140.26,
          "probability": 0.994140625
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.14073291240922556,
        "compression_ratio": 1.1727574750830565,
        "no_speech_prob": 2.7202724448005533e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "高级设置里还能填user agent和资金一请求头不需要Key的服务把Key留空就行另外设置里还提供了一个可选的Javi在线辅助功能打开之后可以帮你做智能断句和一文检查所以平时用起来其实就三步填好你的大模型API并测试连接允许权限然后点一下开始字幕就可以安心看直播了"
        }
      ],
      "id": 64,
      "review_annotation": "",
      "row_id": "64",
      "bounded_row_id": 64,
      "parent_row_ids": [
        64
      ],
      "parent_row_id": 64,
      "packet_id": 5,
      "LID_evidence": [
        {
          "region_id": 25,
          "probabilities": {
            "zh": 0.8050072607066896
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        25
      ],
      "language_confidence": 0.8050072607066896,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 132.7,
      "packet_end": 161.3,
      "raw_text": "不需要Key的服务把Key留空就行",
      "selected_text": "不需要Key的服务把Key留空就行",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            64
          ],
          "packet_id": 5,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        5
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 64
    },
    {
      "start": 141.48,
      "end": 143.89999999999998,
      "text": "另外设置里还提供了一个可选的Javi",
      "raw_asr_text": "另外设置里还提供了一个可选的Javi",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 5,
      "words": [
        {
          "word": "另外",
          "start": 141.48,
          "end": 141.82,
          "probability": 0.998046875
        },
        {
          "word": "设",
          "start": 141.82,
          "end": 142.16,
          "probability": 0.947265625
        },
        {
          "word": "置",
          "start": 142.16,
          "end": 142.29999999999998,
          "probability": 1.0
        },
        {
          "word": "里",
          "start": 142.29999999999998,
          "end": 142.42,
          "probability": 1.0
        },
        {
          "word": "还",
          "start": 142.42,
          "end": 142.54,
          "probability": 1.0
        },
        {
          "word": "提",
          "start": 142.54,
          "end": 142.72,
          "probability": 0.99951171875
        },
        {
          "word": "供",
          "start": 142.72,
          "end": 142.83999999999997,
          "probability": 0.99951171875
        },
        {
          "word": "了",
          "start": 142.83999999999997,
          "end": 142.94,
          "probability": 0.9990234375
        },
        {
          "word": "一个",
          "start": 142.94,
          "end": 143.1,
          "probability": 0.9990234375
        },
        {
          "word": "可",
          "start": 143.1,
          "end": 143.29999999999998,
          "probability": 0.99951171875
        },
        {
          "word": "选",
          "start": 143.29999999999998,
          "end": 143.44,
          "probability": 1.0
        },
        {
          "word": "的",
          "start": 143.44,
          "end": 143.57999999999998,
          "probability": 0.99951171875
        },
        {
          "word": "J",
          "start": 143.57999999999998,
          "end": 143.72,
          "probability": 0.775390625
        },
        {
          "word": "avi",
          "start": 143.72,
          "end": 143.89999999999998,
          "probability": 0.935546875
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.14073291240922556,
        "compression_ratio": 1.1727574750830565,
        "no_speech_prob": 2.7202724448005533e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "高级设置里还能填user agent和资金一请求头不需要Key的服务把Key留空就行另外设置里还提供了一个可选的Javi在线辅助功能打开之后可以帮你做智能断句和一文检查所以平时用起来其实就三步填好你的大模型API并测试连接允许权限然后点一下开始字幕就可以安心看直播了"
        }
      ],
      "id": 65,
      "review_annotation": "待確認",
      "row_id": "65",
      "bounded_row_id": 65,
      "parent_row_ids": [
        65
      ],
      "parent_row_id": 65,
      "packet_id": 5,
      "LID_evidence": [
        {
          "region_id": 26,
          "probabilities": {
            "zh": 0.9087536748912599
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        26
      ],
      "language_confidence": 0.9087536748912599,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 132.7,
      "packet_end": 161.3,
      "raw_text": "另外设置里还提供了一个可选的Javi",
      "selected_text": "另外设置里还提供了一个可选的Javi",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            65
          ],
          "packet_id": 5,
          "engine": "mlx-whisper"
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "65",
          "before": "另外设置里还提供了一个可选的Javi",
          "after": "另外设置里还提供了一个可选的Javi",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "Javi疑似Java或Java API误识",
          "context_evidence": "上下文为在线辅助功能，可能指Java相关",
          "independent_ASR": {
            "status": "REVIEW_ONLY_RESCUE_BUDGET"
          },
          "suppression_proposed": null
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "65",
          "before": "另外设置里还提供了一个可选的Javi",
          "after": "另外设置里还提供了一个可选的Javi",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "Name error: 'Javi' vs 'Java'",
          "context_evidence": "Context: '可选的Java'",
          "independent_ASR": {
            "status": "REVIEW_ONLY_RESCUE_BUDGET"
          },
          "suppression_proposed": null
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "65",
          "before": "另外设置里还提供了一个可选的Javi",
          "after": "另外设置里还提供了一个可选的Javi",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "Name error: 'Javi' vs 'Java'",
          "context_evidence": "Context: '可选的Java'",
          "independent_ASR": {
            "status": "REVIEW_ONLY_RESCUE_BUDGET"
          },
          "suppression_proposed": null
        }
      ],
      "language": "zh",
      "packet_ids": [
        5
      ],
      "classification": "BEST_SUPPORTED_WITH_REVIEW",
      "independent_ASR": {
        "status": "REVIEW_ONLY_RESCUE_BUDGET"
      },
      "source_text": "另外设置里还提供了一个可选的Javi",
      "review_status": "BEST_SUPPORTED_WITH_REVIEW",
      "canonical_timing_evidence": "CUE_ANCHORED",
      "index": 65
    },
    {
      "start": 144.45999999999998,
      "end": 145.7,
      "text": "在线辅助功能",
      "raw_asr_text": "在线辅助功能",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 5,
      "words": [
        {
          "word": "在",
          "start": 144.45999999999998,
          "end": 144.79999999999998,
          "probability": 0.95703125
        },
        {
          "word": "线",
          "start": 144.79999999999998,
          "end": 145.0,
          "probability": 1.0
        },
        {
          "word": "辅",
          "start": 145.0,
          "end": 145.17999999999998,
          "probability": 0.98876953125
        },
        {
          "word": "助",
          "start": 145.17999999999998,
          "end": 145.32,
          "probability": 0.99951171875
        },
        {
          "word": "功",
          "start": 145.32,
          "end": 145.51999999999998,
          "probability": 0.99853515625
        },
        {
          "word": "能",
          "start": 145.51999999999998,
          "end": 145.7,
          "probability": 0.99951171875
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.14073291240922556,
        "compression_ratio": 1.1727574750830565,
        "no_speech_prob": 2.7202724448005533e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "高级设置里还能填user agent和资金一请求头不需要Key的服务把Key留空就行另外设置里还提供了一个可选的Javi在线辅助功能打开之后可以帮你做智能断句和一文检查所以平时用起来其实就三步填好你的大模型API并测试连接允许权限然后点一下开始字幕就可以安心看直播了"
        }
      ],
      "id": 66,
      "review_annotation": "",
      "row_id": "66",
      "bounded_row_id": 66,
      "parent_row_ids": [
        66
      ],
      "parent_row_id": 66,
      "packet_id": 5,
      "LID_evidence": [
        {
          "region_id": 27,
          "probabilities": {
            "zh": 0.9976596747125898
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        27
      ],
      "language_confidence": 0.9976596747125898,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 132.7,
      "packet_end": 161.3,
      "raw_text": "在线辅助功能",
      "selected_text": "在线辅助功能",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            66
          ],
          "packet_id": 5,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        5
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 66
    },
    {
      "start": 145.7,
      "end": 149.16,
      "text": "一文检查",
      "raw_asr_text": "打开之后可以帮你做智能断句和一文检查",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 5,
      "words": [],
      "asr_metrics": {
        "avg_logprob": -0.14073291240922556,
        "compression_ratio": 1.1727574750830565,
        "no_speech_prob": 2.7202724448005533e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "高级设置里还能填user agent和资金一请求头不需要Key的服务把Key留空就行另外设置里还提供了一个可选的Javi在线辅助功能打开之后可以帮你做智能断句和一文检查所以平时用起来其实就三步填好你的大模型API并测试连接允许权限然后点一下开始字幕就可以安心看直播了"
        }
      ],
      "id": 67,
      "review_annotation": "待確認",
      "row_id": "67",
      "bounded_row_id": 67,
      "parent_row_ids": [
        67
      ],
      "parent_row_id": 67,
      "packet_id": 5,
      "LID_evidence": [
        {
          "region_id": 27,
          "probabilities": {
            "zh": 0.9976596747125898
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        27
      ],
      "language_confidence": 0.9976596747125898,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 132.7,
      "packet_end": 161.3,
      "raw_text": "打开之后可以帮你做智能断句和一文检查",
      "selected_text": "一文检查",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            67
          ],
          "packet_id": 5,
          "engine": "mlx-whisper"
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "67",
          "before": "打开之后可以帮你做智能断句和一文检查",
          "after": "一文检查",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "一文检查疑似文案检查误识",
          "context_evidence": "上下文为智能辅助，'文案'更符合逻辑",
          "independent_ASR": {
            "status": "REVIEW_ONLY_RESCUE_BUDGET"
          },
          "suppression_proposed": null
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "67",
          "before": "一文检查",
          "after": "一文检查",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "Nonsense: '一文检查' vs '语法检查' (grammar check)",
          "context_evidence": "Context: '语法检查'",
          "independent_ASR": {
            "status": "REVIEW_ONLY_RESCUE_BUDGET"
          },
          "suppression_proposed": null
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "67",
          "before": "一文检查",
          "after": "一文检查",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "Nonsense: '一文检查' vs '语法检查' (grammar check)",
          "context_evidence": "Context: '语法检查'",
          "independent_ASR": {
            "status": "REVIEW_ONLY_RESCUE_BUDGET"
          },
          "suppression_proposed": null
        }
      ],
      "language": "zh",
      "packet_ids": [
        5
      ],
      "classification": "BEST_SUPPORTED_WITH_REVIEW",
      "independent_ASR": {
        "status": "REVIEW_ONLY_RESCUE_BUDGET"
      },
      "raw_word_timestamps": [
        {
          "word": "打",
          "start": 145.7,
          "end": 146.23999999999998,
          "probability": 0.99169921875
        },
        {
          "word": "开",
          "start": 146.23999999999998,
          "end": 146.38,
          "probability": 0.99951171875
        },
        {
          "word": "之",
          "start": 146.38,
          "end": 146.56,
          "probability": 0.99951171875
        },
        {
          "word": "后",
          "start": 146.56,
          "end": 146.67999999999998,
          "probability": 1.0
        },
        {
          "word": "可以",
          "start": 146.67999999999998,
          "end": 147.1,
          "probability": 0.9990234375
        },
        {
          "word": "帮",
          "start": 147.1,
          "end": 147.23999999999998,
          "probability": 1.0
        },
        {
          "word": "你",
          "start": 147.23999999999998,
          "end": 147.38,
          "probability": 0.99951171875
        },
        {
          "word": "做",
          "start": 147.38,
          "end": 147.54,
          "probability": 0.9990234375
        },
        {
          "word": "智",
          "start": 147.54,
          "end": 147.7,
          "probability": 0.97802734375
        },
        {
          "word": "能",
          "start": 147.7,
          "end": 147.82,
          "probability": 0.99169921875
        },
        {
          "word": "断",
          "start": 147.82,
          "end": 148.0,
          "probability": 0.9462890625
        },
        {
          "word": "句",
          "start": 148.0,
          "end": 148.16,
          "probability": 0.44140625
        },
        {
          "word": "和",
          "start": 148.16,
          "end": 148.45999999999998,
          "probability": 0.99951171875
        },
        {
          "word": "一",
          "start": 148.45999999999998,
          "end": 148.57999999999998,
          "probability": 0.59130859375
        },
        {
          "word": "文",
          "start": 148.57999999999998,
          "end": 148.7,
          "probability": 0.96484375
        },
        {
          "word": "检",
          "start": 148.7,
          "end": 148.89999999999998,
          "probability": 0.99951171875
        },
        {
          "word": "查",
          "start": 148.89999999999998,
          "end": 149.16,
          "probability": 0.9970703125
        }
      ],
      "word_timestamps": [],
      "source_text": "一文检查",
      "review_status": "BEST_SUPPORTED_WITH_REVIEW",
      "canonical_timing_evidence": "CUE_ANCHORED",
      "index": 67
    },
    {
      "start": 150.14,
      "end": 151.94,
      "text": "所以平时用起来其实就三步",
      "raw_asr_text": "所以平时用起来其实就三步",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 5,
      "words": [
        {
          "word": "所以",
          "start": 150.14,
          "end": 150.48,
          "probability": 0.98828125
        },
        {
          "word": "平",
          "start": 150.48,
          "end": 150.64,
          "probability": 0.99951171875
        },
        {
          "word": "时",
          "start": 150.64,
          "end": 150.77999999999997,
          "probability": 0.998046875
        },
        {
          "word": "用",
          "start": 150.77999999999997,
          "end": 150.92,
          "probability": 0.99951171875
        },
        {
          "word": "起",
          "start": 150.92,
          "end": 151.04,
          "probability": 1.0
        },
        {
          "word": "来",
          "start": 151.04,
          "end": 151.16,
          "probability": 1.0
        },
        {
          "word": "其实",
          "start": 151.16,
          "end": 151.39999999999998,
          "probability": 0.994140625
        },
        {
          "word": "就",
          "start": 151.39999999999998,
          "end": 151.57999999999998,
          "probability": 0.99365234375
        },
        {
          "word": "三",
          "start": 151.57999999999998,
          "end": 151.7,
          "probability": 0.9560546875
        },
        {
          "word": "步",
          "start": 151.7,
          "end": 151.94,
          "probability": 0.955078125
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.14073291240922556,
        "compression_ratio": 1.1727574750830565,
        "no_speech_prob": 2.7202724448005533e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "高级设置里还能填user agent和资金一请求头不需要Key的服务把Key留空就行另外设置里还提供了一个可选的Javi在线辅助功能打开之后可以帮你做智能断句和一文检查所以平时用起来其实就三步填好你的大模型API并测试连接允许权限然后点一下开始字幕就可以安心看直播了"
        }
      ],
      "id": 68,
      "review_annotation": "",
      "row_id": "68",
      "bounded_row_id": 68,
      "parent_row_ids": [
        68
      ],
      "parent_row_id": 68,
      "packet_id": 5,
      "LID_evidence": [
        {
          "region_id": 28,
          "probabilities": {
            "zh": 0.9981437474489212
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        28
      ],
      "language_confidence": 0.9981437474489212,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 132.7,
      "packet_end": 161.3,
      "raw_text": "所以平时用起来其实就三步",
      "selected_text": "所以平时用起来其实就三步",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            68
          ],
          "packet_id": 5,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        5
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 68
    },
    {
      "start": 152.28,
      "end": 155.66,
      "text": "填好你的大模型API并测试连接允许权限",
      "raw_asr_text": "填好你的大模型API并测试连接允许权限",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 5,
      "words": [
        {
          "word": "填",
          "start": 152.28,
          "end": 152.62,
          "probability": 0.99658203125
        },
        {
          "word": "好",
          "start": 152.62,
          "end": 152.82,
          "probability": 0.99853515625
        },
        {
          "word": "你的",
          "start": 152.82,
          "end": 153.01999999999998,
          "probability": 0.99951171875
        },
        {
          "word": "大",
          "start": 153.01999999999998,
          "end": 153.17999999999998,
          "probability": 1.0
        },
        {
          "word": "模",
          "start": 153.17999999999998,
          "end": 153.32,
          "probability": 0.99853515625
        },
        {
          "word": "型",
          "start": 153.32,
          "end": 153.5,
          "probability": 0.9990234375
        },
        {
          "word": "AP",
          "start": 153.5,
          "end": 153.76,
          "probability": 0.94580078125
        },
        {
          "word": "I",
          "start": 153.76,
          "end": 154.01999999999998,
          "probability": 1.0
        },
        {
          "word": "并",
          "start": 154.01999999999998,
          "end": 154.29999999999998,
          "probability": 0.9990234375
        },
        {
          "word": "测",
          "start": 154.29999999999998,
          "end": 154.42,
          "probability": 0.99951171875
        },
        {
          "word": "试",
          "start": 154.42,
          "end": 154.56,
          "probability": 0.9990234375
        },
        {
          "word": "连",
          "start": 154.56,
          "end": 154.67999999999998,
          "probability": 0.93017578125
        },
        {
          "word": "接",
          "start": 154.67999999999998,
          "end": 154.88,
          "probability": 0.9970703125
        },
        {
          "word": "允",
          "start": 154.88,
          "end": 155.2,
          "probability": 0.986328125
        },
        {
          "word": "许",
          "start": 155.2,
          "end": 155.32,
          "probability": 0.99951171875
        },
        {
          "word": "权",
          "start": 155.32,
          "end": 155.48,
          "probability": 0.9150390625
        },
        {
          "word": "限",
          "start": 155.48,
          "end": 155.66,
          "probability": 0.9833984375
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.14073291240922556,
        "compression_ratio": 1.1727574750830565,
        "no_speech_prob": 2.7202724448005533e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "高级设置里还能填user agent和资金一请求头不需要Key的服务把Key留空就行另外设置里还提供了一个可选的Javi在线辅助功能打开之后可以帮你做智能断句和一文检查所以平时用起来其实就三步填好你的大模型API并测试连接允许权限然后点一下开始字幕就可以安心看直播了"
        }
      ],
      "id": 69,
      "review_annotation": "",
      "row_id": "69",
      "bounded_row_id": 69,
      "parent_row_ids": [
        69
      ],
      "parent_row_id": 69,
      "packet_id": 5,
      "LID_evidence": [
        {
          "region_id": 29,
          "probabilities": {
            "zh": 0.9979983766873678
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        29
      ],
      "language_confidence": 0.9979983766873678,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 132.7,
      "packet_end": 161.3,
      "raw_text": "填好你的大模型API并测试连接允许权限",
      "selected_text": "填好你的大模型API并测试连接允许权限",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            69
          ],
          "packet_id": 5,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        5
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 69
    },
    {
      "start": 155.66,
      "end": 157.62,
      "text": "然后点一下开始字幕",
      "raw_asr_text": "然后点一下开始字幕",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 5,
      "words": [
        {
          "word": "然后",
          "start": 155.66,
          "end": 156.16,
          "probability": 0.99658203125
        },
        {
          "word": "点",
          "start": 156.16,
          "end": 156.57999999999998,
          "probability": 0.99853515625
        },
        {
          "word": "一下",
          "start": 156.57999999999998,
          "end": 156.88,
          "probability": 0.99609375
        },
        {
          "word": "开始",
          "start": 156.88,
          "end": 157.27999999999997,
          "probability": 0.9951171875
        },
        {
          "word": "字",
          "start": 157.27999999999997,
          "end": 157.44,
          "probability": 0.98583984375
        },
        {
          "word": "幕",
          "start": 157.44,
          "end": 157.62,
          "probability": 0.98974609375
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.14073291240922556,
        "compression_ratio": 1.1727574750830565,
        "no_speech_prob": 2.7202724448005533e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "高级设置里还能填user agent和资金一请求头不需要Key的服务把Key留空就行另外设置里还提供了一个可选的Javi在线辅助功能打开之后可以帮你做智能断句和一文检查所以平时用起来其实就三步填好你的大模型API并测试连接允许权限然后点一下开始字幕就可以安心看直播了"
        }
      ],
      "id": 70,
      "review_annotation": "",
      "row_id": "70",
      "bounded_row_id": 70,
      "parent_row_ids": [
        70
      ],
      "parent_row_id": 70,
      "packet_id": 5,
      "LID_evidence": [
        {
          "region_id": 29,
          "probabilities": {
            "zh": 0.9979983766873678
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        29
      ],
      "language_confidence": 0.9979983766873678,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 132.7,
      "packet_end": 161.3,
      "raw_text": "然后点一下开始字幕",
      "selected_text": "然后点一下开始字幕",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            70
          ],
          "packet_id": 5,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        5
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 70
    },
    {
      "start": 158.12,
      "end": 159.62,
      "text": "就可以安心看直播了",
      "raw_asr_text": "就可以安心看直播了",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 5,
      "words": [
        {
          "word": "就可以",
          "start": 158.12,
          "end": 158.45999999999998,
          "probability": 0.998046875
        },
        {
          "word": "安",
          "start": 158.45999999999998,
          "end": 158.79999999999998,
          "probability": 0.99951171875
        },
        {
          "word": "心",
          "start": 158.79999999999998,
          "end": 159.0,
          "probability": 0.99951171875
        },
        {
          "word": "看",
          "start": 159.0,
          "end": 159.17999999999998,
          "probability": 0.99951171875
        },
        {
          "word": "直",
          "start": 159.17999999999998,
          "end": 159.29999999999998,
          "probability": 0.99755859375
        },
        {
          "word": "播",
          "start": 159.29999999999998,
          "end": 159.45999999999998,
          "probability": 1.0
        },
        {
          "word": "了",
          "start": 159.45999999999998,
          "end": 159.62,
          "probability": 0.99853515625
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.14073291240922556,
        "compression_ratio": 1.1727574750830565,
        "no_speech_prob": 2.7202724448005533e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "高级设置里还能填user agent和资金一请求头不需要Key的服务把Key留空就行另外设置里还提供了一个可选的Javi在线辅助功能打开之后可以帮你做智能断句和一文检查所以平时用起来其实就三步填好你的大模型API并测试连接允许权限然后点一下开始字幕就可以安心看直播了"
        }
      ],
      "id": 71,
      "review_annotation": "",
      "row_id": "71",
      "bounded_row_id": 71,
      "parent_row_ids": [
        71
      ],
      "parent_row_id": 71,
      "packet_id": 5,
      "LID_evidence": [
        {
          "region_id": 30,
          "probabilities": {
            "zh": 0.9977339982986451
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        30
      ],
      "language_confidence": 0.9977339982986451,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 132.7,
      "packet_end": 161.3,
      "raw_text": "就可以安心看直播了",
      "selected_text": "就可以安心看直播了",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            71
          ],
          "packet_id": 5,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        5
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 71
    },
    {
      "start": 164.0,
      "end": 165.64,
      "text": "老朋友可能已经发现了",
      "raw_asr_text": "老朋友可能已经发现了",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 6,
      "words": [
        {
          "word": "老",
          "start": 163.88,
          "end": 164.23999999999998,
          "probability": 0.970703125
        },
        {
          "word": "朋友",
          "start": 164.23999999999998,
          "end": 164.44,
          "probability": 1.0
        },
        {
          "word": "可能",
          "start": 164.44,
          "end": 164.72,
          "probability": 0.99951171875
        },
        {
          "word": "已经",
          "start": 164.72,
          "end": 164.95999999999998,
          "probability": 0.9990234375
        },
        {
          "word": "发",
          "start": 164.95999999999998,
          "end": 165.16,
          "probability": 1.0
        },
        {
          "word": "现",
          "start": 165.16,
          "end": 165.42,
          "probability": 0.99951171875
        },
        {
          "word": "了",
          "start": 165.42,
          "end": 165.64,
          "probability": 1.0
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.12088049108331854,
        "compression_ratio": 1.105263157894737,
        "no_speech_prob": 1.2315977231114328e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "老朋友可能已经发现了1.0版本除了加新功能我还做了几个减法首先是移除了之前内置的index免费翻译它是B占官方提供的接口用的人一多就经常被挤爆不稳定翻译效果其实也不如其他主流大模型"
        }
      ],
      "id": 72,
      "review_annotation": "",
      "row_id": "72",
      "bounded_row_id": 72,
      "parent_row_ids": [
        72
      ],
      "parent_row_id": 72,
      "packet_id": 6,
      "LID_evidence": [
        {
          "region_id": 31,
          "probabilities": {
            "zh": 0.9979968965053558
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        31
      ],
      "language_confidence": 0.9979968965053558,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 163.2,
      "packet_end": 185.70000000000002,
      "raw_text": "老朋友可能已经发现了",
      "selected_text": "老朋友可能已经发现了",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            72
          ],
          "packet_id": 6,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        6
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 72
    },
    {
      "start": 166.29999999999998,
      "end": 168.78,
      "text": "1.0版本除了加新功能",
      "raw_asr_text": "1.0版本除了加新功能",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 6,
      "words": [
        {
          "word": "1.",
          "start": 166.29999999999998,
          "end": 166.66,
          "probability": 0.7783203125
        },
        {
          "word": "0",
          "start": 166.82,
          "end": 167.0,
          "probability": 0.99951171875
        },
        {
          "word": "版",
          "start": 167.0,
          "end": 167.2,
          "probability": 0.99658203125
        },
        {
          "word": "本",
          "start": 167.2,
          "end": 167.48,
          "probability": 1.0
        },
        {
          "word": "除",
          "start": 167.48,
          "end": 167.88,
          "probability": 0.9736328125
        },
        {
          "word": "了",
          "start": 167.88,
          "end": 168.0,
          "probability": 1.0
        },
        {
          "word": "加",
          "start": 168.0,
          "end": 168.16,
          "probability": 0.99365234375
        },
        {
          "word": "新",
          "start": 168.16,
          "end": 168.39999999999998,
          "probability": 0.96875
        },
        {
          "word": "功",
          "start": 168.39999999999998,
          "end": 168.57999999999998,
          "probability": 0.99853515625
        },
        {
          "word": "能",
          "start": 168.57999999999998,
          "end": 168.78,
          "probability": 1.0
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.12088049108331854,
        "compression_ratio": 1.105263157894737,
        "no_speech_prob": 1.2315977231114328e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "老朋友可能已经发现了1.0版本除了加新功能我还做了几个减法首先是移除了之前内置的index免费翻译它是B占官方提供的接口用的人一多就经常被挤爆不稳定翻译效果其实也不如其他主流大模型"
        }
      ],
      "id": 73,
      "review_annotation": "",
      "row_id": "73",
      "bounded_row_id": 73,
      "parent_row_ids": [
        73
      ],
      "parent_row_id": 73,
      "packet_id": 6,
      "LID_evidence": [
        {
          "region_id": 32,
          "probabilities": {
            "zh": 0.9978189468383789
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        32
      ],
      "language_confidence": 0.9978189468383789,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 163.2,
      "packet_end": 185.70000000000002,
      "raw_text": "1.0版本除了加新功能",
      "selected_text": "1.0版本除了加新功能",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            73
          ],
          "packet_id": 6,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        6
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 73
    },
    {
      "start": 169.33999999999997,
      "end": 170.82,
      "text": "我还做了几个减法",
      "raw_asr_text": "我还做了几个减法",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 6,
      "words": [
        {
          "word": "我",
          "start": 169.33999999999997,
          "end": 169.7,
          "probability": 0.97705078125
        },
        {
          "word": "还",
          "start": 169.7,
          "end": 169.85999999999999,
          "probability": 1.0
        },
        {
          "word": "做",
          "start": 169.85999999999999,
          "end": 170.01999999999998,
          "probability": 0.9990234375
        },
        {
          "word": "了",
          "start": 170.01999999999998,
          "end": 170.14,
          "probability": 1.0
        },
        {
          "word": "几",
          "start": 170.14,
          "end": 170.28,
          "probability": 1.0
        },
        {
          "word": "个",
          "start": 170.28,
          "end": 170.39999999999998,
          "probability": 1.0
        },
        {
          "word": "减",
          "start": 170.39999999999998,
          "end": 170.57999999999998,
          "probability": 0.9892578125
        },
        {
          "word": "法",
          "start": 170.57999999999998,
          "end": 170.82,
          "probability": 0.9970703125
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.12088049108331854,
        "compression_ratio": 1.105263157894737,
        "no_speech_prob": 1.2315977231114328e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "老朋友可能已经发现了1.0版本除了加新功能我还做了几个减法首先是移除了之前内置的index免费翻译它是B占官方提供的接口用的人一多就经常被挤爆不稳定翻译效果其实也不如其他主流大模型"
        }
      ],
      "id": 74,
      "review_annotation": "",
      "row_id": "74",
      "bounded_row_id": 74,
      "parent_row_ids": [
        74
      ],
      "parent_row_id": 74,
      "packet_id": 6,
      "LID_evidence": [
        {
          "region_id": 33,
          "probabilities": {
            "zh": 0.9972352623939514
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        33
      ],
      "language_confidence": 0.9972352623939514,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 163.2,
      "packet_end": 185.70000000000002,
      "raw_text": "我还做了几个减法",
      "selected_text": "我还做了几个减法",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            74
          ],
          "packet_id": 6,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        6
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 74
    },
    {
      "start": 171.85999999999999,
      "end": 175.1,
      "text": "首先是移除了之前内置的index免费翻译",
      "raw_asr_text": "首先是移除了之前内置的index免费翻译",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 6,
      "words": [
        {
          "word": "首先",
          "start": 171.85999999999999,
          "end": 172.22,
          "probability": 0.99951171875
        },
        {
          "word": "是",
          "start": 172.22,
          "end": 172.57999999999998,
          "probability": 0.94482421875
        },
        {
          "word": "移",
          "start": 172.57999999999998,
          "end": 172.72,
          "probability": 0.9892578125
        },
        {
          "word": "除",
          "start": 172.72,
          "end": 172.88,
          "probability": 0.9990234375
        },
        {
          "word": "了",
          "start": 172.88,
          "end": 173.01999999999998,
          "probability": 1.0
        },
        {
          "word": "之前",
          "start": 173.01999999999998,
          "end": 173.29999999999998,
          "probability": 0.99951171875
        },
        {
          "word": "内",
          "start": 173.29999999999998,
          "end": 173.5,
          "probability": 0.64501953125
        },
        {
          "word": "置",
          "start": 173.5,
          "end": 173.61999999999998,
          "probability": 0.994140625
        },
        {
          "word": "的",
          "start": 173.61999999999998,
          "end": 173.72,
          "probability": 1.0
        },
        {
          "word": "ind",
          "start": 173.72,
          "end": 173.94,
          "probability": 0.80810546875
        },
        {
          "word": "ex",
          "start": 173.94,
          "end": 174.16,
          "probability": 0.99853515625
        },
        {
          "word": "免",
          "start": 174.16,
          "end": 174.51999999999998,
          "probability": 0.9970703125
        },
        {
          "word": "费",
          "start": 174.51999999999998,
          "end": 174.7,
          "probability": 1.0
        },
        {
          "word": "翻",
          "start": 174.7,
          "end": 174.85999999999999,
          "probability": 0.99951171875
        },
        {
          "word": "译",
          "start": 174.85999999999999,
          "end": 175.1,
          "probability": 0.9990234375
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.12088049108331854,
        "compression_ratio": 1.105263157894737,
        "no_speech_prob": 1.2315977231114328e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "老朋友可能已经发现了1.0版本除了加新功能我还做了几个减法首先是移除了之前内置的index免费翻译它是B占官方提供的接口用的人一多就经常被挤爆不稳定翻译效果其实也不如其他主流大模型"
        }
      ],
      "id": 75,
      "review_annotation": "",
      "row_id": "75",
      "bounded_row_id": 75,
      "parent_row_ids": [
        75
      ],
      "parent_row_id": 75,
      "packet_id": 6,
      "LID_evidence": [
        {
          "region_id": 33,
          "probabilities": {
            "zh": 0.9972352623939514
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        33
      ],
      "language_confidence": 0.9972352623939514,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 163.2,
      "packet_end": 185.70000000000002,
      "raw_text": "首先是移除了之前内置的index免费翻译",
      "selected_text": "首先是移除了之前内置的index免费翻译",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            75
          ],
          "packet_id": 6,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        6
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 75
    },
    {
      "start": 175.1,
      "end": 176.07999999999998,
      "text": "它是B",
      "raw_asr_text": "它是B",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 6,
      "words": [
        {
          "word": "它",
          "start": 175.1,
          "end": 175.76,
          "probability": 0.98974609375
        },
        {
          "word": "是",
          "start": 175.76,
          "end": 175.89999999999998,
          "probability": 1.0
        },
        {
          "word": "B",
          "start": 175.89999999999998,
          "end": 176.07999999999998,
          "probability": 0.55322265625
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.12088049108331854,
        "compression_ratio": 1.105263157894737,
        "no_speech_prob": 1.2315977231114328e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "老朋友可能已经发现了1.0版本除了加新功能我还做了几个减法首先是移除了之前内置的index免费翻译它是B占官方提供的接口用的人一多就经常被挤爆不稳定翻译效果其实也不如其他主流大模型"
        }
      ],
      "id": 76,
      "review_annotation": "待確認",
      "row_id": "76",
      "bounded_row_id": 76,
      "parent_row_ids": [
        76
      ],
      "parent_row_id": 76,
      "packet_id": 6,
      "LID_evidence": [
        {
          "region_id": 33,
          "probabilities": {
            "zh": 0.9972352623939514
          },
          "status": "STABLE"
        },
        {
          "region_id": 34,
          "probabilities": {
            "zh": 0.9976210514704387
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        33,
        34
      ],
      "language_confidence": 0.9976210514704387,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 163.2,
      "packet_end": 185.70000000000002,
      "raw_text": "它是B",
      "selected_text": "它是B",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            76
          ],
          "packet_id": 6,
          "engine": "mlx-whisper"
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "76",
          "before": "它是B",
          "after": "它是B",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "它是B疑似它是免费或它是B站误识",
          "context_evidence": "上下文为翻译服务性质，语义不完整",
          "independent_ASR": {
            "status": "REVIEW_ONLY_RESCUE_BUDGET"
          },
          "suppression_proposed": null
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "76",
          "before": "它是B",
          "after": "它是B",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "Incomplete: '它是B' vs '它是免费' (it is free)",
          "context_evidence": "Context: '它是免费提供的接口'",
          "independent_ASR": {
            "status": "REVIEW_ONLY_RESCUE_BUDGET"
          },
          "suppression_proposed": null
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "76",
          "before": "它是B",
          "after": "它是B",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "Incomplete: '它是B' vs '它是免费' (it is free)",
          "context_evidence": "Context: '它是免费提供的接口'",
          "independent_ASR": {
            "status": "REVIEW_ONLY_RESCUE_BUDGET"
          },
          "suppression_proposed": null
        }
      ],
      "language": "zh",
      "packet_ids": [
        6
      ],
      "classification": "BEST_SUPPORTED_WITH_REVIEW",
      "independent_ASR": {
        "status": "REVIEW_ONLY_RESCUE_BUDGET"
      },
      "source_text": "它是B",
      "review_status": "BEST_SUPPORTED_WITH_REVIEW",
      "canonical_timing_evidence": "CUE_ANCHORED",
      "index": 76
    },
    {
      "start": 176.07999999999998,
      "end": 177.76,
      "text": "占官方提供的接口",
      "raw_asr_text": "占官方提供的接口",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 6,
      "words": [
        {
          "word": "占",
          "start": 176.07999999999998,
          "end": 176.64,
          "probability": 0.826171875
        },
        {
          "word": "官",
          "start": 176.64,
          "end": 176.82,
          "probability": 0.99853515625
        },
        {
          "word": "方",
          "start": 176.82,
          "end": 177.01999999999998,
          "probability": 0.99951171875
        },
        {
          "word": "提",
          "start": 177.01999999999998,
          "end": 177.16,
          "probability": 0.9990234375
        },
        {
          "word": "供",
          "start": 177.16,
          "end": 177.29999999999998,
          "probability": 1.0
        },
        {
          "word": "的",
          "start": 177.29999999999998,
          "end": 177.42,
          "probability": 1.0
        },
        {
          "word": "接",
          "start": 177.42,
          "end": 177.54,
          "probability": 0.7509765625
        },
        {
          "word": "口",
          "start": 177.54,
          "end": 177.76,
          "probability": 0.9990234375
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.12088049108331854,
        "compression_ratio": 1.105263157894737,
        "no_speech_prob": 1.2315977231114328e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "老朋友可能已经发现了1.0版本除了加新功能我还做了几个减法首先是移除了之前内置的index免费翻译它是B占官方提供的接口用的人一多就经常被挤爆不稳定翻译效果其实也不如其他主流大模型"
        }
      ],
      "id": 77,
      "review_annotation": "待確認",
      "row_id": "77",
      "bounded_row_id": 77,
      "parent_row_ids": [
        77
      ],
      "parent_row_id": 77,
      "packet_id": 6,
      "LID_evidence": [
        {
          "region_id": 34,
          "probabilities": {
            "zh": 0.9976210514704387
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        34
      ],
      "language_confidence": 0.9976210514704387,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 163.2,
      "packet_end": 185.70000000000002,
      "raw_text": "占官方提供的接口",
      "selected_text": "占官方提供的接口",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            77
          ],
          "packet_id": 6,
          "engine": "mlx-whisper"
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "77",
          "before": "占官方提供的接口",
          "after": "占官方提供的接口",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "占疑似调用或占用误识",
          "context_evidence": "上下文为接口使用，'调用'或'占用'更符合",
          "independent_ASR": {
            "status": "REVIEW_ONLY_RESCUE_BUDGET"
          },
          "suppression_proposed": null
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "77",
          "before": "占官方提供的接口",
          "after": "占官方提供的接口",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "Nonsense: '占' vs '官方' (official)",
          "context_evidence": "Context: '官方提供的接口'",
          "independent_ASR": {
            "status": "REVIEW_ONLY_RESCUE_BUDGET"
          },
          "suppression_proposed": null
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "77",
          "before": "占官方提供的接口",
          "after": "占官方提供的接口",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "Nonsense: '占' vs '官方' (official)",
          "context_evidence": "Context: '官方提供的接口'",
          "independent_ASR": {
            "status": "REVIEW_ONLY_RESCUE_BUDGET"
          },
          "suppression_proposed": null
        }
      ],
      "language": "zh",
      "packet_ids": [
        6
      ],
      "classification": "BEST_SUPPORTED_WITH_REVIEW",
      "independent_ASR": {
        "status": "REVIEW_ONLY_RESCUE_BUDGET"
      },
      "source_text": "占官方提供的接口",
      "review_status": "BEST_SUPPORTED_WITH_REVIEW",
      "canonical_timing_evidence": "CUE_ANCHORED",
      "index": 77
    },
    {
      "start": 177.76,
      "end": 180.06,
      "text": "用的人一多",
      "raw_asr_text": "用的人一多就经常被挤爆",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 6,
      "words": [],
      "asr_metrics": {
        "avg_logprob": -0.12088049108331854,
        "compression_ratio": 1.105263157894737,
        "no_speech_prob": 1.2315977231114328e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "老朋友可能已经发现了1.0版本除了加新功能我还做了几个减法首先是移除了之前内置的index免费翻译它是B占官方提供的接口用的人一多就经常被挤爆不稳定翻译效果其实也不如其他主流大模型"
        }
      ],
      "id": 78,
      "review_annotation": "待確認",
      "row_id": "78",
      "bounded_row_id": 78,
      "parent_row_ids": [
        78
      ],
      "parent_row_id": 78,
      "packet_id": 6,
      "LID_evidence": [
        {
          "region_id": 34,
          "probabilities": {
            "zh": 0.9976210514704387
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        34
      ],
      "language_confidence": 0.9976210514704387,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 163.2,
      "packet_end": 185.70000000000002,
      "raw_text": "用的人一多就经常被挤爆",
      "selected_text": "用的人一多",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            78
          ],
          "packet_id": 6,
          "engine": "mlx-whisper"
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "78",
          "before": "用的人一多就经常被挤爆",
          "after": "用的人一多",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "人一多疑似人多误识",
          "context_evidence": "口语中'人'与'一'连读可能导致误识",
          "independent_ASR": {
            "status": "REVIEW_ONLY_RESCUE_BUDGET"
          },
          "suppression_proposed": null
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "78",
          "before": "用的人一多",
          "after": "用的人一多",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "Nonsense: '用的人一多' vs '用的人一多' (when many use)",
          "context_evidence": "Context: '用的人一多'",
          "independent_ASR": {
            "status": "REVIEW_ONLY_RESCUE_BUDGET"
          },
          "suppression_proposed": null
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "78",
          "before": "用的人一多",
          "after": "用的人一多",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "Nonsense: '用的人一多' vs '用的人一多' (when many use)",
          "context_evidence": "Context: '用的人一多'",
          "independent_ASR": {
            "status": "REVIEW_ONLY_RESCUE_BUDGET"
          },
          "suppression_proposed": null
        }
      ],
      "language": "zh",
      "packet_ids": [
        6
      ],
      "classification": "BEST_SUPPORTED_WITH_REVIEW",
      "independent_ASR": {
        "status": "REVIEW_ONLY_RESCUE_BUDGET"
      },
      "raw_word_timestamps": [
        {
          "word": "用",
          "start": 177.76,
          "end": 178.33999999999997,
          "probability": 0.99951171875
        },
        {
          "word": "的人",
          "start": 178.33999999999997,
          "end": 178.48,
          "probability": 0.9970703125
        },
        {
          "word": "一",
          "start": 178.48,
          "end": 178.72,
          "probability": 0.99951171875
        },
        {
          "word": "多",
          "start": 178.72,
          "end": 178.89999999999998,
          "probability": 0.99951171875
        },
        {
          "word": "就",
          "start": 178.89999999999998,
          "end": 179.23999999999998,
          "probability": 0.99951171875
        },
        {
          "word": "经",
          "start": 179.23999999999998,
          "end": 179.39999999999998,
          "probability": 1.0
        },
        {
          "word": "常",
          "start": 179.39999999999998,
          "end": 179.57999999999998,
          "probability": 1.0
        },
        {
          "word": "被",
          "start": 179.57999999999998,
          "end": 179.73999999999998,
          "probability": 1.0
        },
        {
          "word": "挤",
          "start": 179.73999999999998,
          "end": 179.89999999999998,
          "probability": 0.955078125
        },
        {
          "word": "爆",
          "start": 179.89999999999998,
          "end": 180.06,
          "probability": 0.98046875
        }
      ],
      "word_timestamps": [],
      "source_text": "用的人一多",
      "review_status": "BEST_SUPPORTED_WITH_REVIEW",
      "canonical_timing_evidence": "CUE_ANCHORED",
      "index": 78
    },
    {
      "start": 180.06,
      "end": 180.98,
      "text": "不稳定",
      "raw_asr_text": "不稳定",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 6,
      "words": [
        {
          "word": "不",
          "start": 180.06,
          "end": 180.6,
          "probability": 0.9814453125
        },
        {
          "word": "稳",
          "start": 180.6,
          "end": 180.76,
          "probability": 1.0
        },
        {
          "word": "定",
          "start": 180.76,
          "end": 180.98,
          "probability": 1.0
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.12088049108331854,
        "compression_ratio": 1.105263157894737,
        "no_speech_prob": 1.2315977231114328e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "老朋友可能已经发现了1.0版本除了加新功能我还做了几个减法首先是移除了之前内置的index免费翻译它是B占官方提供的接口用的人一多就经常被挤爆不稳定翻译效果其实也不如其他主流大模型"
        }
      ],
      "id": 79,
      "review_annotation": "",
      "row_id": "79",
      "bounded_row_id": 79,
      "parent_row_ids": [
        79
      ],
      "parent_row_id": 79,
      "packet_id": 6,
      "LID_evidence": [
        {
          "region_id": 34,
          "probabilities": {
            "zh": 0.9976210514704387
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        34
      ],
      "language_confidence": 0.9976210514704387,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 163.2,
      "packet_end": 185.70000000000002,
      "raw_text": "不稳定",
      "selected_text": "不稳定",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            79
          ],
          "packet_id": 6,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        6
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 79
    },
    {
      "start": 181.7,
      "end": 184.6,
      "text": "翻译效果其实也不如其他主流大模型",
      "raw_asr_text": "翻译效果其实也不如其他主流大模型",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 6,
      "words": [
        {
          "word": "翻",
          "start": 181.7,
          "end": 182.06,
          "probability": 0.99951171875
        },
        {
          "word": "译",
          "start": 182.06,
          "end": 182.2,
          "probability": 1.0
        },
        {
          "word": "效",
          "start": 182.2,
          "end": 182.35999999999999,
          "probability": 0.99951171875
        },
        {
          "word": "果",
          "start": 182.35999999999999,
          "end": 182.6,
          "probability": 1.0
        },
        {
          "word": "其实",
          "start": 182.6,
          "end": 182.98,
          "probability": 0.99755859375
        },
        {
          "word": "也",
          "start": 182.98,
          "end": 183.2,
          "probability": 1.0
        },
        {
          "word": "不",
          "start": 183.2,
          "end": 183.33999999999997,
          "probability": 0.99951171875
        },
        {
          "word": "如",
          "start": 183.33999999999997,
          "end": 183.48,
          "probability": 0.99951171875
        },
        {
          "word": "其",
          "start": 183.48,
          "end": 183.64,
          "probability": 1.0
        },
        {
          "word": "他",
          "start": 183.64,
          "end": 183.76,
          "probability": 0.98681640625
        },
        {
          "word": "主",
          "start": 183.76,
          "end": 183.92,
          "probability": 1.0
        },
        {
          "word": "流",
          "start": 183.92,
          "end": 184.06,
          "probability": 0.9990234375
        },
        {
          "word": "大",
          "start": 184.06,
          "end": 184.2,
          "probability": 0.98095703125
        },
        {
          "word": "模",
          "start": 184.2,
          "end": 184.35999999999999,
          "probability": 0.88525390625
        },
        {
          "word": "型",
          "start": 184.35999999999999,
          "end": 184.6,
          "probability": 0.99951171875
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.12088049108331854,
        "compression_ratio": 1.105263157894737,
        "no_speech_prob": 1.2315977231114328e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "老朋友可能已经发现了1.0版本除了加新功能我还做了几个减法首先是移除了之前内置的index免费翻译它是B占官方提供的接口用的人一多就经常被挤爆不稳定翻译效果其实也不如其他主流大模型"
        }
      ],
      "id": 80,
      "review_annotation": "",
      "row_id": "80",
      "bounded_row_id": 80,
      "parent_row_ids": [
        80
      ],
      "parent_row_id": 80,
      "packet_id": 6,
      "LID_evidence": [
        {
          "region_id": 35,
          "probabilities": {
            "zh": 0.9976518154144287
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        35
      ],
      "language_confidence": 0.9976518154144287,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 163.2,
      "packet_end": 185.70000000000002,
      "raw_text": "翻译效果其实也不如其他主流大模型",
      "selected_text": "翻译效果其实也不如其他主流大模型",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            80
          ],
          "packet_id": 6,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        6
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 80
    },
    {
      "start": 185.5,
      "end": 187.07999999999998,
      "text": "我后来想明白了一件事",
      "raw_asr_text": "我后来想明白了一件事",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 7,
      "words": [
        {
          "word": "我",
          "start": 185.44,
          "end": 185.76,
          "probability": 0.978515625
        },
        {
          "word": "后",
          "start": 185.76,
          "end": 185.89999999999998,
          "probability": 0.998046875
        },
        {
          "word": "来",
          "start": 185.89999999999998,
          "end": 186.06,
          "probability": 0.99951171875
        },
        {
          "word": "想",
          "start": 186.06,
          "end": 186.2,
          "probability": 0.98974609375
        },
        {
          "word": "明白",
          "start": 186.2,
          "end": 186.44,
          "probability": 0.9814453125
        },
        {
          "word": "了",
          "start": 186.44,
          "end": 186.6,
          "probability": 1.0
        },
        {
          "word": "一",
          "start": 186.6,
          "end": 186.67999999999998,
          "probability": 0.99951171875
        },
        {
          "word": "件",
          "start": 186.67999999999998,
          "end": 186.82,
          "probability": 0.9970703125
        },
        {
          "word": "事",
          "start": 186.82,
          "end": 187.07999999999998,
          "probability": 0.99951171875
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.12329312100436518,
        "compression_ratio": 1.3413793103448275,
        "no_speech_prob": 2.021101688787219e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "我后来想明白了一件事其他提供免费翻译的软件它们本身是做付费服务的商业软件对免费功能的可用性也是有保障义务的当免费翻译用不了的时候用户去反馈指责甚至骂两句那是商业软件应该受着的但同传字幕是个开源项目Index是逼不赚提供的服务它不稳定的问题本来不应该反馈到我这里可只要"
        }
      ],
      "id": 81,
      "review_annotation": "",
      "row_id": "81",
      "bounded_row_id": 81,
      "parent_row_ids": [
        81
      ],
      "parent_row_id": 81,
      "packet_id": 7,
      "LID_evidence": [
        {
          "region_id": 36,
          "probabilities": {
            "zh": 0.9977045164388769
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        36
      ],
      "language_confidence": 0.9977045164388769,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 184.7,
      "packet_end": 212.70000000000002,
      "raw_text": "我后来想明白了一件事",
      "selected_text": "我后来想明白了一件事",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            81
          ],
          "packet_id": 7,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        7
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 81
    },
    {
      "start": 187.07999999999998,
      "end": 189.1,
      "text": "其他提供免费翻译的软件",
      "raw_asr_text": "其他提供免费翻译的软件",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 7,
      "words": [
        {
          "word": "其",
          "start": 187.07999999999998,
          "end": 187.64,
          "probability": 0.88818359375
        },
        {
          "word": "他",
          "start": 187.64,
          "end": 187.76,
          "probability": 0.97607421875
        },
        {
          "word": "提",
          "start": 187.76,
          "end": 187.92,
          "probability": 1.0
        },
        {
          "word": "供",
          "start": 187.92,
          "end": 188.06,
          "probability": 1.0
        },
        {
          "word": "免",
          "start": 188.06,
          "end": 188.23999999999998,
          "probability": 1.0
        },
        {
          "word": "费",
          "start": 188.23999999999998,
          "end": 188.35999999999999,
          "probability": 0.9990234375
        },
        {
          "word": "翻",
          "start": 188.35999999999999,
          "end": 188.5,
          "probability": 0.9990234375
        },
        {
          "word": "译",
          "start": 188.5,
          "end": 188.61999999999998,
          "probability": 0.99853515625
        },
        {
          "word": "的",
          "start": 188.61999999999998,
          "end": 188.72,
          "probability": 1.0
        },
        {
          "word": "软",
          "start": 188.72,
          "end": 188.89999999999998,
          "probability": 0.9990234375
        },
        {
          "word": "件",
          "start": 188.89999999999998,
          "end": 189.1,
          "probability": 0.99951171875
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.12329312100436518,
        "compression_ratio": 1.3413793103448275,
        "no_speech_prob": 2.021101688787219e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "我后来想明白了一件事其他提供免费翻译的软件它们本身是做付费服务的商业软件对免费功能的可用性也是有保障义务的当免费翻译用不了的时候用户去反馈指责甚至骂两句那是商业软件应该受着的但同传字幕是个开源项目Index是逼不赚提供的服务它不稳定的问题本来不应该反馈到我这里可只要"
        }
      ],
      "id": 82,
      "review_annotation": "",
      "row_id": "82",
      "bounded_row_id": 82,
      "parent_row_ids": [
        82
      ],
      "parent_row_id": 82,
      "packet_id": 7,
      "LID_evidence": [
        {
          "region_id": 36,
          "probabilities": {
            "zh": 0.9977045164388769
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        36
      ],
      "language_confidence": 0.9977045164388769,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 184.7,
      "packet_end": 212.70000000000002,
      "raw_text": "其他提供免费翻译的软件",
      "selected_text": "其他提供免费翻译的软件",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            82
          ],
          "packet_id": 7,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        7
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 82
    },
    {
      "start": 189.1,
      "end": 191.76,
      "text": "它们本身是做付费服务的商业软件",
      "raw_asr_text": "它们本身是做付费服务的商业软件",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 7,
      "words": [
        {
          "word": "它",
          "start": 189.1,
          "end": 189.56,
          "probability": 0.544921875
        },
        {
          "word": "们",
          "start": 189.56,
          "end": 189.67999999999998,
          "probability": 0.99951171875
        },
        {
          "word": "本",
          "start": 189.67999999999998,
          "end": 189.83999999999997,
          "probability": 1.0
        },
        {
          "word": "身",
          "start": 189.83999999999997,
          "end": 190.01999999999998,
          "probability": 1.0
        },
        {
          "word": "是",
          "start": 190.01999999999998,
          "end": 190.17999999999998,
          "probability": 1.0
        },
        {
          "word": "做",
          "start": 190.17999999999998,
          "end": 190.32,
          "probability": 0.94775390625
        },
        {
          "word": "付",
          "start": 190.32,
          "end": 190.48,
          "probability": 0.88720703125
        },
        {
          "word": "费",
          "start": 190.48,
          "end": 190.64,
          "probability": 0.9541015625
        },
        {
          "word": "服",
          "start": 190.64,
          "end": 190.79999999999998,
          "probability": 1.0
        },
        {
          "word": "务",
          "start": 190.79999999999998,
          "end": 190.92,
          "probability": 1.0
        },
        {
          "word": "的",
          "start": 190.92,
          "end": 191.04,
          "probability": 1.0
        },
        {
          "word": "商",
          "start": 191.04,
          "end": 191.2,
          "probability": 0.99853515625
        },
        {
          "word": "业",
          "start": 191.2,
          "end": 191.35999999999999,
          "probability": 0.99951171875
        },
        {
          "word": "软",
          "start": 191.35999999999999,
          "end": 191.54,
          "probability": 1.0
        },
        {
          "word": "件",
          "start": 191.54,
          "end": 191.76,
          "probability": 1.0
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.12329312100436518,
        "compression_ratio": 1.3413793103448275,
        "no_speech_prob": 2.021101688787219e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "我后来想明白了一件事其他提供免费翻译的软件它们本身是做付费服务的商业软件对免费功能的可用性也是有保障义务的当免费翻译用不了的时候用户去反馈指责甚至骂两句那是商业软件应该受着的但同传字幕是个开源项目Index是逼不赚提供的服务它不稳定的问题本来不应该反馈到我这里可只要"
        }
      ],
      "id": 83,
      "review_annotation": "",
      "row_id": "83",
      "bounded_row_id": 83,
      "parent_row_ids": [
        83
      ],
      "parent_row_id": 83,
      "packet_id": 7,
      "LID_evidence": [
        {
          "region_id": 36,
          "probabilities": {
            "zh": 0.9977045164388769
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        36
      ],
      "language_confidence": 0.9977045164388769,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 184.7,
      "packet_end": 212.70000000000002,
      "raw_text": "它们本身是做付费服务的商业软件",
      "selected_text": "它们本身是做付费服务的商业软件",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            83
          ],
          "packet_id": 7,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        7
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 83
    },
    {
      "start": 192.56,
      "end": 194.22,
      "text": "对免费功能的可用性",
      "raw_asr_text": "对免费功能的可用性",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 7,
      "words": [
        {
          "word": "对",
          "start": 192.56,
          "end": 192.88,
          "probability": 0.99560546875
        },
        {
          "word": "免",
          "start": 192.88,
          "end": 193.07999999999998,
          "probability": 1.0
        },
        {
          "word": "费",
          "start": 193.07999999999998,
          "end": 193.2,
          "probability": 1.0
        },
        {
          "word": "功",
          "start": 193.2,
          "end": 193.38,
          "probability": 0.99853515625
        },
        {
          "word": "能",
          "start": 193.38,
          "end": 193.51999999999998,
          "probability": 1.0
        },
        {
          "word": "的",
          "start": 193.51999999999998,
          "end": 193.64,
          "probability": 1.0
        },
        {
          "word": "可",
          "start": 193.64,
          "end": 193.82,
          "probability": 0.99951171875
        },
        {
          "word": "用",
          "start": 193.82,
          "end": 194.01999999999998,
          "probability": 1.0
        },
        {
          "word": "性",
          "start": 194.01999999999998,
          "end": 194.22,
          "probability": 1.0
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.12329312100436518,
        "compression_ratio": 1.3413793103448275,
        "no_speech_prob": 2.021101688787219e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "我后来想明白了一件事其他提供免费翻译的软件它们本身是做付费服务的商业软件对免费功能的可用性也是有保障义务的当免费翻译用不了的时候用户去反馈指责甚至骂两句那是商业软件应该受着的但同传字幕是个开源项目Index是逼不赚提供的服务它不稳定的问题本来不应该反馈到我这里可只要"
        }
      ],
      "id": 84,
      "review_annotation": "",
      "row_id": "84",
      "bounded_row_id": 84,
      "parent_row_ids": [
        84
      ],
      "parent_row_id": 84,
      "packet_id": 7,
      "LID_evidence": [
        {
          "region_id": 37,
          "probabilities": {
            "zh": 0.9980977177619934
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        37
      ],
      "language_confidence": 0.9980977177619934,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 184.7,
      "packet_end": 212.70000000000002,
      "raw_text": "对免费功能的可用性",
      "selected_text": "对免费功能的可用性",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            84
          ],
          "packet_id": 7,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        7
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 84
    },
    {
      "start": 194.22,
      "end": 195.61999999999998,
      "text": "也是有保障义务的",
      "raw_asr_text": "也是有保障义务的",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 7,
      "words": [
        {
          "word": "也是",
          "start": 194.22,
          "end": 194.67999999999998,
          "probability": 0.99755859375
        },
        {
          "word": "有",
          "start": 194.67999999999998,
          "end": 194.82,
          "probability": 1.0
        },
        {
          "word": "保",
          "start": 194.82,
          "end": 194.95999999999998,
          "probability": 0.99951171875
        },
        {
          "word": "障",
          "start": 194.95999999999998,
          "end": 195.16,
          "probability": 0.998046875
        },
        {
          "word": "义",
          "start": 195.16,
          "end": 195.32,
          "probability": 0.9951171875
        },
        {
          "word": "务",
          "start": 195.32,
          "end": 195.45999999999998,
          "probability": 0.998046875
        },
        {
          "word": "的",
          "start": 195.45999999999998,
          "end": 195.61999999999998,
          "probability": 1.0
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.12329312100436518,
        "compression_ratio": 1.3413793103448275,
        "no_speech_prob": 2.021101688787219e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "我后来想明白了一件事其他提供免费翻译的软件它们本身是做付费服务的商业软件对免费功能的可用性也是有保障义务的当免费翻译用不了的时候用户去反馈指责甚至骂两句那是商业软件应该受着的但同传字幕是个开源项目Index是逼不赚提供的服务它不稳定的问题本来不应该反馈到我这里可只要"
        }
      ],
      "id": 85,
      "review_annotation": "",
      "row_id": "85",
      "bounded_row_id": 85,
      "parent_row_ids": [
        85
      ],
      "parent_row_id": 85,
      "packet_id": 7,
      "LID_evidence": [
        {
          "region_id": 37,
          "probabilities": {
            "zh": 0.9980977177619934
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        37
      ],
      "language_confidence": 0.9980977177619934,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 184.7,
      "packet_end": 212.70000000000002,
      "raw_text": "也是有保障义务的",
      "selected_text": "也是有保障义务的",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            85
          ],
          "packet_id": 7,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        7
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 85
    },
    {
      "start": 196.16,
      "end": 197.83999999999997,
      "text": "当免费翻译用不了的时候",
      "raw_asr_text": "当免费翻译用不了的时候",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 7,
      "words": [
        {
          "word": "当",
          "start": 196.16,
          "end": 196.48,
          "probability": 0.99609375
        },
        {
          "word": "免",
          "start": 196.48,
          "end": 196.66,
          "probability": 1.0
        },
        {
          "word": "费",
          "start": 196.66,
          "end": 196.79999999999998,
          "probability": 1.0
        },
        {
          "word": "翻",
          "start": 196.79999999999998,
          "end": 196.95999999999998,
          "probability": 1.0
        },
        {
          "word": "译",
          "start": 196.95999999999998,
          "end": 197.1,
          "probability": 1.0
        },
        {
          "word": "用",
          "start": 197.1,
          "end": 197.23999999999998,
          "probability": 0.9990234375
        },
        {
          "word": "不了",
          "start": 197.23999999999998,
          "end": 197.45999999999998,
          "probability": 1.0
        },
        {
          "word": "的时候",
          "start": 197.45999999999998,
          "end": 197.83999999999997,
          "probability": 0.99853515625
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.12329312100436518,
        "compression_ratio": 1.3413793103448275,
        "no_speech_prob": 2.021101688787219e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "我后来想明白了一件事其他提供免费翻译的软件它们本身是做付费服务的商业软件对免费功能的可用性也是有保障义务的当免费翻译用不了的时候用户去反馈指责甚至骂两句那是商业软件应该受着的但同传字幕是个开源项目Index是逼不赚提供的服务它不稳定的问题本来不应该反馈到我这里可只要"
        }
      ],
      "id": 86,
      "review_annotation": "",
      "row_id": "86",
      "bounded_row_id": 86,
      "parent_row_ids": [
        86
      ],
      "parent_row_id": 86,
      "packet_id": 7,
      "LID_evidence": [
        {
          "region_id": 38,
          "probabilities": {
            "zh": 0.9979860153463151
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        38
      ],
      "language_confidence": 0.9979860153463151,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 184.7,
      "packet_end": 212.70000000000002,
      "raw_text": "当免费翻译用不了的时候",
      "selected_text": "当免费翻译用不了的时候",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            86
          ],
          "packet_id": 7,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        7
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 86
    },
    {
      "start": 198.2,
      "end": 199.22,
      "text": "用户去反馈",
      "raw_asr_text": "用户去反馈",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 7,
      "words": [
        {
          "word": "用",
          "start": 198.2,
          "end": 198.51999999999998,
          "probability": 0.99853515625
        },
        {
          "word": "户",
          "start": 198.51999999999998,
          "end": 198.67999999999998,
          "probability": 0.998046875
        },
        {
          "word": "去",
          "start": 198.67999999999998,
          "end": 198.82,
          "probability": 0.9970703125
        },
        {
          "word": "反",
          "start": 198.82,
          "end": 198.98,
          "probability": 1.0
        },
        {
          "word": "馈",
          "start": 198.98,
          "end": 199.22,
          "probability": 0.998046875
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.12329312100436518,
        "compression_ratio": 1.3413793103448275,
        "no_speech_prob": 2.021101688787219e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "我后来想明白了一件事其他提供免费翻译的软件它们本身是做付费服务的商业软件对免费功能的可用性也是有保障义务的当免费翻译用不了的时候用户去反馈指责甚至骂两句那是商业软件应该受着的但同传字幕是个开源项目Index是逼不赚提供的服务它不稳定的问题本来不应该反馈到我这里可只要"
        }
      ],
      "id": 87,
      "review_annotation": "",
      "row_id": "87",
      "bounded_row_id": 87,
      "parent_row_ids": [
        87
      ],
      "parent_row_id": 87,
      "packet_id": 7,
      "LID_evidence": [
        {
          "region_id": 38,
          "probabilities": {
            "zh": 0.9979860153463151
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        38
      ],
      "language_confidence": 0.9979860153463151,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 184.7,
      "packet_end": 212.70000000000002,
      "raw_text": "用户去反馈",
      "selected_text": "用户去反馈",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            87
          ],
          "packet_id": 7,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        7
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 87
    },
    {
      "start": 199.22,
      "end": 200.67999999999998,
      "text": "指责甚至骂两句",
      "raw_asr_text": "指责甚至骂两句",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 7,
      "words": [
        {
          "word": "指",
          "start": 199.22,
          "end": 199.72,
          "probability": 0.837890625
        },
        {
          "word": "责",
          "start": 199.72,
          "end": 199.92,
          "probability": 0.99951171875
        },
        {
          "word": "甚",
          "start": 199.92,
          "end": 200.06,
          "probability": 0.9990234375
        },
        {
          "word": "至",
          "start": 200.06,
          "end": 200.17999999999998,
          "probability": 1.0
        },
        {
          "word": "骂",
          "start": 200.17999999999998,
          "end": 200.35999999999999,
          "probability": 0.9990234375
        },
        {
          "word": "两",
          "start": 200.35999999999999,
          "end": 200.48,
          "probability": 0.99853515625
        },
        {
          "word": "句",
          "start": 200.48,
          "end": 200.67999999999998,
          "probability": 1.0
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.12329312100436518,
        "compression_ratio": 1.3413793103448275,
        "no_speech_prob": 2.021101688787219e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "我后来想明白了一件事其他提供免费翻译的软件它们本身是做付费服务的商业软件对免费功能的可用性也是有保障义务的当免费翻译用不了的时候用户去反馈指责甚至骂两句那是商业软件应该受着的但同传字幕是个开源项目Index是逼不赚提供的服务它不稳定的问题本来不应该反馈到我这里可只要"
        }
      ],
      "id": 88,
      "review_annotation": "",
      "row_id": "88",
      "bounded_row_id": 88,
      "parent_row_ids": [
        88
      ],
      "parent_row_id": 88,
      "packet_id": 7,
      "LID_evidence": [
        {
          "region_id": 38,
          "probabilities": {
            "zh": 0.9979860153463151
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        38
      ],
      "language_confidence": 0.9979860153463151,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 184.7,
      "packet_end": 212.70000000000002,
      "raw_text": "指责甚至骂两句",
      "selected_text": "指责甚至骂两句",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            88
          ],
          "packet_id": 7,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        7
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 88
    },
    {
      "start": 200.67999999999998,
      "end": 203.2,
      "text": "那是商业软件应该受着的",
      "raw_asr_text": "那是商业软件应该受着的",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 7,
      "words": [
        {
          "word": "那",
          "start": 200.67999999999998,
          "end": 201.27999999999997,
          "probability": 0.9990234375
        },
        {
          "word": "是",
          "start": 201.27999999999997,
          "end": 201.39999999999998,
          "probability": 0.99853515625
        },
        {
          "word": "商",
          "start": 201.39999999999998,
          "end": 201.56,
          "probability": 1.0
        },
        {
          "word": "业",
          "start": 201.56,
          "end": 201.73999999999998,
          "probability": 1.0
        },
        {
          "word": "软",
          "start": 201.73999999999998,
          "end": 201.92,
          "probability": 1.0
        },
        {
          "word": "件",
          "start": 201.92,
          "end": 202.1,
          "probability": 1.0
        },
        {
          "word": "应",
          "start": 202.1,
          "end": 202.5,
          "probability": 0.998046875
        },
        {
          "word": "该",
          "start": 202.5,
          "end": 202.6,
          "probability": 1.0
        },
        {
          "word": "受",
          "start": 202.6,
          "end": 202.73999999999998,
          "probability": 0.93359375
        },
        {
          "word": "着",
          "start": 202.73999999999998,
          "end": 202.89999999999998,
          "probability": 0.998046875
        },
        {
          "word": "的",
          "start": 202.89999999999998,
          "end": 203.2,
          "probability": 1.0
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.12329312100436518,
        "compression_ratio": 1.3413793103448275,
        "no_speech_prob": 2.021101688787219e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "我后来想明白了一件事其他提供免费翻译的软件它们本身是做付费服务的商业软件对免费功能的可用性也是有保障义务的当免费翻译用不了的时候用户去反馈指责甚至骂两句那是商业软件应该受着的但同传字幕是个开源项目Index是逼不赚提供的服务它不稳定的问题本来不应该反馈到我这里可只要"
        }
      ],
      "id": 89,
      "review_annotation": "",
      "row_id": "89",
      "bounded_row_id": 89,
      "parent_row_ids": [
        89
      ],
      "parent_row_id": 89,
      "packet_id": 7,
      "LID_evidence": [
        {
          "region_id": 38,
          "probabilities": {
            "zh": 0.9979860153463151
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        38
      ],
      "language_confidence": 0.9979860153463151,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 184.7,
      "packet_end": 212.70000000000002,
      "raw_text": "那是商业软件应该受着的",
      "selected_text": "那是商业软件应该受着的",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            89
          ],
          "packet_id": 7,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        7
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 89
    },
    {
      "start": 204.0,
      "end": 205.98,
      "text": "但同传字幕是个开源项目",
      "raw_asr_text": "但同传字幕是个开源项目",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 7,
      "words": [
        {
          "word": "但",
          "start": 204.0,
          "end": 204.32,
          "probability": 0.99560546875
        },
        {
          "word": "同",
          "start": 204.32,
          "end": 204.48,
          "probability": 0.9736328125
        },
        {
          "word": "传",
          "start": 204.48,
          "end": 204.67999999999998,
          "probability": 0.994140625
        },
        {
          "word": "字",
          "start": 204.67999999999998,
          "end": 204.85999999999999,
          "probability": 0.9990234375
        },
        {
          "word": "幕",
          "start": 204.85999999999999,
          "end": 204.98,
          "probability": 0.7421875
        },
        {
          "word": "是",
          "start": 204.98,
          "end": 205.12,
          "probability": 0.99951171875
        },
        {
          "word": "个",
          "start": 205.12,
          "end": 205.22,
          "probability": 0.99560546875
        },
        {
          "word": "开",
          "start": 205.22,
          "end": 205.35999999999999,
          "probability": 0.99951171875
        },
        {
          "word": "源",
          "start": 205.35999999999999,
          "end": 205.51999999999998,
          "probability": 0.99169921875
        },
        {
          "word": "项",
          "start": 205.51999999999998,
          "end": 205.73999999999998,
          "probability": 1.0
        },
        {
          "word": "目",
          "start": 205.73999999999998,
          "end": 205.98,
          "probability": 0.99951171875
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.12329312100436518,
        "compression_ratio": 1.3413793103448275,
        "no_speech_prob": 2.021101688787219e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "我后来想明白了一件事其他提供免费翻译的软件它们本身是做付费服务的商业软件对免费功能的可用性也是有保障义务的当免费翻译用不了的时候用户去反馈指责甚至骂两句那是商业软件应该受着的但同传字幕是个开源项目Index是逼不赚提供的服务它不稳定的问题本来不应该反馈到我这里可只要"
        }
      ],
      "id": 90,
      "review_annotation": "",
      "row_id": "90",
      "bounded_row_id": 90,
      "parent_row_ids": [
        90
      ],
      "parent_row_id": 90,
      "packet_id": 7,
      "LID_evidence": [
        {
          "region_id": 39,
          "probabilities": {
            "zh": 0.9978515356779099
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        39
      ],
      "language_confidence": 0.9978515356779099,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 184.7,
      "packet_end": 212.70000000000002,
      "raw_text": "但同传字幕是个开源项目",
      "selected_text": "但同传字幕是个开源项目",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            90
          ],
          "packet_id": 7,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        7
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 90
    },
    {
      "start": 205.98,
      "end": 208.76,
      "text": "Index是逼不赚提供的服务",
      "raw_asr_text": "Index是逼不赚提供的服务",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 7,
      "words": [
        {
          "word": "Ind",
          "start": 205.98,
          "end": 206.57999999999998,
          "probability": 0.462890625
        },
        {
          "word": "ex",
          "start": 206.57999999999998,
          "end": 206.88,
          "probability": 0.998046875
        },
        {
          "word": "是",
          "start": 206.88,
          "end": 207.14,
          "probability": 0.9970703125
        },
        {
          "word": "逼",
          "start": 207.14,
          "end": 207.32,
          "probability": 0.7373046875
        },
        {
          "word": "不",
          "start": 207.32,
          "end": 207.44,
          "probability": 0.93798828125
        },
        {
          "word": "赚",
          "start": 207.44,
          "end": 207.57999999999998,
          "probability": 0.828125
        },
        {
          "word": "提",
          "start": 207.57999999999998,
          "end": 207.73999999999998,
          "probability": 0.998046875
        },
        {
          "word": "供",
          "start": 207.73999999999998,
          "end": 207.89999999999998,
          "probability": 1.0
        },
        {
          "word": "的",
          "start": 207.89999999999998,
          "end": 208.0,
          "probability": 1.0
        },
        {
          "word": "服",
          "start": 208.0,
          "end": 208.14,
          "probability": 0.99951171875
        },
        {
          "word": "务",
          "start": 208.14,
          "end": 208.76,
          "probability": 1.0
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.12329312100436518,
        "compression_ratio": 1.3413793103448275,
        "no_speech_prob": 2.021101688787219e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "我后来想明白了一件事其他提供免费翻译的软件它们本身是做付费服务的商业软件对免费功能的可用性也是有保障义务的当免费翻译用不了的时候用户去反馈指责甚至骂两句那是商业软件应该受着的但同传字幕是个开源项目Index是逼不赚提供的服务它不稳定的问题本来不应该反馈到我这里可只要"
        }
      ],
      "id": 91,
      "review_annotation": "待確認",
      "row_id": "91",
      "bounded_row_id": 91,
      "parent_row_ids": [
        91
      ],
      "parent_row_id": 91,
      "packet_id": 7,
      "LID_evidence": [
        {
          "region_id": 39,
          "probabilities": {
            "zh": 0.9978515356779099
          },
          "status": "STABLE"
        },
        {
          "region_id": 40,
          "probabilities": {
            "zh": 0.9977081338564555
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        39,
        40
      ],
      "language_confidence": 0.9978515356779099,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 184.7,
      "packet_end": 212.70000000000002,
      "raw_text": "Index是逼不赚提供的服务",
      "selected_text": "Index是逼不赚提供的服务",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            91
          ],
          "packet_id": 7,
          "engine": "mlx-whisper"
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "91",
          "before": "Index是逼不赚提供的服务",
          "after": "Index是逼不赚提供的服务",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "逼不赚疑似B站或Bilibili误识",
          "context_evidence": "上下文为B站生态，'B站'更符合逻辑",
          "independent_ASR": {
            "status": "REVIEW_ONLY_RESCUE_BUDGET"
          },
          "suppression_proposed": null
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "91",
          "before": "Index是逼不赚提供的服务",
          "after": "Index是逼不赚提供的服务",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "Hallucination: '逼不赚' vs '免费' (free)",
          "context_evidence": "Context: 'Index是免费提供的服务'",
          "independent_ASR": {
            "status": "REVIEW_ONLY_RESCUE_BUDGET"
          },
          "suppression_proposed": null
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "91",
          "before": "Index是逼不赚提供的服务",
          "after": "Index是逼不赚提供的服务",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "Hallucination: '逼不赚' vs '免费' (free)",
          "context_evidence": "Context: 'Index是免费提供的服务'",
          "independent_ASR": {
            "status": "REVIEW_ONLY_RESCUE_BUDGET"
          },
          "suppression_proposed": null
        }
      ],
      "language": "zh",
      "packet_ids": [
        7
      ],
      "classification": "BEST_SUPPORTED_WITH_REVIEW",
      "independent_ASR": {
        "status": "REVIEW_ONLY_RESCUE_BUDGET"
      },
      "source_text": "Index是逼不赚提供的服务",
      "review_status": "BEST_SUPPORTED_WITH_REVIEW",
      "canonical_timing_evidence": "CUE_ANCHORED",
      "index": 91
    },
    {
      "start": 208.76,
      "end": 209.79999999999998,
      "text": "它不稳定的问题",
      "raw_asr_text": "它不稳定的问题",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 7,
      "words": [
        {
          "word": "它",
          "start": 208.76,
          "end": 209.0,
          "probability": 0.94482421875
        },
        {
          "word": "不",
          "start": 209.0,
          "end": 209.16,
          "probability": 1.0
        },
        {
          "word": "稳",
          "start": 209.16,
          "end": 209.27999999999997,
          "probability": 0.99951171875
        },
        {
          "word": "定",
          "start": 209.27999999999997,
          "end": 209.42,
          "probability": 1.0
        },
        {
          "word": "的",
          "start": 209.42,
          "end": 209.51999999999998,
          "probability": 1.0
        },
        {
          "word": "问题",
          "start": 209.51999999999998,
          "end": 209.79999999999998,
          "probability": 1.0
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.12329312100436518,
        "compression_ratio": 1.3413793103448275,
        "no_speech_prob": 2.021101688787219e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "我后来想明白了一件事其他提供免费翻译的软件它们本身是做付费服务的商业软件对免费功能的可用性也是有保障义务的当免费翻译用不了的时候用户去反馈指责甚至骂两句那是商业软件应该受着的但同传字幕是个开源项目Index是逼不赚提供的服务它不稳定的问题本来不应该反馈到我这里可只要"
        }
      ],
      "id": 92,
      "review_annotation": "",
      "row_id": "92",
      "bounded_row_id": 92,
      "parent_row_ids": [
        92
      ],
      "parent_row_id": 92,
      "packet_id": 7,
      "LID_evidence": [
        {
          "region_id": 40,
          "probabilities": {
            "zh": 0.9977081338564555
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        40
      ],
      "language_confidence": 0.9977081338564555,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 184.7,
      "packet_end": 212.70000000000002,
      "raw_text": "它不稳定的问题",
      "selected_text": "它不稳定的问题",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            92
          ],
          "packet_id": 7,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        7
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 92
    },
    {
      "start": 209.79999999999998,
      "end": 211.6,
      "text": "本来不应该反馈到我这里",
      "raw_asr_text": "本来不应该反馈到我这里",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 7,
      "words": [
        {
          "word": "本",
          "start": 209.79999999999998,
          "end": 210.26,
          "probability": 0.9990234375
        },
        {
          "word": "来",
          "start": 210.26,
          "end": 210.39999999999998,
          "probability": 1.0
        },
        {
          "word": "不",
          "start": 210.39999999999998,
          "end": 210.54,
          "probability": 0.99951171875
        },
        {
          "word": "应",
          "start": 210.54,
          "end": 210.7,
          "probability": 1.0
        },
        {
          "word": "该",
          "start": 210.7,
          "end": 210.79999999999998,
          "probability": 0.99951171875
        },
        {
          "word": "反",
          "start": 210.79999999999998,
          "end": 210.92,
          "probability": 0.99951171875
        },
        {
          "word": "馈",
          "start": 210.92,
          "end": 211.06,
          "probability": 0.9990234375
        },
        {
          "word": "到",
          "start": 211.06,
          "end": 211.17999999999998,
          "probability": 0.9990234375
        },
        {
          "word": "我",
          "start": 211.17999999999998,
          "end": 211.32,
          "probability": 1.0
        },
        {
          "word": "这里",
          "start": 211.32,
          "end": 211.6,
          "probability": 0.99951171875
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.12329312100436518,
        "compression_ratio": 1.3413793103448275,
        "no_speech_prob": 2.021101688787219e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "我后来想明白了一件事其他提供免费翻译的软件它们本身是做付费服务的商业软件对免费功能的可用性也是有保障义务的当免费翻译用不了的时候用户去反馈指责甚至骂两句那是商业软件应该受着的但同传字幕是个开源项目Index是逼不赚提供的服务它不稳定的问题本来不应该反馈到我这里可只要"
        }
      ],
      "id": 93,
      "review_annotation": "",
      "row_id": "93",
      "bounded_row_id": 93,
      "parent_row_ids": [
        93
      ],
      "parent_row_id": 93,
      "packet_id": 7,
      "LID_evidence": [
        {
          "region_id": 40,
          "probabilities": {
            "zh": 0.9977081338564555
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        40
      ],
      "language_confidence": 0.9977081338564555,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 184.7,
      "packet_end": 212.70000000000002,
      "raw_text": "本来不应该反馈到我这里",
      "selected_text": "本来不应该反馈到我这里",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            93
          ],
          "packet_id": 7,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        7
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 93
    },
    {
      "start": 212.1,
      "end": 214.1,
      "text": "可只要我把它内置在软件里",
      "raw_asr_text": "可只要我把它内置在软件里",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 8,
      "words": [
        {
          "word": "可",
          "start": 212.01999999999998,
          "end": 212.38,
          "probability": 0.09185791015625
        },
        {
          "word": "只",
          "start": 212.38,
          "end": 212.57999999999998,
          "probability": 0.99609375
        },
        {
          "word": "要",
          "start": 212.57999999999998,
          "end": 212.7,
          "probability": 0.99951171875
        },
        {
          "word": "我",
          "start": 212.7,
          "end": 212.82,
          "probability": 0.9990234375
        },
        {
          "word": "把它",
          "start": 212.82,
          "end": 213.04,
          "probability": 0.96533203125
        },
        {
          "word": "内",
          "start": 213.04,
          "end": 213.26,
          "probability": 0.97412109375
        },
        {
          "word": "置",
          "start": 213.26,
          "end": 213.38,
          "probability": 0.99609375
        },
        {
          "word": "在",
          "start": 213.38,
          "end": 213.49999999999997,
          "probability": 0.99755859375
        },
        {
          "word": "软",
          "start": 213.49999999999997,
          "end": 213.7,
          "probability": 0.9990234375
        },
        {
          "word": "件",
          "start": 213.7,
          "end": 213.92,
          "probability": 1.0
        },
        {
          "word": "里",
          "start": 213.92,
          "end": 214.1,
          "probability": 0.99951171875
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.14027634287268165,
        "compression_ratio": 1.2254098360655739,
        "no_speech_prob": 4.3821491574336235e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "可只要我把它内置在软件里大家自然会觉得这是我提供的服务我就对它有义务所以我决定把它移除掉当然你依然可以通过现有的自定义配置自主合规地建入从前的index翻译接口但那是你亲手填上去用的你会清楚这是你自己的选择"
        }
      ],
      "id": 94,
      "review_annotation": "",
      "row_id": "94",
      "bounded_row_id": 94,
      "parent_row_ids": [
        94
      ],
      "parent_row_id": 94,
      "packet_id": 8,
      "LID_evidence": [
        {
          "region_id": 41,
          "probabilities": {
            "zh": 0.9978593746821086
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        41
      ],
      "language_confidence": 0.9978593746821086,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 211.29999999999998,
      "packet_end": 235.5,
      "raw_text": "可只要我把它内置在软件里",
      "selected_text": "可只要我把它内置在软件里",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            94
          ],
          "packet_id": 8,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        8
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 94
    },
    {
      "start": 214.1,
      "end": 216.43999999999997,
      "text": "大家自然会觉得这是我提供的服务",
      "raw_asr_text": "大家自然会觉得这是我提供的服务",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 8,
      "words": [
        {
          "word": "大家",
          "start": 214.1,
          "end": 214.57999999999998,
          "probability": 0.82666015625
        },
        {
          "word": "自",
          "start": 214.57999999999998,
          "end": 214.83999999999997,
          "probability": 1.0
        },
        {
          "word": "然",
          "start": 214.83999999999997,
          "end": 214.98,
          "probability": 0.99951171875
        },
        {
          "word": "会",
          "start": 214.98,
          "end": 215.1,
          "probability": 0.99951171875
        },
        {
          "word": "觉得",
          "start": 215.1,
          "end": 215.29999999999998,
          "probability": 1.0
        },
        {
          "word": "这是",
          "start": 215.29999999999998,
          "end": 215.49999999999997,
          "probability": 0.9990234375
        },
        {
          "word": "我",
          "start": 215.49999999999997,
          "end": 215.67999999999998,
          "probability": 1.0
        },
        {
          "word": "提",
          "start": 215.67999999999998,
          "end": 215.83999999999997,
          "probability": 1.0
        },
        {
          "word": "供",
          "start": 215.83999999999997,
          "end": 215.98,
          "probability": 1.0
        },
        {
          "word": "的",
          "start": 215.98,
          "end": 216.07999999999998,
          "probability": 1.0
        },
        {
          "word": "服",
          "start": 216.07999999999998,
          "end": 216.23999999999998,
          "probability": 0.99951171875
        },
        {
          "word": "务",
          "start": 216.23999999999998,
          "end": 216.43999999999997,
          "probability": 1.0
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.14027634287268165,
        "compression_ratio": 1.2254098360655739,
        "no_speech_prob": 4.3821491574336235e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "可只要我把它内置在软件里大家自然会觉得这是我提供的服务我就对它有义务所以我决定把它移除掉当然你依然可以通过现有的自定义配置自主合规地建入从前的index翻译接口但那是你亲手填上去用的你会清楚这是你自己的选择"
        }
      ],
      "id": 95,
      "review_annotation": "",
      "row_id": "95",
      "bounded_row_id": 95,
      "parent_row_ids": [
        95
      ],
      "parent_row_id": 95,
      "packet_id": 8,
      "LID_evidence": [
        {
          "region_id": 41,
          "probabilities": {
            "zh": 0.9978593746821086
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        41
      ],
      "language_confidence": 0.9978593746821086,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 211.29999999999998,
      "packet_end": 235.5,
      "raw_text": "大家自然会觉得这是我提供的服务",
      "selected_text": "大家自然会觉得这是我提供的服务",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            95
          ],
          "packet_id": 8,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        8
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 95
    },
    {
      "start": 216.43999999999997,
      "end": 217.64,
      "text": "我就对它有义务",
      "raw_asr_text": "我就对它有义务",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 8,
      "words": [
        {
          "word": "我就",
          "start": 216.43999999999997,
          "end": 216.83999999999997,
          "probability": 0.9931640625
        },
        {
          "word": "对",
          "start": 216.83999999999997,
          "end": 216.99999999999997,
          "probability": 1.0
        },
        {
          "word": "它",
          "start": 216.99999999999997,
          "end": 217.11999999999998,
          "probability": 0.98095703125
        },
        {
          "word": "有",
          "start": 217.11999999999998,
          "end": 217.29999999999998,
          "probability": 0.99951171875
        },
        {
          "word": "义",
          "start": 217.29999999999998,
          "end": 217.48,
          "probability": 0.99609375
        },
        {
          "word": "务",
          "start": 217.48,
          "end": 217.64,
          "probability": 0.9990234375
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.14027634287268165,
        "compression_ratio": 1.2254098360655739,
        "no_speech_prob": 4.3821491574336235e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "可只要我把它内置在软件里大家自然会觉得这是我提供的服务我就对它有义务所以我决定把它移除掉当然你依然可以通过现有的自定义配置自主合规地建入从前的index翻译接口但那是你亲手填上去用的你会清楚这是你自己的选择"
        }
      ],
      "id": 96,
      "review_annotation": "",
      "row_id": "96",
      "bounded_row_id": 96,
      "parent_row_ids": [
        96
      ],
      "parent_row_id": 96,
      "packet_id": 8,
      "LID_evidence": [
        {
          "region_id": 41,
          "probabilities": {
            "zh": 0.9978593746821086
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        41
      ],
      "language_confidence": 0.9978593746821086,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 211.29999999999998,
      "packet_end": 235.5,
      "raw_text": "我就对它有义务",
      "selected_text": "我就对它有义务",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            96
          ],
          "packet_id": 8,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        8
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 96
    },
    {
      "start": 218.44,
      "end": 220.07999999999998,
      "text": "所以我决定把它移除掉",
      "raw_asr_text": "所以我决定把它移除掉",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 8,
      "words": [
        {
          "word": "所以",
          "start": 218.44,
          "end": 218.79999999999998,
          "probability": 0.99609375
        },
        {
          "word": "我",
          "start": 218.79999999999998,
          "end": 219.16,
          "probability": 0.94921875
        },
        {
          "word": "决",
          "start": 219.16,
          "end": 219.27999999999997,
          "probability": 1.0
        },
        {
          "word": "定",
          "start": 219.27999999999997,
          "end": 219.38,
          "probability": 0.99951171875
        },
        {
          "word": "把它",
          "start": 219.38,
          "end": 219.55999999999997,
          "probability": 0.9990234375
        },
        {
          "word": "移",
          "start": 219.55999999999997,
          "end": 219.76,
          "probability": 0.99609375
        },
        {
          "word": "除",
          "start": 219.76,
          "end": 219.89999999999998,
          "probability": 0.99951171875
        },
        {
          "word": "掉",
          "start": 219.89999999999998,
          "end": 220.07999999999998,
          "probability": 1.0
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.14027634287268165,
        "compression_ratio": 1.2254098360655739,
        "no_speech_prob": 4.3821491574336235e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "可只要我把它内置在软件里大家自然会觉得这是我提供的服务我就对它有义务所以我决定把它移除掉当然你依然可以通过现有的自定义配置自主合规地建入从前的index翻译接口但那是你亲手填上去用的你会清楚这是你自己的选择"
        }
      ],
      "id": 97,
      "review_annotation": "",
      "row_id": "97",
      "bounded_row_id": 97,
      "parent_row_ids": [
        97
      ],
      "parent_row_id": 97,
      "packet_id": 8,
      "LID_evidence": [
        {
          "region_id": 42,
          "probabilities": {
            "zh": 0.9979430511593819
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        42
      ],
      "language_confidence": 0.9979430511593819,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 211.29999999999998,
      "packet_end": 235.5,
      "raw_text": "所以我决定把它移除掉",
      "selected_text": "所以我决定把它移除掉",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            97
          ],
          "packet_id": 8,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        8
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 97
    },
    {
      "start": 221.17999999999998,
      "end": 221.54,
      "text": "当然",
      "raw_asr_text": "当然",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 8,
      "words": [
        {
          "word": "当然",
          "start": 221.17999999999998,
          "end": 221.54,
          "probability": 0.9970703125
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.14027634287268165,
        "compression_ratio": 1.2254098360655739,
        "no_speech_prob": 4.3821491574336235e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "可只要我把它内置在软件里大家自然会觉得这是我提供的服务我就对它有义务所以我决定把它移除掉当然你依然可以通过现有的自定义配置自主合规地建入从前的index翻译接口但那是你亲手填上去用的你会清楚这是你自己的选择"
        }
      ],
      "id": 98,
      "review_annotation": "",
      "row_id": "98",
      "bounded_row_id": 98,
      "parent_row_ids": [
        98
      ],
      "parent_row_id": 98,
      "packet_id": 8,
      "LID_evidence": [
        {
          "region_id": 43,
          "probabilities": {
            "zh": 0.9921650290489197
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        43
      ],
      "language_confidence": 0.9921650290489197,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 211.29999999999998,
      "packet_end": 235.5,
      "raw_text": "当然",
      "selected_text": "当然",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            98
          ],
          "packet_id": 8,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        8
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 98
    },
    {
      "start": 222.33999999999997,
      "end": 224.65999999999997,
      "text": "你依然可以通过现有的自定义配置",
      "raw_asr_text": "你依然可以通过现有的自定义配置",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 8,
      "words": [
        {
          "word": "你",
          "start": 222.33999999999997,
          "end": 222.7,
          "probability": 0.98486328125
        },
        {
          "word": "依",
          "start": 222.7,
          "end": 222.85999999999999,
          "probability": 1.0
        },
        {
          "word": "然",
          "start": 222.85999999999999,
          "end": 222.95999999999998,
          "probability": 0.98876953125
        },
        {
          "word": "可以",
          "start": 222.95999999999998,
          "end": 223.15999999999997,
          "probability": 1.0
        },
        {
          "word": "通",
          "start": 223.15999999999997,
          "end": 223.32,
          "probability": 0.99951171875
        },
        {
          "word": "过",
          "start": 223.32,
          "end": 223.48,
          "probability": 0.99951171875
        },
        {
          "word": "现",
          "start": 223.48,
          "end": 223.64,
          "probability": 0.99658203125
        },
        {
          "word": "有",
          "start": 223.64,
          "end": 223.77999999999997,
          "probability": 1.0
        },
        {
          "word": "的",
          "start": 223.77999999999997,
          "end": 223.85999999999999,
          "probability": 1.0
        },
        {
          "word": "自",
          "start": 223.85999999999999,
          "end": 223.99999999999997,
          "probability": 0.99853515625
        },
        {
          "word": "定",
          "start": 223.99999999999997,
          "end": 224.15999999999997,
          "probability": 0.96923828125
        },
        {
          "word": "义",
          "start": 224.15999999999997,
          "end": 224.26,
          "probability": 0.998046875
        },
        {
          "word": "配",
          "start": 224.26,
          "end": 224.42,
          "probability": 0.9951171875
        },
        {
          "word": "置",
          "start": 224.42,
          "end": 224.65999999999997,
          "probability": 0.98583984375
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.14027634287268165,
        "compression_ratio": 1.2254098360655739,
        "no_speech_prob": 4.3821491574336235e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "可只要我把它内置在软件里大家自然会觉得这是我提供的服务我就对它有义务所以我决定把它移除掉当然你依然可以通过现有的自定义配置自主合规地建入从前的index翻译接口但那是你亲手填上去用的你会清楚这是你自己的选择"
        }
      ],
      "id": 99,
      "review_annotation": "",
      "row_id": "99",
      "bounded_row_id": 99,
      "parent_row_ids": [
        99
      ],
      "parent_row_id": 99,
      "packet_id": 8,
      "LID_evidence": [
        {
          "region_id": 44,
          "probabilities": {
            "zh": 0.98795114670481
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        44
      ],
      "language_confidence": 0.98795114670481,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 211.29999999999998,
      "packet_end": 235.5,
      "raw_text": "你依然可以通过现有的自定义配置",
      "selected_text": "你依然可以通过现有的自定义配置",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            99
          ],
          "packet_id": 8,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        8
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 99
    },
    {
      "start": 224.65999999999997,
      "end": 228.42,
      "text": "自主合规地建入从前的index翻译接口",
      "raw_asr_text": "自主合规地建入从前的index翻译接口",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 8,
      "words": [
        {
          "word": "自",
          "start": 224.65999999999997,
          "end": 225.26,
          "probability": 0.99853515625
        },
        {
          "word": "主",
          "start": 225.26,
          "end": 225.44,
          "probability": 0.99951171875
        },
        {
          "word": "合",
          "start": 225.44,
          "end": 225.57999999999998,
          "probability": 0.9873046875
        },
        {
          "word": "规",
          "start": 225.57999999999998,
          "end": 225.76,
          "probability": 0.986328125
        },
        {
          "word": "地",
          "start": 225.76,
          "end": 225.88,
          "probability": 0.74853515625
        },
        {
          "word": "建",
          "start": 225.88,
          "end": 226.05999999999997,
          "probability": 0.50048828125
        },
        {
          "word": "入",
          "start": 226.05999999999997,
          "end": 226.2,
          "probability": 0.94189453125
        },
        {
          "word": "从",
          "start": 226.2,
          "end": 226.35999999999999,
          "probability": 0.998046875
        },
        {
          "word": "前",
          "start": 226.35999999999999,
          "end": 226.57999999999998,
          "probability": 1.0
        },
        {
          "word": "的",
          "start": 226.57999999999998,
          "end": 226.7,
          "probability": 1.0
        },
        {
          "word": "ind",
          "start": 226.7,
          "end": 226.94,
          "probability": 0.83642578125
        },
        {
          "word": "ex",
          "start": 226.94,
          "end": 227.26,
          "probability": 0.99853515625
        },
        {
          "word": "翻",
          "start": 227.26,
          "end": 227.85999999999999,
          "probability": 0.9921875
        },
        {
          "word": "译",
          "start": 227.85999999999999,
          "end": 228.04,
          "probability": 0.99609375
        },
        {
          "word": "接",
          "start": 228.04,
          "end": 228.15999999999997,
          "probability": 0.87060546875
        },
        {
          "word": "口",
          "start": 228.15999999999997,
          "end": 228.42,
          "probability": 0.9990234375
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.14027634287268165,
        "compression_ratio": 1.2254098360655739,
        "no_speech_prob": 4.3821491574336235e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "可只要我把它内置在软件里大家自然会觉得这是我提供的服务我就对它有义务所以我决定把它移除掉当然你依然可以通过现有的自定义配置自主合规地建入从前的index翻译接口但那是你亲手填上去用的你会清楚这是你自己的选择"
        }
      ],
      "id": 100,
      "review_annotation": "",
      "row_id": "100",
      "bounded_row_id": 100,
      "parent_row_ids": [
        100
      ],
      "parent_row_id": 100,
      "packet_id": 8,
      "LID_evidence": [
        {
          "region_id": 44,
          "probabilities": {
            "zh": 0.98795114670481
          },
          "status": "STABLE"
        },
        {
          "region_id": 45,
          "probabilities": {
            "zh": 0.9774258732795715
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        44,
        45
      ],
      "language_confidence": 0.98795114670481,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 211.29999999999998,
      "packet_end": 235.5,
      "raw_text": "自主合规地建入从前的index翻译接口",
      "selected_text": "自主合规地建入从前的index翻译接口",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            100
          ],
          "packet_id": 8,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        8
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 100
    },
    {
      "start": 228.89999999999998,
      "end": 230.88,
      "text": "但那是你亲手填上去用的",
      "raw_asr_text": "但那是你亲手填上去用的",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 8,
      "words": [
        {
          "word": "但",
          "start": 228.89999999999998,
          "end": 229.26,
          "probability": 0.9833984375
        },
        {
          "word": "那",
          "start": 229.26,
          "end": 229.39999999999998,
          "probability": 0.84619140625
        },
        {
          "word": "是",
          "start": 229.39999999999998,
          "end": 229.55999999999997,
          "probability": 0.99853515625
        },
        {
          "word": "你",
          "start": 229.55999999999997,
          "end": 229.7,
          "probability": 1.0
        },
        {
          "word": "亲",
          "start": 229.7,
          "end": 229.89999999999998,
          "probability": 0.9970703125
        },
        {
          "word": "手",
          "start": 229.89999999999998,
          "end": 230.1,
          "probability": 0.99951171875
        },
        {
          "word": "填",
          "start": 230.1,
          "end": 230.23999999999998,
          "probability": 0.99267578125
        },
        {
          "word": "上",
          "start": 230.23999999999998,
          "end": 230.39999999999998,
          "probability": 1.0
        },
        {
          "word": "去",
          "start": 230.39999999999998,
          "end": 230.54,
          "probability": 1.0
        },
        {
          "word": "用",
          "start": 230.54,
          "end": 230.71999999999997,
          "probability": 1.0
        },
        {
          "word": "的",
          "start": 230.71999999999997,
          "end": 230.88,
          "probability": 1.0
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.14027634287268165,
        "compression_ratio": 1.2254098360655739,
        "no_speech_prob": 4.3821491574336235e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "可只要我把它内置在软件里大家自然会觉得这是我提供的服务我就对它有义务所以我决定把它移除掉当然你依然可以通过现有的自定义配置自主合规地建入从前的index翻译接口但那是你亲手填上去用的你会清楚这是你自己的选择"
        }
      ],
      "id": 101,
      "review_annotation": "",
      "row_id": "101",
      "bounded_row_id": 101,
      "parent_row_ids": [
        101
      ],
      "parent_row_id": 101,
      "packet_id": 8,
      "LID_evidence": [
        {
          "region_id": 45,
          "probabilities": {
            "zh": 0.9774258732795715
          },
          "status": "STABLE"
        },
        {
          "region_id": 46,
          "probabilities": {
            "zh": 0.9953979690869649
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        45,
        46
      ],
      "language_confidence": 0.9953979690869649,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 211.29999999999998,
      "packet_end": 235.5,
      "raw_text": "但那是你亲手填上去用的",
      "selected_text": "但那是你亲手填上去用的",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            101
          ],
          "packet_id": 8,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        8
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 101
    },
    {
      "start": 231.6,
      "end": 232.48,
      "text": "你会清楚",
      "raw_asr_text": "你会清楚",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 8,
      "words": [
        {
          "word": "你",
          "start": 231.6,
          "end": 231.95999999999998,
          "probability": 0.9990234375
        },
        {
          "word": "会",
          "start": 231.95999999999998,
          "end": 232.11999999999998,
          "probability": 0.99951171875
        },
        {
          "word": "清楚",
          "start": 232.11999999999998,
          "end": 232.48,
          "probability": 0.99658203125
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.14027634287268165,
        "compression_ratio": 1.2254098360655739,
        "no_speech_prob": 4.3821491574336235e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "可只要我把它内置在软件里大家自然会觉得这是我提供的服务我就对它有义务所以我决定把它移除掉当然你依然可以通过现有的自定义配置自主合规地建入从前的index翻译接口但那是你亲手填上去用的你会清楚这是你自己的选择"
        }
      ],
      "id": 102,
      "review_annotation": "",
      "row_id": "102",
      "bounded_row_id": 102,
      "parent_row_ids": [
        102
      ],
      "parent_row_id": 102,
      "packet_id": 8,
      "LID_evidence": [
        {
          "region_id": 46,
          "probabilities": {
            "zh": 0.9953979690869649
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        46
      ],
      "language_confidence": 0.9953979690869649,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 211.29999999999998,
      "packet_end": 235.5,
      "raw_text": "你会清楚",
      "selected_text": "你会清楚",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            102
          ],
          "packet_id": 8,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        8
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 102
    },
    {
      "start": 232.48,
      "end": 234.27999999999997,
      "text": "这是你自己的选择",
      "raw_asr_text": "这是你自己的选择",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 8,
      "words": [
        {
          "word": "这是",
          "start": 232.48,
          "end": 233.15999999999997,
          "probability": 0.9951171875
        },
        {
          "word": "你",
          "start": 233.15999999999997,
          "end": 233.39999999999998,
          "probability": 1.0
        },
        {
          "word": "自己",
          "start": 233.39999999999998,
          "end": 233.7,
          "probability": 1.0
        },
        {
          "word": "的",
          "start": 233.7,
          "end": 233.83999999999997,
          "probability": 1.0
        },
        {
          "word": "选",
          "start": 233.83999999999997,
          "end": 234.04,
          "probability": 1.0
        },
        {
          "word": "择",
          "start": 234.04,
          "end": 234.27999999999997,
          "probability": 1.0
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.14027634287268165,
        "compression_ratio": 1.2254098360655739,
        "no_speech_prob": 4.3821491574336235e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "可只要我把它内置在软件里大家自然会觉得这是我提供的服务我就对它有义务所以我决定把它移除掉当然你依然可以通过现有的自定义配置自主合规地建入从前的index翻译接口但那是你亲手填上去用的你会清楚这是你自己的选择"
        }
      ],
      "id": 103,
      "review_annotation": "",
      "row_id": "103",
      "bounded_row_id": 103,
      "parent_row_ids": [
        103
      ],
      "parent_row_id": 103,
      "packet_id": 8,
      "LID_evidence": [
        {
          "region_id": 46,
          "probabilities": {
            "zh": 0.9953979690869649
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        46
      ],
      "language_confidence": 0.9953979690869649,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 211.29999999999998,
      "packet_end": 235.5,
      "raw_text": "这是你自己的选择",
      "selected_text": "这是你自己的选择",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            103
          ],
          "packet_id": 8,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        8
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 103
    },
    {
      "start": 235.14,
      "end": 239.39999999999998,
      "text": "第二个减法是移除了本机大模型部署",
      "raw_asr_text": "第二个减法是移除了本机大模型部署",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 9,
      "words": [
        {
          "word": "第二",
          "start": 235.14,
          "end": 235.54,
          "probability": 0.8486328125
        },
        {
          "word": "个",
          "start": 235.54,
          "end": 235.72,
          "probability": 0.998046875
        },
        {
          "word": "减",
          "start": 235.72,
          "end": 235.88,
          "probability": 0.880859375
        },
        {
          "word": "法",
          "start": 235.88,
          "end": 236.11999999999998,
          "probability": 0.99658203125
        },
        {
          "word": "是",
          "start": 236.11999999999998,
          "end": 236.82,
          "probability": 0.95556640625
        },
        {
          "word": "移",
          "start": 236.82,
          "end": 237.01999999999998,
          "probability": 0.99609375
        },
        {
          "word": "除",
          "start": 237.01999999999998,
          "end": 237.17999999999998,
          "probability": 0.99951171875
        },
        {
          "word": "了",
          "start": 237.17999999999998,
          "end": 237.39999999999998,
          "probability": 0.99853515625
        },
        {
          "word": "本",
          "start": 237.39999999999998,
          "end": 237.89999999999998,
          "probability": 0.99755859375
        },
        {
          "word": "机",
          "start": 237.89999999999998,
          "end": 238.16,
          "probability": 0.546875
        },
        {
          "word": "大",
          "start": 238.16,
          "end": 238.6,
          "probability": 0.99658203125
        },
        {
          "word": "模",
          "start": 238.6,
          "end": 238.79999999999998,
          "probability": 0.99365234375
        },
        {
          "word": "型",
          "start": 238.79999999999998,
          "end": 238.98,
          "probability": 0.99658203125
        },
        {
          "word": "部",
          "start": 238.98,
          "end": 239.16,
          "probability": 0.9990234375
        },
        {
          "word": "署",
          "start": 239.16,
          "end": 239.39999999999998,
          "probability": 0.98388671875
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.10223612934350967,
        "compression_ratio": 1.1865079365079365,
        "no_speech_prob": 1.525546529379529e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "第二个减法是移除了本机大模型部署因为这个项目的重心是做好工具本身也就是模型调用和翻译相关功能的开发优化而在安卓上面对五花八门的系统SoC CPU GPU和NPU让手机连续跑本地模型做实时翻译本身就是个性能功耗和负载炸弹"
        }
      ],
      "id": 104,
      "review_annotation": "",
      "row_id": "104",
      "bounded_row_id": 104,
      "parent_row_ids": [
        104
      ],
      "parent_row_id": 104,
      "packet_id": 9,
      "LID_evidence": [
        {
          "region_id": 47,
          "probabilities": {
            "zh": 0.9968255277602903
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        47
      ],
      "language_confidence": 0.9968255277602903,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 234.2,
      "packet_end": 263.1,
      "raw_text": "第二个减法是移除了本机大模型部署",
      "selected_text": "第二个减法是移除了本机大模型部署",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            104
          ],
          "packet_id": 9,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        9
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 104
    },
    {
      "start": 240.06,
      "end": 243.11999999999998,
      "text": "因为这个项目的重心是做好工具本身",
      "raw_asr_text": "因为这个项目的重心是做好工具本身",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 9,
      "words": [
        {
          "word": "因为",
          "start": 240.06,
          "end": 240.45999999999998,
          "probability": 0.97314453125
        },
        {
          "word": "这个",
          "start": 240.45999999999998,
          "end": 240.73999999999998,
          "probability": 0.99951171875
        },
        {
          "word": "项",
          "start": 240.73999999999998,
          "end": 240.94,
          "probability": 1.0
        },
        {
          "word": "目",
          "start": 240.94,
          "end": 241.04,
          "probability": 1.0
        },
        {
          "word": "的",
          "start": 241.04,
          "end": 241.14,
          "probability": 1.0
        },
        {
          "word": "重",
          "start": 241.14,
          "end": 241.26,
          "probability": 0.9990234375
        },
        {
          "word": "心",
          "start": 241.26,
          "end": 241.54,
          "probability": 0.99365234375
        },
        {
          "word": "是",
          "start": 241.54,
          "end": 242.01999999999998,
          "probability": 0.9990234375
        },
        {
          "word": "做",
          "start": 242.01999999999998,
          "end": 242.2,
          "probability": 0.9990234375
        },
        {
          "word": "好",
          "start": 242.2,
          "end": 242.38,
          "probability": 1.0
        },
        {
          "word": "工",
          "start": 242.38,
          "end": 242.6,
          "probability": 0.99951171875
        },
        {
          "word": "具",
          "start": 242.6,
          "end": 242.7,
          "probability": 0.99951171875
        },
        {
          "word": "本",
          "start": 242.7,
          "end": 242.85999999999999,
          "probability": 0.99951171875
        },
        {
          "word": "身",
          "start": 242.85999999999999,
          "end": 243.11999999999998,
          "probability": 1.0
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.10223612934350967,
        "compression_ratio": 1.1865079365079365,
        "no_speech_prob": 1.525546529379529e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "第二个减法是移除了本机大模型部署因为这个项目的重心是做好工具本身也就是模型调用和翻译相关功能的开发优化而在安卓上面对五花八门的系统SoC CPU GPU和NPU让手机连续跑本地模型做实时翻译本身就是个性能功耗和负载炸弹"
        }
      ],
      "id": 105,
      "review_annotation": "",
      "row_id": "105",
      "bounded_row_id": 105,
      "parent_row_ids": [
        105
      ],
      "parent_row_id": 105,
      "packet_id": 9,
      "LID_evidence": [
        {
          "region_id": 47,
          "probabilities": {
            "zh": 0.9968255277602903
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        47
      ],
      "language_confidence": 0.9968255277602903,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 234.2,
      "packet_end": 263.1,
      "raw_text": "因为这个项目的重心是做好工具本身",
      "selected_text": "因为这个项目的重心是做好工具本身",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            105
          ],
          "packet_id": 9,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        9
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 105
    },
    {
      "start": 243.11999999999998,
      "end": 248.64,
      "text": "也就是模型调用和翻译相关功能的开发优化",
      "raw_asr_text": "也就是模型调用和翻译相关功能的开发优化",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 9,
      "words": [
        {
          "word": "也",
          "start": 243.11999999999998,
          "end": 243.79999999999998,
          "probability": 0.9853515625
        },
        {
          "word": "就是",
          "start": 243.79999999999998,
          "end": 244.2,
          "probability": 1.0
        },
        {
          "word": "模",
          "start": 244.2,
          "end": 244.76,
          "probability": 0.99951171875
        },
        {
          "word": "型",
          "start": 244.76,
          "end": 245.04,
          "probability": 1.0
        },
        {
          "word": "调",
          "start": 245.04,
          "end": 245.32,
          "probability": 0.9951171875
        },
        {
          "word": "用",
          "start": 245.32,
          "end": 245.54,
          "probability": 0.9990234375
        },
        {
          "word": "和",
          "start": 245.54,
          "end": 245.92,
          "probability": 0.9990234375
        },
        {
          "word": "翻",
          "start": 245.92,
          "end": 246.14,
          "probability": 0.99951171875
        },
        {
          "word": "译",
          "start": 246.14,
          "end": 246.39999999999998,
          "probability": 0.998046875
        },
        {
          "word": "相",
          "start": 246.39999999999998,
          "end": 246.78,
          "probability": 0.99951171875
        },
        {
          "word": "关",
          "start": 246.78,
          "end": 247.0,
          "probability": 0.99951171875
        },
        {
          "word": "功",
          "start": 247.0,
          "end": 247.17999999999998,
          "probability": 0.9990234375
        },
        {
          "word": "能",
          "start": 247.17999999999998,
          "end": 247.35999999999999,
          "probability": 0.99951171875
        },
        {
          "word": "的",
          "start": 247.35999999999999,
          "end": 247.54,
          "probability": 0.99951171875
        },
        {
          "word": "开",
          "start": 247.54,
          "end": 247.94,
          "probability": 0.99951171875
        },
        {
          "word": "发",
          "start": 247.94,
          "end": 248.22,
          "probability": 0.9990234375
        },
        {
          "word": "优",
          "start": 248.22,
          "end": 248.38,
          "probability": 0.99951171875
        },
        {
          "word": "化",
          "start": 248.38,
          "end": 248.64,
          "probability": 1.0
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.10223612934350967,
        "compression_ratio": 1.1865079365079365,
        "no_speech_prob": 1.525546529379529e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "第二个减法是移除了本机大模型部署因为这个项目的重心是做好工具本身也就是模型调用和翻译相关功能的开发优化而在安卓上面对五花八门的系统SoC CPU GPU和NPU让手机连续跑本地模型做实时翻译本身就是个性能功耗和负载炸弹"
        }
      ],
      "id": 106,
      "review_annotation": "",
      "row_id": "106",
      "bounded_row_id": 106,
      "parent_row_ids": [
        106
      ],
      "parent_row_id": 106,
      "packet_id": 9,
      "LID_evidence": [
        {
          "region_id": 47,
          "probabilities": {
            "zh": 0.9968255277602903
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        47
      ],
      "language_confidence": 0.9968255277602903,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 234.2,
      "packet_end": 263.1,
      "raw_text": "也就是模型调用和翻译相关功能的开发优化",
      "selected_text": "也就是模型调用和翻译相关功能的开发优化",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            106
          ],
          "packet_id": 9,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        9
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 106
    },
    {
      "start": 249.07999999999998,
      "end": 251.32,
      "text": "而在安卓上面对五花八门的系统",
      "raw_asr_text": "而在安卓上面对五花八门的系统",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 9,
      "words": [
        {
          "word": "而",
          "start": 249.07999999999998,
          "end": 249.48,
          "probability": 0.9990234375
        },
        {
          "word": "在",
          "start": 249.48,
          "end": 249.64,
          "probability": 1.0
        },
        {
          "word": "安",
          "start": 249.64,
          "end": 249.78,
          "probability": 0.94921875
        },
        {
          "word": "卓",
          "start": 249.78,
          "end": 249.89999999999998,
          "probability": 0.99755859375
        },
        {
          "word": "上面",
          "start": 249.89999999999998,
          "end": 250.17999999999998,
          "probability": 0.99072265625
        },
        {
          "word": "对",
          "start": 250.17999999999998,
          "end": 250.35999999999999,
          "probability": 0.94970703125
        },
        {
          "word": "五",
          "start": 250.35999999999999,
          "end": 250.48,
          "probability": 0.8916015625
        },
        {
          "word": "花",
          "start": 250.48,
          "end": 250.6,
          "probability": 0.990234375
        },
        {
          "word": "八",
          "start": 250.6,
          "end": 250.76,
          "probability": 0.96875
        },
        {
          "word": "门",
          "start": 250.76,
          "end": 250.83999999999997,
          "probability": 0.984375
        },
        {
          "word": "的",
          "start": 250.83999999999997,
          "end": 250.98,
          "probability": 0.9990234375
        },
        {
          "word": "系",
          "start": 250.98,
          "end": 251.12,
          "probability": 1.0
        },
        {
          "word": "统",
          "start": 251.12,
          "end": 251.32,
          "probability": 0.99951171875
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.10223612934350967,
        "compression_ratio": 1.1865079365079365,
        "no_speech_prob": 1.525546529379529e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "第二个减法是移除了本机大模型部署因为这个项目的重心是做好工具本身也就是模型调用和翻译相关功能的开发优化而在安卓上面对五花八门的系统SoC CPU GPU和NPU让手机连续跑本地模型做实时翻译本身就是个性能功耗和负载炸弹"
        }
      ],
      "id": 107,
      "review_annotation": "",
      "row_id": "107",
      "bounded_row_id": 107,
      "parent_row_ids": [
        107
      ],
      "parent_row_id": 107,
      "packet_id": 9,
      "LID_evidence": [
        {
          "region_id": 48,
          "probabilities": {
            "zh": 0.8436801614158828
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        48
      ],
      "language_confidence": 0.8436801614158828,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 234.2,
      "packet_end": 263.1,
      "raw_text": "而在安卓上面对五花八门的系统",
      "selected_text": "而在安卓上面对五花八门的系统",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            107
          ],
          "packet_id": 9,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        9
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 107
    },
    {
      "start": 251.32,
      "end": 254.95999999999998,
      "text": "SoC CPU GPU和NPU",
      "raw_asr_text": "SoC CPU GPU和NPU",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 9,
      "words": [
        {
          "word": "So",
          "start": 251.32,
          "end": 251.88,
          "probability": 0.642578125
        },
        {
          "word": "C",
          "start": 251.88,
          "end": 252.22,
          "probability": 0.9990234375
        },
        {
          "word": " CPU",
          "start": 252.22,
          "end": 253.07999999999998,
          "probability": 0.7939453125
        },
        {
          "word": " GPU",
          "start": 253.07999999999998,
          "end": 253.92,
          "probability": 0.9921875
        },
        {
          "word": "和",
          "start": 253.92,
          "end": 254.54,
          "probability": 0.96923828125
        },
        {
          "word": "N",
          "start": 254.54,
          "end": 254.67999999999998,
          "probability": 0.970703125
        },
        {
          "word": "PU",
          "start": 254.67999999999998,
          "end": 254.95999999999998,
          "probability": 0.99462890625
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.10223612934350967,
        "compression_ratio": 1.1865079365079365,
        "no_speech_prob": 1.525546529379529e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "第二个减法是移除了本机大模型部署因为这个项目的重心是做好工具本身也就是模型调用和翻译相关功能的开发优化而在安卓上面对五花八门的系统SoC CPU GPU和NPU让手机连续跑本地模型做实时翻译本身就是个性能功耗和负载炸弹"
        }
      ],
      "id": 108,
      "review_annotation": "",
      "row_id": "108",
      "bounded_row_id": 108,
      "parent_row_ids": [
        108
      ],
      "parent_row_id": 108,
      "packet_id": 9,
      "LID_evidence": [
        {
          "region_id": 48,
          "probabilities": {
            "zh": 0.8436801614158828
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        48
      ],
      "language_confidence": 0.8436801614158828,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 234.2,
      "packet_end": 263.1,
      "raw_text": "SoC CPU GPU和NPU",
      "selected_text": "SoC CPU GPU和NPU",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            108
          ],
          "packet_id": 9,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        9
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "independent_ASR": {
        "engine": "whisper-large-v3",
        "model_path": "/Users/sagawa/AI_WORK/cache/huggingface/hub/models--mlx-community--whisper-large-v3",
        "language": "zh",
        "audio_interval": [
          249.32,
          256.96
        ],
        "target_interval": [
          251.32,
          254.95999999999998
        ],
        "text": "SoCCPUGPU和NPU",
        "words": [
          {
            "word": "So",
            "start": 251.23999999999998,
            "end": 251.76,
            "probability": 0.3916015625
          },
          {
            "word": "C",
            "start": 251.76,
            "end": 252.07999999999998,
            "probability": 0.9990234375
          },
          {
            "word": " CPU",
            "start": 252.07999999999998,
            "end": 252.79999999999998,
            "probability": 0.576171875
          },
          {
            "word": " GPU",
            "start": 252.79999999999998,
            "end": 253.73999999999998,
            "probability": 0.96435546875
          },
          {
            "word": "和",
            "start": 253.73999999999998,
            "end": 254.44,
            "probability": 0.94775390625
          },
          {
            "word": "N",
            "start": 254.44,
            "end": 254.6,
            "probability": 0.83544921875
          },
          {
            "word": "PU",
            "start": 254.6,
            "end": 254.84,
            "probability": 0.998046875
          }
        ],
        "raw_result": {
          "text": "而在安卓上面对五花八门的系统SoC CPU GPU和NPU让手机连续跑本地",
          "segments": [
            {
              "id": 0,
              "seek": 0,
              "start": 0.0,
              "end": 1.92,
              "text": "而在安卓上面对五花八门的系统",
              "tokens": [
                50365,
                11070,
                3581,
                16206,
                5322,
                241,
                49750,
                8713,
                21001,
                20127,
                33453,
                8259,
                101,
                1546,
                25368,
                10115,
                253,
                50464
              ],
              "temperature": 0.0,
              "avg_logprob": -0.09962893724441528,
              "compression_ratio": 0.8829787234042553,
              "no_speech_prob": 0.2201370894908905,
              "words": [
                {
                  "word": "而",
                  "start": 0.0,
                  "end": 0.14,
                  "probability": 0.96875
                },
                {
                  "word": "在",
                  "start": 0.14,
                  "end": 0.24,
                  "probability": 0.99267578125
                },
                {
                  "word": "安",
                  "start": 0.24,
                  "end": 0.36,
                  "probability": 0.962890625
                },
                {
                  "word": "卓",
                  "start": 0.36,
                  "end": 0.6,
                  "probability": 0.9990234375
                },
                {
                  "word": "上面",
                  "start": 0.6,
                  "end": 0.76,
                  "probability": 0.93505859375
                },
                {
                  "word": "对",
                  "start": 0.76,
                  "end": 0.98,
                  "probability": 0.91650390625
                },
                {
                  "word": "五",
                  "start": 0.98,
                  "end": 1.1,
                  "probability": 0.84033203125
                },
                {
                  "word": "花",
                  "start": 1.1,
                  "end": 1.24,
                  "probability": 0.96826171875
                },
                {
                  "word": "八",
                  "start": 1.24,
                  "end": 1.36,
                  "probability": 0.99951171875
                },
                {
                  "word": "门",
                  "start": 1.36,
                  "end": 1.54,
                  "probability": 0.998046875
                },
                {
                  "word": "的",
                  "start": 1.54,
                  "end": 1.6,
                  "probability": 0.98193359375
                },
                {
                  "word": "系",
                  "start": 1.6,
                  "end": 1.7,
                  "probability": 0.99755859375
                },
                {
                  "word": "统",
                  "start": 1.7,
                  "end": 1.92,
                  "probability": 1.0
                }
              ]
            },
            {
              "id": 1,
              "seek": 0,
              "start": 1.92,
              "end": 5.52,
              "text": "SoC CPU GPU和NPU",
              "tokens": [
                50464,
                6455,
                34,
                13199,
                18407,
                12565,
                45,
                8115,
                50651
              ],
              "temperature": 0.0,
              "avg_logprob": -0.09962893724441528,
              "compression_ratio": 0.8829787234042553,
              "no_speech_prob": 0.2201370894908905,
              "words": [
                {
                  "word": "So",
                  "start": 1.92,
                  "end": 2.44,
                  "probability": 0.3916015625
                },
                {
                  "word": "C",
                  "start": 2.44,
                  "end": 2.76,
                  "probability": 0.9990234375
                },
                {
                  "word": " CPU",
                  "start": 2.76,
                  "end": 3.48,
                  "probability": 0.576171875
                },
                {
                  "word": " GPU",
                  "start": 3.48,
                  "end": 4.42,
                  "probability": 0.96435546875
                },
                {
                  "word": "和",
                  "start": 4.42,
                  "end": 5.12,
                  "probability": 0.94775390625
                },
                {
                  "word": "N",
                  "start": 5.12,
                  "end": 5.28,
                  "probability": 0.83544921875
                },
                {
                  "word": "PU",
                  "start": 5.28,
                  "end": 5.52,
                  "probability": 0.998046875
                }
              ]
            },
            {
              "id": 2,
              "seek": 0,
              "start": 6.0200000000000005,
              "end": 7.5,
              "text": "让手机连续跑本地",
              "tokens": [
                50651,
                33650,
                11389,
                37960,
                3316,
                252,
                10115,
                255,
                32585,
                8802,
                10928,
                50744
              ],
              "temperature": 0.0,
              "avg_logprob": -0.09962893724441528,
              "compression_ratio": 0.8829787234042553,
              "no_speech_prob": 0.2201370894908905,
              "words": [
                {
                  "word": "让",
                  "start": 6.0200000000000005,
                  "end": 6.38,
                  "probability": 0.9912109375
                },
                {
                  "word": "手",
                  "start": 6.38,
                  "end": 6.56,
                  "probability": 0.99951171875
                },
                {
                  "word": "机",
                  "start": 6.56,
                  "end": 6.78,
                  "probability": 0.99951171875
                },
                {
                  "word": "连",
                  "start": 6.78,
                  "end": 6.96,
                  "probability": 0.9990234375
                },
                {
                  "word": "续",
                  "start": 6.96,
                  "end": 7.08,
                  "probability": 1.0
                },
                {
                  "word": "跑",
                  "start": 7.08,
                  "end": 7.16,
                  "probability": 0.99951171875
                },
                {
                  "word": "本",
                  "start": 7.16,
                  "end": 7.34,
                  "probability": 0.98974609375
                },
                {
                  "word": "地",
                  "start": 7.34,
                  "end": 7.5,
                  "probability": 0.9951171875
                }
              ]
            }
          ],
          "language": "zh",
          "duration": 7.64
        },
        "status": "COMPLETE",
        "requested_language": "zh"
      },
      "index": 108
    },
    {
      "start": 255.39999999999998,
      "end": 258.38,
      "text": "让手机连续跑本地模型做实时翻译",
      "raw_asr_text": "让手机连续跑本地模型做实时翻译",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 9,
      "words": [
        {
          "word": "让",
          "start": 255.39999999999998,
          "end": 255.79999999999998,
          "probability": 0.99853515625
        },
        {
          "word": "手",
          "start": 255.79999999999998,
          "end": 255.98,
          "probability": 1.0
        },
        {
          "word": "机",
          "start": 255.98,
          "end": 256.12,
          "probability": 0.99951171875
        },
        {
          "word": "连",
          "start": 256.12,
          "end": 256.28,
          "probability": 0.99755859375
        },
        {
          "word": "续",
          "start": 256.28,
          "end": 256.4,
          "probability": 0.9990234375
        },
        {
          "word": "跑",
          "start": 256.4,
          "end": 256.53999999999996,
          "probability": 0.9970703125
        },
        {
          "word": "本",
          "start": 256.53999999999996,
          "end": 256.74,
          "probability": 0.99951171875
        },
        {
          "word": "地",
          "start": 256.74,
          "end": 256.88,
          "probability": 0.59423828125
        },
        {
          "word": "模",
          "start": 256.88,
          "end": 257.03999999999996,
          "probability": 0.99951171875
        },
        {
          "word": "型",
          "start": 257.03999999999996,
          "end": 257.3,
          "probability": 1.0
        },
        {
          "word": "做",
          "start": 257.3,
          "end": 257.65999999999997,
          "probability": 0.994140625
        },
        {
          "word": "实",
          "start": 257.65999999999997,
          "end": 257.84,
          "probability": 0.98828125
        },
        {
          "word": "时",
          "start": 257.84,
          "end": 257.98,
          "probability": 0.77587890625
        },
        {
          "word": "翻",
          "start": 257.98,
          "end": 258.15999999999997,
          "probability": 1.0
        },
        {
          "word": "译",
          "start": 258.15999999999997,
          "end": 258.38,
          "probability": 0.9990234375
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.10223612934350967,
        "compression_ratio": 1.1865079365079365,
        "no_speech_prob": 1.525546529379529e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "第二个减法是移除了本机大模型部署因为这个项目的重心是做好工具本身也就是模型调用和翻译相关功能的开发优化而在安卓上面对五花八门的系统SoC CPU GPU和NPU让手机连续跑本地模型做实时翻译本身就是个性能功耗和负载炸弹"
        }
      ],
      "id": 109,
      "review_annotation": "",
      "row_id": "109",
      "bounded_row_id": 109,
      "parent_row_ids": [
        109
      ],
      "parent_row_id": 109,
      "packet_id": 9,
      "LID_evidence": [
        {
          "region_id": 48,
          "probabilities": {
            "zh": 0.8436801614158828
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        48
      ],
      "language_confidence": 0.8436801614158828,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 234.2,
      "packet_end": 263.1,
      "raw_text": "让手机连续跑本地模型做实时翻译",
      "selected_text": "让手机连续跑本地模型做实时翻译",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            109
          ],
          "packet_id": 9,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        9
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 109
    },
    {
      "start": 258.38,
      "end": 261.98,
      "text": "本身就是个性能功耗和负载炸弹",
      "raw_asr_text": "本身就是个性能功耗和负载炸弹",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 9,
      "words": [
        {
          "word": "本",
          "start": 258.38,
          "end": 259.02,
          "probability": 0.99609375
        },
        {
          "word": "身",
          "start": 259.02,
          "end": 259.26,
          "probability": 0.99951171875
        },
        {
          "word": "就是",
          "start": 259.26,
          "end": 259.76,
          "probability": 0.99951171875
        },
        {
          "word": "个",
          "start": 259.76,
          "end": 259.94,
          "probability": 0.99853515625
        },
        {
          "word": "性",
          "start": 259.94,
          "end": 260.12,
          "probability": 0.9990234375
        },
        {
          "word": "能",
          "start": 260.12,
          "end": 260.36,
          "probability": 0.99951171875
        },
        {
          "word": "功",
          "start": 260.36,
          "end": 260.78,
          "probability": 0.81884765625
        },
        {
          "word": "耗",
          "start": 260.78,
          "end": 261.02,
          "probability": 0.99609375
        },
        {
          "word": "和",
          "start": 261.02,
          "end": 261.4,
          "probability": 0.99365234375
        },
        {
          "word": "负",
          "start": 261.4,
          "end": 261.52,
          "probability": 0.9814453125
        },
        {
          "word": "载",
          "start": 261.52,
          "end": 261.65999999999997,
          "probability": 0.9990234375
        },
        {
          "word": "炸",
          "start": 261.65999999999997,
          "end": 261.82,
          "probability": 0.95703125
        },
        {
          "word": "弹",
          "start": 261.82,
          "end": 261.98,
          "probability": 0.9990234375
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.10223612934350967,
        "compression_ratio": 1.1865079365079365,
        "no_speech_prob": 1.525546529379529e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "第二个减法是移除了本机大模型部署因为这个项目的重心是做好工具本身也就是模型调用和翻译相关功能的开发优化而在安卓上面对五花八门的系统SoC CPU GPU和NPU让手机连续跑本地模型做实时翻译本身就是个性能功耗和负载炸弹"
        }
      ],
      "id": 110,
      "review_annotation": "",
      "row_id": "110",
      "bounded_row_id": 110,
      "parent_row_ids": [
        110
      ],
      "parent_row_id": 110,
      "packet_id": 9,
      "LID_evidence": [
        {
          "region_id": 48,
          "probabilities": {
            "zh": 0.8436801614158828
          },
          "status": "STABLE"
        },
        {
          "region_id": 49,
          "probabilities": {
            "zh": 0.9065634808079763
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        48,
        49
      ],
      "language_confidence": 0.9065634808079763,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 234.2,
      "packet_end": 263.1,
      "raw_text": "本身就是个性能功耗和负载炸弹",
      "selected_text": "本身就是个性能功耗和负载炸弹",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            110
          ],
          "packet_id": 9,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        9
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 110
    },
    {
      "start": 262.96,
      "end": 265.09999999999997,
      "text": "ows和Mac上也是一样",
      "raw_asr_text": "Windows和Mac上也是一样",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 10,
      "words": [],
      "asr_metrics": {
        "avg_logprob": -0.09083681895320576,
        "compression_ratio": 1.189922480620155,
        "no_speech_prob": 2.283331683450296e-11
      },
      "lid_status": "UNCERTAIN",
      "alternatives": [
        {
          "language": "zh",
          "text": "Windows和Mac上也是一样不仅要适配各种显卡和统一内存而且上了桌面端大家肯定会想要更强的模型而不只是两臂的量化小模型继续支持这些会给项目引入大量非必要的开发成本如果你想在本地部署模型其实有很多更优秀的专门项目用它"
        }
      ],
      "id": 111,
      "review_annotation": "待確認",
      "row_id": "111",
      "bounded_row_id": 111,
      "parent_row_ids": [
        111
      ],
      "parent_row_id": 111,
      "packet_id": 10,
      "LID_evidence": [
        {
          "region_id": 50,
          "probabilities": {
            "zh": 0.6633929526433349
          },
          "status": "UNCERTAIN"
        },
        {
          "region_id": 51,
          "probabilities": {
            "zh": 0.7127072313534362
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        50,
        51
      ],
      "language_confidence": 0.7127072313534362,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 261.9,
      "packet_end": 288.40000000000003,
      "raw_text": "Windows和Mac上也是一样",
      "selected_text": "ows和Mac上也是一样",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            111
          ],
          "packet_id": 10,
          "engine": "mlx-whisper"
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "111",
          "before": "Windows和Mac上也是一样",
          "after": "ows和Mac上也是一样",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "STRONG_RESTORATION",
          "reason": "Audio-backed final language reconciliation",
          "context_evidence": "whisper-large-v3",
          "independent_ASR": {
            "engine": "whisper-large-v3",
            "model_path": "/Users/sagawa/AI_WORK/cache/huggingface/hub/models--mlx-community--whisper-large-v3",
            "language": "zh",
            "audio_interval": [
              260.96,
              267.09999999999997
            ],
            "target_interval": [
              262.96,
              265.09999999999997
            ],
            "text": "ows和Mac上也是一样",
            "words": [
              {
                "word": "ows",
                "start": 263.14,
                "end": 263.53999999999996,
                "probability": 0.99755859375
              },
              {
                "word": "和",
                "start": 263.53999999999996,
                "end": 263.97999999999996,
                "probability": 0.98779296875
              },
              {
                "word": "Mac",
                "start": 263.97999999999996,
                "end": 264.26,
                "probability": 0.98779296875
              },
              {
                "word": "上",
                "start": 264.26,
                "end": 264.44,
                "probability": 0.994140625
              },
              {
                "word": "也是",
                "start": 264.44,
                "end": 264.65999999999997,
                "probability": 0.99853515625
              },
              {
                "word": "一",
                "start": 264.65999999999997,
                "end": 264.85999999999996,
                "probability": 0.99951171875
              },
              {
                "word": "样",
                "start": 264.85999999999996,
                "end": 265.06,
                "probability": 1.0
              }
            ],
            "raw_result": {
              "text": "和负载炸弹Windows和Mac上也是一样不仅要适配各种显卡",
              "segments": [
                {
                  "id": 0,
                  "seek": 0,
                  "start": 0.0,
                  "end": 1.0,
                  "text": "和负载炸弹",
                  "tokens": [
                    50365,
                    12565,
                    18464,
                    253,
                    17819,
                    121,
                    4804,
                    116,
                    5702,
                    117,
                    50418
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.09749541680018108,
                  "compression_ratio": 0.8641975308641975,
                  "no_speech_prob": 0.7773045897483826,
                  "words": [
                    {
                      "word": "和",
                      "start": 0.0,
                      "end": 0.38,
                      "probability": 0.72021484375
                    },
                    {
                      "word": "负",
                      "start": 0.38,
                      "end": 0.54,
                      "probability": 0.84619140625
                    },
                    {
                      "word": "载",
                      "start": 0.54,
                      "end": 0.7,
                      "probability": 0.998046875
                    },
                    {
                      "word": "炸",
                      "start": 0.7,
                      "end": 0.86,
                      "probability": 0.99755859375
                    },
                    {
                      "word": "弹",
                      "start": 0.86,
                      "end": 1.0,
                      "probability": 1.0
                    }
                  ]
                },
                {
                  "id": 1,
                  "seek": 0,
                  "start": 1.7800000000000002,
                  "end": 4.1,
                  "text": "Windows和Mac上也是一样",
                  "tokens": [
                    50418,
                    45813,
                    1509,
                    12565,
                    34355,
                    5708,
                    22021,
                    2257,
                    14496,
                    50574
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.09749541680018108,
                  "compression_ratio": 0.8641975308641975,
                  "no_speech_prob": 0.7773045897483826,
                  "words": [
                    {
                      "word": "Wind",
                      "start": 1.7800000000000002,
                      "end": 2.18,
                      "probability": 0.93701171875
                    },
                    {
                      "word": "ows",
                      "start": 2.18,
                      "end": 2.58,
                      "probability": 0.99755859375
                    },
                    {
                      "word": "和",
                      "start": 2.58,
                      "end": 3.02,
                      "probability": 0.98779296875
                    },
                    {
                      "word": "Mac",
                      "start": 3.02,
                      "end": 3.3,
                      "probability": 0.98779296875
                    },
                    {
                      "word": "上",
                      "start": 3.3,
                      "end": 3.48,
                      "probability": 0.994140625
                    },
                    {
                      "word": "也是",
                      "start": 3.48,
                      "end": 3.7,
                      "probability": 0.99853515625
                    },
                    {
                      "word": "一",
                      "start": 3.7,
                      "end": 3.9,
                      "probability": 0.99951171875
                    },
                    {
                      "word": "样",
                      "start": 3.9,
                      "end": 4.1,
                      "probability": 1.0
                    }
                  ]
                },
                {
                  "id": 2,
                  "seek": 0,
                  "start": 4.1,
                  "end": 6.06,
                  "text": "不仅要适配各种显卡",
                  "tokens": [
                    50574,
                    1960,
                    1550,
                    227,
                    4275,
                    2215,
                    224,
                    38846,
                    17516,
                    39810,
                    1431,
                    122,
                    32681,
                    50672
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.09749541680018108,
                  "compression_ratio": 0.8641975308641975,
                  "no_speech_prob": 0.7773045897483826,
                  "words": [
                    {
                      "word": "不",
                      "start": 4.1,
                      "end": 4.76,
                      "probability": 0.98486328125
                    },
                    {
                      "word": "仅",
                      "start": 4.76,
                      "end": 4.98,
                      "probability": 0.99560546875
                    },
                    {
                      "word": "要",
                      "start": 4.98,
                      "end": 5.08,
                      "probability": 0.9990234375
                    },
                    {
                      "word": "适",
                      "start": 5.08,
                      "end": 5.32,
                      "probability": 1.0
                    },
                    {
                      "word": "配",
                      "start": 5.32,
                      "end": 5.46,
                      "probability": 0.99853515625
                    },
                    {
                      "word": "各",
                      "start": 5.46,
                      "end": 5.6,
                      "probability": 0.9990234375
                    },
                    {
                      "word": "种",
                      "start": 5.6,
                      "end": 5.8,
                      "probability": 1.0
                    },
                    {
                      "word": "显",
                      "start": 5.8,
                      "end": 6.02,
                      "probability": 0.998046875
                    },
                    {
                      "word": "卡",
                      "start": 6.02,
                      "end": 6.06,
                      "probability": 0.994140625
                    }
                  ]
                }
              ],
              "language": "zh",
              "duration": 6.14
            },
            "status": "COMPLETE",
            "requested_language": "zh"
          },
          "suppression_proposed": null
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "111",
          "before": "ows和Mac上也是一样",
          "after": "ows和Mac上也是一样",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "ows疑似iOS误识",
          "context_evidence": "上下文为Mac和iOS平台，'iOS'为常见系统",
          "independent_ASR": {
            "engine": "whisper-large-v3",
            "model_path": "/Users/sagawa/AI_WORK/cache/huggingface/hub/models--mlx-community--whisper-large-v3",
            "language": "zh",
            "audio_interval": [
              260.96,
              267.09999999999997
            ],
            "target_interval": [
              262.96,
              265.09999999999997
            ],
            "text": "ows和Mac上也是一样",
            "words": [
              {
                "word": "ows",
                "start": 263.14,
                "end": 263.53999999999996,
                "probability": 0.99755859375
              },
              {
                "word": "和",
                "start": 263.53999999999996,
                "end": 263.97999999999996,
                "probability": 0.98779296875
              },
              {
                "word": "Mac",
                "start": 263.97999999999996,
                "end": 264.26,
                "probability": 0.98779296875
              },
              {
                "word": "上",
                "start": 264.26,
                "end": 264.44,
                "probability": 0.994140625
              },
              {
                "word": "也是",
                "start": 264.44,
                "end": 264.65999999999997,
                "probability": 0.99853515625
              },
              {
                "word": "一",
                "start": 264.65999999999997,
                "end": 264.85999999999996,
                "probability": 0.99951171875
              },
              {
                "word": "样",
                "start": 264.85999999999996,
                "end": 265.06,
                "probability": 1.0
              }
            ],
            "raw_result": {
              "text": "和负载炸弹Windows和Mac上也是一样不仅要适配各种显卡",
              "segments": [
                {
                  "id": 0,
                  "seek": 0,
                  "start": 0.0,
                  "end": 1.0,
                  "text": "和负载炸弹",
                  "tokens": [
                    50365,
                    12565,
                    18464,
                    253,
                    17819,
                    121,
                    4804,
                    116,
                    5702,
                    117,
                    50418
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.09749541680018108,
                  "compression_ratio": 0.8641975308641975,
                  "no_speech_prob": 0.7773045897483826,
                  "words": [
                    {
                      "word": "和",
                      "start": 0.0,
                      "end": 0.38,
                      "probability": 0.72021484375
                    },
                    {
                      "word": "负",
                      "start": 0.38,
                      "end": 0.54,
                      "probability": 0.84619140625
                    },
                    {
                      "word": "载",
                      "start": 0.54,
                      "end": 0.7,
                      "probability": 0.998046875
                    },
                    {
                      "word": "炸",
                      "start": 0.7,
                      "end": 0.86,
                      "probability": 0.99755859375
                    },
                    {
                      "word": "弹",
                      "start": 0.86,
                      "end": 1.0,
                      "probability": 1.0
                    }
                  ]
                },
                {
                  "id": 1,
                  "seek": 0,
                  "start": 1.7800000000000002,
                  "end": 4.1,
                  "text": "Windows和Mac上也是一样",
                  "tokens": [
                    50418,
                    45813,
                    1509,
                    12565,
                    34355,
                    5708,
                    22021,
                    2257,
                    14496,
                    50574
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.09749541680018108,
                  "compression_ratio": 0.8641975308641975,
                  "no_speech_prob": 0.7773045897483826,
                  "words": [
                    {
                      "word": "Wind",
                      "start": 1.7800000000000002,
                      "end": 2.18,
                      "probability": 0.93701171875
                    },
                    {
                      "word": "ows",
                      "start": 2.18,
                      "end": 2.58,
                      "probability": 0.99755859375
                    },
                    {
                      "word": "和",
                      "start": 2.58,
                      "end": 3.02,
                      "probability": 0.98779296875
                    },
                    {
                      "word": "Mac",
                      "start": 3.02,
                      "end": 3.3,
                      "probability": 0.98779296875
                    },
                    {
                      "word": "上",
                      "start": 3.3,
                      "end": 3.48,
                      "probability": 0.994140625
                    },
                    {
                      "word": "也是",
                      "start": 3.48,
                      "end": 3.7,
                      "probability": 0.99853515625
                    },
                    {
                      "word": "一",
                      "start": 3.7,
                      "end": 3.9,
                      "probability": 0.99951171875
                    },
                    {
                      "word": "样",
                      "start": 3.9,
                      "end": 4.1,
                      "probability": 1.0
                    }
                  ]
                },
                {
                  "id": 2,
                  "seek": 0,
                  "start": 4.1,
                  "end": 6.06,
                  "text": "不仅要适配各种显卡",
                  "tokens": [
                    50574,
                    1960,
                    1550,
                    227,
                    4275,
                    2215,
                    224,
                    38846,
                    17516,
                    39810,
                    1431,
                    122,
                    32681,
                    50672
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.09749541680018108,
                  "compression_ratio": 0.8641975308641975,
                  "no_speech_prob": 0.7773045897483826,
                  "words": [
                    {
                      "word": "不",
                      "start": 4.1,
                      "end": 4.76,
                      "probability": 0.98486328125
                    },
                    {
                      "word": "仅",
                      "start": 4.76,
                      "end": 4.98,
                      "probability": 0.99560546875
                    },
                    {
                      "word": "要",
                      "start": 4.98,
                      "end": 5.08,
                      "probability": 0.9990234375
                    },
                    {
                      "word": "适",
                      "start": 5.08,
                      "end": 5.32,
                      "probability": 1.0
                    },
                    {
                      "word": "配",
                      "start": 5.32,
                      "end": 5.46,
                      "probability": 0.99853515625
                    },
                    {
                      "word": "各",
                      "start": 5.46,
                      "end": 5.6,
                      "probability": 0.9990234375
                    },
                    {
                      "word": "种",
                      "start": 5.6,
                      "end": 5.8,
                      "probability": 1.0
                    },
                    {
                      "word": "显",
                      "start": 5.8,
                      "end": 6.02,
                      "probability": 0.998046875
                    },
                    {
                      "word": "卡",
                      "start": 6.02,
                      "end": 6.06,
                      "probability": 0.994140625
                    }
                  ]
                }
              ],
              "language": "zh",
              "duration": 6.14
            },
            "status": "COMPLETE",
            "requested_language": "zh"
          },
          "suppression_proposed": null
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "111",
          "before": "ows和Mac上也是一样",
          "after": "ows和Mac上也是一样",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "Name error: 'ows' vs 'Windows'",
          "context_evidence": "Context: 'Windows和Mac上也是一样'",
          "independent_ASR": {
            "engine": "whisper-large-v3",
            "model_path": "/Users/sagawa/AI_WORK/cache/huggingface/hub/models--mlx-community--whisper-large-v3",
            "language": "zh",
            "audio_interval": [
              260.96,
              267.09999999999997
            ],
            "target_interval": [
              262.96,
              265.09999999999997
            ],
            "text": "ows和Mac上也是一样",
            "words": [
              {
                "word": "ows",
                "start": 263.14,
                "end": 263.53999999999996,
                "probability": 0.99755859375
              },
              {
                "word": "和",
                "start": 263.53999999999996,
                "end": 263.97999999999996,
                "probability": 0.98779296875
              },
              {
                "word": "Mac",
                "start": 263.97999999999996,
                "end": 264.26,
                "probability": 0.98779296875
              },
              {
                "word": "上",
                "start": 264.26,
                "end": 264.44,
                "probability": 0.994140625
              },
              {
                "word": "也是",
                "start": 264.44,
                "end": 264.65999999999997,
                "probability": 0.99853515625
              },
              {
                "word": "一",
                "start": 264.65999999999997,
                "end": 264.85999999999996,
                "probability": 0.99951171875
              },
              {
                "word": "样",
                "start": 264.85999999999996,
                "end": 265.06,
                "probability": 1.0
              }
            ],
            "raw_result": {
              "text": "和负载炸弹Windows和Mac上也是一样不仅要适配各种显卡",
              "segments": [
                {
                  "id": 0,
                  "seek": 0,
                  "start": 0.0,
                  "end": 1.0,
                  "text": "和负载炸弹",
                  "tokens": [
                    50365,
                    12565,
                    18464,
                    253,
                    17819,
                    121,
                    4804,
                    116,
                    5702,
                    117,
                    50418
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.09749541680018108,
                  "compression_ratio": 0.8641975308641975,
                  "no_speech_prob": 0.7773045897483826,
                  "words": [
                    {
                      "word": "和",
                      "start": 0.0,
                      "end": 0.38,
                      "probability": 0.72021484375
                    },
                    {
                      "word": "负",
                      "start": 0.38,
                      "end": 0.54,
                      "probability": 0.84619140625
                    },
                    {
                      "word": "载",
                      "start": 0.54,
                      "end": 0.7,
                      "probability": 0.998046875
                    },
                    {
                      "word": "炸",
                      "start": 0.7,
                      "end": 0.86,
                      "probability": 0.99755859375
                    },
                    {
                      "word": "弹",
                      "start": 0.86,
                      "end": 1.0,
                      "probability": 1.0
                    }
                  ]
                },
                {
                  "id": 1,
                  "seek": 0,
                  "start": 1.7800000000000002,
                  "end": 4.1,
                  "text": "Windows和Mac上也是一样",
                  "tokens": [
                    50418,
                    45813,
                    1509,
                    12565,
                    34355,
                    5708,
                    22021,
                    2257,
                    14496,
                    50574
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.09749541680018108,
                  "compression_ratio": 0.8641975308641975,
                  "no_speech_prob": 0.7773045897483826,
                  "words": [
                    {
                      "word": "Wind",
                      "start": 1.7800000000000002,
                      "end": 2.18,
                      "probability": 0.93701171875
                    },
                    {
                      "word": "ows",
                      "start": 2.18,
                      "end": 2.58,
                      "probability": 0.99755859375
                    },
                    {
                      "word": "和",
                      "start": 2.58,
                      "end": 3.02,
                      "probability": 0.98779296875
                    },
                    {
                      "word": "Mac",
                      "start": 3.02,
                      "end": 3.3,
                      "probability": 0.98779296875
                    },
                    {
                      "word": "上",
                      "start": 3.3,
                      "end": 3.48,
                      "probability": 0.994140625
                    },
                    {
                      "word": "也是",
                      "start": 3.48,
                      "end": 3.7,
                      "probability": 0.99853515625
                    },
                    {
                      "word": "一",
                      "start": 3.7,
                      "end": 3.9,
                      "probability": 0.99951171875
                    },
                    {
                      "word": "样",
                      "start": 3.9,
                      "end": 4.1,
                      "probability": 1.0
                    }
                  ]
                },
                {
                  "id": 2,
                  "seek": 0,
                  "start": 4.1,
                  "end": 6.06,
                  "text": "不仅要适配各种显卡",
                  "tokens": [
                    50574,
                    1960,
                    1550,
                    227,
                    4275,
                    2215,
                    224,
                    38846,
                    17516,
                    39810,
                    1431,
                    122,
                    32681,
                    50672
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.09749541680018108,
                  "compression_ratio": 0.8641975308641975,
                  "no_speech_prob": 0.7773045897483826,
                  "words": [
                    {
                      "word": "不",
                      "start": 4.1,
                      "end": 4.76,
                      "probability": 0.98486328125
                    },
                    {
                      "word": "仅",
                      "start": 4.76,
                      "end": 4.98,
                      "probability": 0.99560546875
                    },
                    {
                      "word": "要",
                      "start": 4.98,
                      "end": 5.08,
                      "probability": 0.9990234375
                    },
                    {
                      "word": "适",
                      "start": 5.08,
                      "end": 5.32,
                      "probability": 1.0
                    },
                    {
                      "word": "配",
                      "start": 5.32,
                      "end": 5.46,
                      "probability": 0.99853515625
                    },
                    {
                      "word": "各",
                      "start": 5.46,
                      "end": 5.6,
                      "probability": 0.9990234375
                    },
                    {
                      "word": "种",
                      "start": 5.6,
                      "end": 5.8,
                      "probability": 1.0
                    },
                    {
                      "word": "显",
                      "start": 5.8,
                      "end": 6.02,
                      "probability": 0.998046875
                    },
                    {
                      "word": "卡",
                      "start": 6.02,
                      "end": 6.06,
                      "probability": 0.994140625
                    }
                  ]
                }
              ],
              "language": "zh",
              "duration": 6.14
            },
            "status": "COMPLETE",
            "requested_language": "zh"
          },
          "suppression_proposed": null
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "111",
          "before": "ows和Mac上也是一样",
          "after": "ows和Mac上也是一样",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "Name error: 'ows' vs 'Windows'",
          "context_evidence": "Context: 'Windows和Mac上也是一样'",
          "independent_ASR": {
            "engine": "whisper-large-v3",
            "model_path": "/Users/sagawa/AI_WORK/cache/huggingface/hub/models--mlx-community--whisper-large-v3",
            "language": "zh",
            "audio_interval": [
              260.96,
              267.09999999999997
            ],
            "target_interval": [
              262.96,
              265.09999999999997
            ],
            "text": "ows和Mac上也是一样",
            "words": [
              {
                "word": "ows",
                "start": 263.14,
                "end": 263.53999999999996,
                "probability": 0.99755859375
              },
              {
                "word": "和",
                "start": 263.53999999999996,
                "end": 263.97999999999996,
                "probability": 0.98779296875
              },
              {
                "word": "Mac",
                "start": 263.97999999999996,
                "end": 264.26,
                "probability": 0.98779296875
              },
              {
                "word": "上",
                "start": 264.26,
                "end": 264.44,
                "probability": 0.994140625
              },
              {
                "word": "也是",
                "start": 264.44,
                "end": 264.65999999999997,
                "probability": 0.99853515625
              },
              {
                "word": "一",
                "start": 264.65999999999997,
                "end": 264.85999999999996,
                "probability": 0.99951171875
              },
              {
                "word": "样",
                "start": 264.85999999999996,
                "end": 265.06,
                "probability": 1.0
              }
            ],
            "raw_result": {
              "text": "和负载炸弹Windows和Mac上也是一样不仅要适配各种显卡",
              "segments": [
                {
                  "id": 0,
                  "seek": 0,
                  "start": 0.0,
                  "end": 1.0,
                  "text": "和负载炸弹",
                  "tokens": [
                    50365,
                    12565,
                    18464,
                    253,
                    17819,
                    121,
                    4804,
                    116,
                    5702,
                    117,
                    50418
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.09749541680018108,
                  "compression_ratio": 0.8641975308641975,
                  "no_speech_prob": 0.7773045897483826,
                  "words": [
                    {
                      "word": "和",
                      "start": 0.0,
                      "end": 0.38,
                      "probability": 0.72021484375
                    },
                    {
                      "word": "负",
                      "start": 0.38,
                      "end": 0.54,
                      "probability": 0.84619140625
                    },
                    {
                      "word": "载",
                      "start": 0.54,
                      "end": 0.7,
                      "probability": 0.998046875
                    },
                    {
                      "word": "炸",
                      "start": 0.7,
                      "end": 0.86,
                      "probability": 0.99755859375
                    },
                    {
                      "word": "弹",
                      "start": 0.86,
                      "end": 1.0,
                      "probability": 1.0
                    }
                  ]
                },
                {
                  "id": 1,
                  "seek": 0,
                  "start": 1.7800000000000002,
                  "end": 4.1,
                  "text": "Windows和Mac上也是一样",
                  "tokens": [
                    50418,
                    45813,
                    1509,
                    12565,
                    34355,
                    5708,
                    22021,
                    2257,
                    14496,
                    50574
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.09749541680018108,
                  "compression_ratio": 0.8641975308641975,
                  "no_speech_prob": 0.7773045897483826,
                  "words": [
                    {
                      "word": "Wind",
                      "start": 1.7800000000000002,
                      "end": 2.18,
                      "probability": 0.93701171875
                    },
                    {
                      "word": "ows",
                      "start": 2.18,
                      "end": 2.58,
                      "probability": 0.99755859375
                    },
                    {
                      "word": "和",
                      "start": 2.58,
                      "end": 3.02,
                      "probability": 0.98779296875
                    },
                    {
                      "word": "Mac",
                      "start": 3.02,
                      "end": 3.3,
                      "probability": 0.98779296875
                    },
                    {
                      "word": "上",
                      "start": 3.3,
                      "end": 3.48,
                      "probability": 0.994140625
                    },
                    {
                      "word": "也是",
                      "start": 3.48,
                      "end": 3.7,
                      "probability": 0.99853515625
                    },
                    {
                      "word": "一",
                      "start": 3.7,
                      "end": 3.9,
                      "probability": 0.99951171875
                    },
                    {
                      "word": "样",
                      "start": 3.9,
                      "end": 4.1,
                      "probability": 1.0
                    }
                  ]
                },
                {
                  "id": 2,
                  "seek": 0,
                  "start": 4.1,
                  "end": 6.06,
                  "text": "不仅要适配各种显卡",
                  "tokens": [
                    50574,
                    1960,
                    1550,
                    227,
                    4275,
                    2215,
                    224,
                    38846,
                    17516,
                    39810,
                    1431,
                    122,
                    32681,
                    50672
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.09749541680018108,
                  "compression_ratio": 0.8641975308641975,
                  "no_speech_prob": 0.7773045897483826,
                  "words": [
                    {
                      "word": "不",
                      "start": 4.1,
                      "end": 4.76,
                      "probability": 0.98486328125
                    },
                    {
                      "word": "仅",
                      "start": 4.76,
                      "end": 4.98,
                      "probability": 0.99560546875
                    },
                    {
                      "word": "要",
                      "start": 4.98,
                      "end": 5.08,
                      "probability": 0.9990234375
                    },
                    {
                      "word": "适",
                      "start": 5.08,
                      "end": 5.32,
                      "probability": 1.0
                    },
                    {
                      "word": "配",
                      "start": 5.32,
                      "end": 5.46,
                      "probability": 0.99853515625
                    },
                    {
                      "word": "各",
                      "start": 5.46,
                      "end": 5.6,
                      "probability": 0.9990234375
                    },
                    {
                      "word": "种",
                      "start": 5.6,
                      "end": 5.8,
                      "probability": 1.0
                    },
                    {
                      "word": "显",
                      "start": 5.8,
                      "end": 6.02,
                      "probability": 0.998046875
                    },
                    {
                      "word": "卡",
                      "start": 6.02,
                      "end": 6.06,
                      "probability": 0.994140625
                    }
                  ]
                }
              ],
              "language": "zh",
              "duration": 6.14
            },
            "status": "COMPLETE",
            "requested_language": "zh"
          },
          "suppression_proposed": null
        }
      ],
      "language": "zh",
      "packet_ids": [
        10
      ],
      "classification": "BEST_SUPPORTED_WITH_REVIEW",
      "independent_ASR": {
        "engine": "whisper-large-v3",
        "model_path": "/Users/sagawa/AI_WORK/cache/huggingface/hub/models--mlx-community--whisper-large-v3",
        "language": "zh",
        "audio_interval": [
          260.96,
          267.09999999999997
        ],
        "target_interval": [
          262.96,
          265.09999999999997
        ],
        "text": "ows和Mac上也是一样",
        "words": [
          {
            "word": "ows",
            "start": 263.14,
            "end": 263.53999999999996,
            "probability": 0.99755859375
          },
          {
            "word": "和",
            "start": 263.53999999999996,
            "end": 263.97999999999996,
            "probability": 0.98779296875
          },
          {
            "word": "Mac",
            "start": 263.97999999999996,
            "end": 264.26,
            "probability": 0.98779296875
          },
          {
            "word": "上",
            "start": 264.26,
            "end": 264.44,
            "probability": 0.994140625
          },
          {
            "word": "也是",
            "start": 264.44,
            "end": 264.65999999999997,
            "probability": 0.99853515625
          },
          {
            "word": "一",
            "start": 264.65999999999997,
            "end": 264.85999999999996,
            "probability": 0.99951171875
          },
          {
            "word": "样",
            "start": 264.85999999999996,
            "end": 265.06,
            "probability": 1.0
          }
        ],
        "raw_result": {
          "text": "和负载炸弹Windows和Mac上也是一样不仅要适配各种显卡",
          "segments": [
            {
              "id": 0,
              "seek": 0,
              "start": 0.0,
              "end": 1.0,
              "text": "和负载炸弹",
              "tokens": [
                50365,
                12565,
                18464,
                253,
                17819,
                121,
                4804,
                116,
                5702,
                117,
                50418
              ],
              "temperature": 0.0,
              "avg_logprob": -0.09749541680018108,
              "compression_ratio": 0.8641975308641975,
              "no_speech_prob": 0.7773045897483826,
              "words": [
                {
                  "word": "和",
                  "start": 0.0,
                  "end": 0.38,
                  "probability": 0.72021484375
                },
                {
                  "word": "负",
                  "start": 0.38,
                  "end": 0.54,
                  "probability": 0.84619140625
                },
                {
                  "word": "载",
                  "start": 0.54,
                  "end": 0.7,
                  "probability": 0.998046875
                },
                {
                  "word": "炸",
                  "start": 0.7,
                  "end": 0.86,
                  "probability": 0.99755859375
                },
                {
                  "word": "弹",
                  "start": 0.86,
                  "end": 1.0,
                  "probability": 1.0
                }
              ]
            },
            {
              "id": 1,
              "seek": 0,
              "start": 1.7800000000000002,
              "end": 4.1,
              "text": "Windows和Mac上也是一样",
              "tokens": [
                50418,
                45813,
                1509,
                12565,
                34355,
                5708,
                22021,
                2257,
                14496,
                50574
              ],
              "temperature": 0.0,
              "avg_logprob": -0.09749541680018108,
              "compression_ratio": 0.8641975308641975,
              "no_speech_prob": 0.7773045897483826,
              "words": [
                {
                  "word": "Wind",
                  "start": 1.7800000000000002,
                  "end": 2.18,
                  "probability": 0.93701171875
                },
                {
                  "word": "ows",
                  "start": 2.18,
                  "end": 2.58,
                  "probability": 0.99755859375
                },
                {
                  "word": "和",
                  "start": 2.58,
                  "end": 3.02,
                  "probability": 0.98779296875
                },
                {
                  "word": "Mac",
                  "start": 3.02,
                  "end": 3.3,
                  "probability": 0.98779296875
                },
                {
                  "word": "上",
                  "start": 3.3,
                  "end": 3.48,
                  "probability": 0.994140625
                },
                {
                  "word": "也是",
                  "start": 3.48,
                  "end": 3.7,
                  "probability": 0.99853515625
                },
                {
                  "word": "一",
                  "start": 3.7,
                  "end": 3.9,
                  "probability": 0.99951171875
                },
                {
                  "word": "样",
                  "start": 3.9,
                  "end": 4.1,
                  "probability": 1.0
                }
              ]
            },
            {
              "id": 2,
              "seek": 0,
              "start": 4.1,
              "end": 6.06,
              "text": "不仅要适配各种显卡",
              "tokens": [
                50574,
                1960,
                1550,
                227,
                4275,
                2215,
                224,
                38846,
                17516,
                39810,
                1431,
                122,
                32681,
                50672
              ],
              "temperature": 0.0,
              "avg_logprob": -0.09749541680018108,
              "compression_ratio": 0.8641975308641975,
              "no_speech_prob": 0.7773045897483826,
              "words": [
                {
                  "word": "不",
                  "start": 4.1,
                  "end": 4.76,
                  "probability": 0.98486328125
                },
                {
                  "word": "仅",
                  "start": 4.76,
                  "end": 4.98,
                  "probability": 0.99560546875
                },
                {
                  "word": "要",
                  "start": 4.98,
                  "end": 5.08,
                  "probability": 0.9990234375
                },
                {
                  "word": "适",
                  "start": 5.08,
                  "end": 5.32,
                  "probability": 1.0
                },
                {
                  "word": "配",
                  "start": 5.32,
                  "end": 5.46,
                  "probability": 0.99853515625
                },
                {
                  "word": "各",
                  "start": 5.46,
                  "end": 5.6,
                  "probability": 0.9990234375
                },
                {
                  "word": "种",
                  "start": 5.6,
                  "end": 5.8,
                  "probability": 1.0
                },
                {
                  "word": "显",
                  "start": 5.8,
                  "end": 6.02,
                  "probability": 0.998046875
                },
                {
                  "word": "卡",
                  "start": 6.02,
                  "end": 6.06,
                  "probability": 0.994140625
                }
              ]
            }
          ],
          "language": "zh",
          "duration": 6.14
        },
        "status": "COMPLETE",
        "requested_language": "zh"
      },
      "raw_word_timestamps": [
        {
          "word": "Wind",
          "start": 262.96,
          "end": 263.32,
          "probability": 0.8955078125
        },
        {
          "word": "ows",
          "start": 263.32,
          "end": 263.56,
          "probability": 0.998046875
        },
        {
          "word": "和",
          "start": 263.56,
          "end": 264.09999999999997,
          "probability": 0.9912109375
        },
        {
          "word": "Mac",
          "start": 264.09999999999997,
          "end": 264.34,
          "probability": 0.9951171875
        },
        {
          "word": "上",
          "start": 264.34,
          "end": 264.52,
          "probability": 0.982421875
        },
        {
          "word": "也是",
          "start": 264.52,
          "end": 264.78,
          "probability": 0.99951171875
        },
        {
          "word": "一",
          "start": 264.78,
          "end": 264.96,
          "probability": 0.99609375
        },
        {
          "word": "样",
          "start": 264.96,
          "end": 265.09999999999997,
          "probability": 1.0
        }
      ],
      "word_timestamps": [],
      "source_text": "ows和Mac上也是一样",
      "review_status": "BEST_SUPPORTED_WITH_REVIEW",
      "canonical_timing_evidence": "CUE_ANCHORED",
      "index": 111
    },
    {
      "start": 265.47999999999996,
      "end": 267.96,
      "text": "不仅要适配各种显卡和统一内存",
      "raw_asr_text": "不仅要适配各种显卡和统一内存",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 10,
      "words": [
        {
          "word": "不",
          "start": 265.47999999999996,
          "end": 265.84,
          "probability": 0.8955078125
        },
        {
          "word": "仅",
          "start": 265.84,
          "end": 265.96,
          "probability": 0.99951171875
        },
        {
          "word": "要",
          "start": 265.96,
          "end": 266.12,
          "probability": 0.99951171875
        },
        {
          "word": "适",
          "start": 266.12,
          "end": 266.29999999999995,
          "probability": 0.998046875
        },
        {
          "word": "配",
          "start": 266.29999999999995,
          "end": 266.41999999999996,
          "probability": 0.99365234375
        },
        {
          "word": "各",
          "start": 266.41999999999996,
          "end": 266.62,
          "probability": 1.0
        },
        {
          "word": "种",
          "start": 266.62,
          "end": 266.79999999999995,
          "probability": 1.0
        },
        {
          "word": "显",
          "start": 266.79999999999995,
          "end": 266.97999999999996,
          "probability": 0.9970703125
        },
        {
          "word": "卡",
          "start": 266.97999999999996,
          "end": 267.14,
          "probability": 0.99560546875
        },
        {
          "word": "和",
          "start": 267.14,
          "end": 267.28,
          "probability": 0.99951171875
        },
        {
          "word": "统",
          "start": 267.28,
          "end": 267.44,
          "probability": 0.9990234375
        },
        {
          "word": "一",
          "start": 267.44,
          "end": 267.58,
          "probability": 0.99951171875
        },
        {
          "word": "内",
          "start": 267.58,
          "end": 267.71999999999997,
          "probability": 0.99658203125
        },
        {
          "word": "存",
          "start": 267.71999999999997,
          "end": 267.96,
          "probability": 1.0
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.09083681895320576,
        "compression_ratio": 1.189922480620155,
        "no_speech_prob": 2.283331683450296e-11
      },
      "lid_status": "UNCERTAIN",
      "alternatives": [
        {
          "language": "zh",
          "text": "Windows和Mac上也是一样不仅要适配各种显卡和统一内存而且上了桌面端大家肯定会想要更强的模型而不只是两臂的量化小模型继续支持这些会给项目引入大量非必要的开发成本如果你想在本地部署模型其实有很多更优秀的专门项目用它"
        }
      ],
      "id": 112,
      "review_annotation": "",
      "row_id": "112",
      "bounded_row_id": 112,
      "parent_row_ids": [
        112
      ],
      "parent_row_id": 112,
      "packet_id": 10,
      "LID_evidence": [
        {
          "region_id": 52,
          "probabilities": {
            "zh": 0.9982388748062981
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        52
      ],
      "language_confidence": 0.9982388748062981,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 261.9,
      "packet_end": 288.40000000000003,
      "raw_text": "不仅要适配各种显卡和统一内存",
      "selected_text": "不仅要适配各种显卡和统一内存",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            112
          ],
          "packet_id": 10,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        10
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 112
    },
    {
      "start": 268.47999999999996,
      "end": 269.7,
      "text": "而且上了桌面端",
      "raw_asr_text": "而且上了桌面端",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 10,
      "words": [
        {
          "word": "而且",
          "start": 268.47999999999996,
          "end": 268.84,
          "probability": 0.99560546875
        },
        {
          "word": "上",
          "start": 268.84,
          "end": 269.06,
          "probability": 0.9990234375
        },
        {
          "word": "了",
          "start": 269.06,
          "end": 269.15999999999997,
          "probability": 0.99951171875
        },
        {
          "word": "桌",
          "start": 269.15999999999997,
          "end": 269.32,
          "probability": 0.99951171875
        },
        {
          "word": "面",
          "start": 269.32,
          "end": 269.5,
          "probability": 1.0
        },
        {
          "word": "端",
          "start": 269.5,
          "end": 269.7,
          "probability": 0.97705078125
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.09083681895320576,
        "compression_ratio": 1.189922480620155,
        "no_speech_prob": 2.283331683450296e-11
      },
      "lid_status": "UNCERTAIN",
      "alternatives": [
        {
          "language": "zh",
          "text": "Windows和Mac上也是一样不仅要适配各种显卡和统一内存而且上了桌面端大家肯定会想要更强的模型而不只是两臂的量化小模型继续支持这些会给项目引入大量非必要的开发成本如果你想在本地部署模型其实有很多更优秀的专门项目用它"
        }
      ],
      "id": 113,
      "review_annotation": "",
      "row_id": "113",
      "bounded_row_id": 113,
      "parent_row_ids": [
        113
      ],
      "parent_row_id": 113,
      "packet_id": 10,
      "LID_evidence": [
        {
          "region_id": 53,
          "probabilities": {
            "zh": 0.9978324224551519
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        53
      ],
      "language_confidence": 0.9978324224551519,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 261.9,
      "packet_end": 288.40000000000003,
      "raw_text": "而且上了桌面端",
      "selected_text": "而且上了桌面端",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            113
          ],
          "packet_id": 10,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        10
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 113
    },
    {
      "start": 269.7,
      "end": 271.91999999999996,
      "text": "大家肯定会想要更强的模型",
      "raw_asr_text": "大家肯定会想要更强的模型",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 10,
      "words": [
        {
          "word": "大家",
          "start": 269.7,
          "end": 270.2,
          "probability": 0.984375
        },
        {
          "word": "肯",
          "start": 270.2,
          "end": 270.5,
          "probability": 1.0
        },
        {
          "word": "定",
          "start": 270.5,
          "end": 270.59999999999997,
          "probability": 1.0
        },
        {
          "word": "会",
          "start": 270.59999999999997,
          "end": 270.76,
          "probability": 1.0
        },
        {
          "word": "想",
          "start": 270.76,
          "end": 270.91999999999996,
          "probability": 0.9990234375
        },
        {
          "word": "要",
          "start": 270.91999999999996,
          "end": 271.03999999999996,
          "probability": 0.99951171875
        },
        {
          "word": "更",
          "start": 271.03999999999996,
          "end": 271.21999999999997,
          "probability": 1.0
        },
        {
          "word": "强",
          "start": 271.21999999999997,
          "end": 271.41999999999996,
          "probability": 1.0
        },
        {
          "word": "的",
          "start": 271.41999999999996,
          "end": 271.52,
          "probability": 1.0
        },
        {
          "word": "模",
          "start": 271.52,
          "end": 271.65999999999997,
          "probability": 0.99853515625
        },
        {
          "word": "型",
          "start": 271.65999999999997,
          "end": 271.91999999999996,
          "probability": 0.99951171875
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.09083681895320576,
        "compression_ratio": 1.189922480620155,
        "no_speech_prob": 2.283331683450296e-11
      },
      "lid_status": "UNCERTAIN",
      "alternatives": [
        {
          "language": "zh",
          "text": "Windows和Mac上也是一样不仅要适配各种显卡和统一内存而且上了桌面端大家肯定会想要更强的模型而不只是两臂的量化小模型继续支持这些会给项目引入大量非必要的开发成本如果你想在本地部署模型其实有很多更优秀的专门项目用它"
        }
      ],
      "id": 114,
      "review_annotation": "",
      "row_id": "114",
      "bounded_row_id": 114,
      "parent_row_ids": [
        114
      ],
      "parent_row_id": 114,
      "packet_id": 10,
      "LID_evidence": [
        {
          "region_id": 53,
          "probabilities": {
            "zh": 0.9978324224551519
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        53
      ],
      "language_confidence": 0.9978324224551519,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 261.9,
      "packet_end": 288.40000000000003,
      "raw_text": "大家肯定会想要更强的模型",
      "selected_text": "大家肯定会想要更强的模型",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            114
          ],
          "packet_id": 10,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        10
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 114
    },
    {
      "start": 272.38,
      "end": 274.91999999999996,
      "text": "两臂",
      "raw_asr_text": "而不只是两臂的量化小模型",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 10,
      "words": [],
      "asr_metrics": {
        "avg_logprob": -0.09083681895320576,
        "compression_ratio": 1.189922480620155,
        "no_speech_prob": 2.283331683450296e-11
      },
      "lid_status": "UNCERTAIN",
      "alternatives": [
        {
          "language": "zh",
          "text": "Windows和Mac上也是一样不仅要适配各种显卡和统一内存而且上了桌面端大家肯定会想要更强的模型而不只是两臂的量化小模型继续支持这些会给项目引入大量非必要的开发成本如果你想在本地部署模型其实有很多更优秀的专门项目用它"
        }
      ],
      "id": 115,
      "review_annotation": "待確認",
      "row_id": "115",
      "bounded_row_id": 115,
      "parent_row_ids": [
        115
      ],
      "parent_row_id": 115,
      "packet_id": 10,
      "LID_evidence": [
        {
          "region_id": 54,
          "probabilities": {
            "zh": 0.9968667891290452
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        54
      ],
      "language_confidence": 0.9968667891290452,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 261.9,
      "packet_end": 288.40000000000003,
      "raw_text": "而不只是两臂的量化小模型",
      "selected_text": "两臂",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            115
          ],
          "packet_id": 10,
          "engine": "mlx-whisper"
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "115",
          "before": "而不只是两臂的量化小模型",
          "after": "两臂",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "两臂疑似轻量或量化误识",
          "context_evidence": "上下文为模型大小，'轻量'或'量化'更符合",
          "independent_ASR": {
            "status": "REVIEW_ONLY_RESCUE_BUDGET"
          },
          "suppression_proposed": null
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "115",
          "before": "两臂",
          "after": "两臂",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "Nonsense: '两臂' vs '还要' (also need)",
          "context_evidence": "Context: '还要继续支持这些'",
          "independent_ASR": {
            "status": "REVIEW_ONLY_RESCUE_BUDGET"
          },
          "suppression_proposed": null
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "115",
          "before": "两臂",
          "after": "两臂",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "Nonsense: '两臂' vs '还要' (also need)",
          "context_evidence": "Context: '还要继续支持这些'",
          "independent_ASR": {
            "status": "REVIEW_ONLY_RESCUE_BUDGET"
          },
          "suppression_proposed": null
        }
      ],
      "language": "zh",
      "packet_ids": [
        10
      ],
      "classification": "BEST_SUPPORTED_WITH_REVIEW",
      "independent_ASR": {
        "status": "REVIEW_ONLY_RESCUE_BUDGET"
      },
      "raw_word_timestamps": [
        {
          "word": "而",
          "start": 272.38,
          "end": 272.73999999999995,
          "probability": 0.99853515625
        },
        {
          "word": "不",
          "start": 272.73999999999995,
          "end": 272.9,
          "probability": 1.0
        },
        {
          "word": "只是",
          "start": 272.9,
          "end": 273.15999999999997,
          "probability": 0.9716796875
        },
        {
          "word": "两",
          "start": 273.15999999999997,
          "end": 273.64,
          "probability": 0.62939453125
        },
        {
          "word": "臂",
          "start": 273.64,
          "end": 273.84,
          "probability": 0.7265625
        },
        {
          "word": "的",
          "start": 273.84,
          "end": 273.97999999999996,
          "probability": 0.998046875
        },
        {
          "word": "量",
          "start": 273.97999999999996,
          "end": 274.12,
          "probability": 0.9736328125
        },
        {
          "word": "化",
          "start": 274.12,
          "end": 274.28,
          "probability": 0.99951171875
        },
        {
          "word": "小",
          "start": 274.28,
          "end": 274.47999999999996,
          "probability": 0.99609375
        },
        {
          "word": "模",
          "start": 274.47999999999996,
          "end": 274.65999999999997,
          "probability": 1.0
        },
        {
          "word": "型",
          "start": 274.65999999999997,
          "end": 274.91999999999996,
          "probability": 1.0
        }
      ],
      "word_timestamps": [],
      "source_text": "两臂",
      "review_status": "BEST_SUPPORTED_WITH_REVIEW",
      "canonical_timing_evidence": "CUE_ANCHORED",
      "index": 115
    },
    {
      "start": 275.65999999999997,
      "end": 276.88,
      "text": "继续支持这些",
      "raw_asr_text": "继续支持这些",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 10,
      "words": [
        {
          "word": "继",
          "start": 275.65999999999997,
          "end": 276.02,
          "probability": 0.9990234375
        },
        {
          "word": "续",
          "start": 276.02,
          "end": 276.17999999999995,
          "probability": 1.0
        },
        {
          "word": "支持",
          "start": 276.17999999999995,
          "end": 276.4,
          "probability": 0.99951171875
        },
        {
          "word": "这",
          "start": 276.4,
          "end": 276.62,
          "probability": 0.99951171875
        },
        {
          "word": "些",
          "start": 276.62,
          "end": 276.88,
          "probability": 1.0
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.09083681895320576,
        "compression_ratio": 1.189922480620155,
        "no_speech_prob": 2.283331683450296e-11
      },
      "lid_status": "UNCERTAIN",
      "alternatives": [
        {
          "language": "zh",
          "text": "Windows和Mac上也是一样不仅要适配各种显卡和统一内存而且上了桌面端大家肯定会想要更强的模型而不只是两臂的量化小模型继续支持这些会给项目引入大量非必要的开发成本如果你想在本地部署模型其实有很多更优秀的专门项目用它"
        }
      ],
      "id": 116,
      "review_annotation": "",
      "row_id": "116",
      "bounded_row_id": 116,
      "parent_row_ids": [
        116
      ],
      "parent_row_id": 116,
      "packet_id": 10,
      "LID_evidence": [
        {
          "region_id": 55,
          "probabilities": {
            "zh": 0.9972164869308472
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        55
      ],
      "language_confidence": 0.9972164869308472,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 261.9,
      "packet_end": 288.40000000000003,
      "raw_text": "继续支持这些",
      "selected_text": "继续支持这些",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            116
          ],
          "packet_id": 10,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        10
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 116
    },
    {
      "start": 277.35999999999996,
      "end": 280.78,
      "text": "会给项目引入大量非必要的开发成本",
      "raw_asr_text": "会给项目引入大量非必要的开发成本",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 10,
      "words": [
        {
          "word": "会",
          "start": 277.35999999999996,
          "end": 277.71999999999997,
          "probability": 0.99755859375
        },
        {
          "word": "给",
          "start": 277.71999999999997,
          "end": 277.84,
          "probability": 1.0
        },
        {
          "word": "项",
          "start": 277.84,
          "end": 278.08,
          "probability": 0.99951171875
        },
        {
          "word": "目",
          "start": 278.08,
          "end": 278.32,
          "probability": 1.0
        },
        {
          "word": "引",
          "start": 278.32,
          "end": 278.82,
          "probability": 0.99462890625
        },
        {
          "word": "入",
          "start": 278.82,
          "end": 278.97999999999996,
          "probability": 0.99951171875
        },
        {
          "word": "大",
          "start": 278.97999999999996,
          "end": 279.17999999999995,
          "probability": 0.99951171875
        },
        {
          "word": "量",
          "start": 279.17999999999995,
          "end": 279.38,
          "probability": 1.0
        },
        {
          "word": "非",
          "start": 279.38,
          "end": 279.62,
          "probability": 0.99658203125
        },
        {
          "word": "必",
          "start": 279.62,
          "end": 279.82,
          "probability": 0.99951171875
        },
        {
          "word": "要",
          "start": 279.82,
          "end": 279.97999999999996,
          "probability": 1.0
        },
        {
          "word": "的",
          "start": 279.97999999999996,
          "end": 280.08,
          "probability": 1.0
        },
        {
          "word": "开",
          "start": 280.08,
          "end": 280.21999999999997,
          "probability": 0.99853515625
        },
        {
          "word": "发",
          "start": 280.21999999999997,
          "end": 280.38,
          "probability": 0.994140625
        },
        {
          "word": "成",
          "start": 280.38,
          "end": 280.56,
          "probability": 1.0
        },
        {
          "word": "本",
          "start": 280.56,
          "end": 280.78,
          "probability": 0.99951171875
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.09083681895320576,
        "compression_ratio": 1.189922480620155,
        "no_speech_prob": 2.283331683450296e-11
      },
      "lid_status": "UNCERTAIN",
      "alternatives": [
        {
          "language": "zh",
          "text": "Windows和Mac上也是一样不仅要适配各种显卡和统一内存而且上了桌面端大家肯定会想要更强的模型而不只是两臂的量化小模型继续支持这些会给项目引入大量非必要的开发成本如果你想在本地部署模型其实有很多更优秀的专门项目用它"
        }
      ],
      "id": 117,
      "review_annotation": "",
      "row_id": "117",
      "bounded_row_id": 117,
      "parent_row_ids": [
        117
      ],
      "parent_row_id": 117,
      "packet_id": 10,
      "LID_evidence": [
        {
          "region_id": 55,
          "probabilities": {
            "zh": 0.9972164869308472
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        55
      ],
      "language_confidence": 0.9972164869308472,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 261.9,
      "packet_end": 288.40000000000003,
      "raw_text": "会给项目引入大量非必要的开发成本",
      "selected_text": "会给项目引入大量非必要的开发成本",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            117
          ],
          "packet_id": 10,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        10
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 117
    },
    {
      "start": 281.73999999999995,
      "end": 283.76,
      "text": "如果你想在本地部署模型",
      "raw_asr_text": "如果你想在本地部署模型",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 10,
      "words": [
        {
          "word": "如果",
          "start": 281.73999999999995,
          "end": 282.09999999999997,
          "probability": 0.99658203125
        },
        {
          "word": "你想",
          "start": 282.09999999999997,
          "end": 282.46,
          "probability": 0.99365234375
        },
        {
          "word": "在",
          "start": 282.46,
          "end": 282.64,
          "probability": 1.0
        },
        {
          "word": "本",
          "start": 282.64,
          "end": 282.82,
          "probability": 1.0
        },
        {
          "word": "地",
          "start": 282.82,
          "end": 283.0,
          "probability": 1.0
        },
        {
          "word": "部",
          "start": 283.0,
          "end": 283.17999999999995,
          "probability": 0.9599609375
        },
        {
          "word": "署",
          "start": 283.17999999999995,
          "end": 283.32,
          "probability": 0.9921875
        },
        {
          "word": "模",
          "start": 283.32,
          "end": 283.47999999999996,
          "probability": 0.99951171875
        },
        {
          "word": "型",
          "start": 283.47999999999996,
          "end": 283.76,
          "probability": 1.0
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.09083681895320576,
        "compression_ratio": 1.189922480620155,
        "no_speech_prob": 2.283331683450296e-11
      },
      "lid_status": "UNCERTAIN",
      "alternatives": [
        {
          "language": "zh",
          "text": "Windows和Mac上也是一样不仅要适配各种显卡和统一内存而且上了桌面端大家肯定会想要更强的模型而不只是两臂的量化小模型继续支持这些会给项目引入大量非必要的开发成本如果你想在本地部署模型其实有很多更优秀的专门项目用它"
        }
      ],
      "id": 118,
      "review_annotation": "",
      "row_id": "118",
      "bounded_row_id": 118,
      "parent_row_ids": [
        118
      ],
      "parent_row_id": 118,
      "packet_id": 10,
      "LID_evidence": [
        {
          "region_id": 56,
          "probabilities": {
            "zh": 0.9978239072693719
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        56
      ],
      "language_confidence": 0.9978239072693719,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 261.9,
      "packet_end": 288.40000000000003,
      "raw_text": "如果你想在本地部署模型",
      "selected_text": "如果你想在本地部署模型",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            118
          ],
          "packet_id": 10,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        10
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 118
    },
    {
      "start": 284.78,
      "end": 287.29999999999995,
      "text": "其实有很多更优秀的专门项目",
      "raw_asr_text": "其实有很多更优秀的专门项目",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 10,
      "words": [
        {
          "word": "其实",
          "start": 284.78,
          "end": 285.14,
          "probability": 0.998046875
        },
        {
          "word": "有",
          "start": 285.14,
          "end": 285.5,
          "probability": 0.99853515625
        },
        {
          "word": "很多",
          "start": 285.5,
          "end": 285.91999999999996,
          "probability": 1.0
        },
        {
          "word": "更",
          "start": 285.91999999999996,
          "end": 286.21999999999997,
          "probability": 0.99755859375
        },
        {
          "word": "优",
          "start": 286.21999999999997,
          "end": 286.4,
          "probability": 0.99951171875
        },
        {
          "word": "秀",
          "start": 286.4,
          "end": 286.53999999999996,
          "probability": 1.0
        },
        {
          "word": "的",
          "start": 286.53999999999996,
          "end": 286.64,
          "probability": 1.0
        },
        {
          "word": "专",
          "start": 286.64,
          "end": 286.78,
          "probability": 1.0
        },
        {
          "word": "门",
          "start": 286.78,
          "end": 286.9,
          "probability": 0.9990234375
        },
        {
          "word": "项",
          "start": 286.9,
          "end": 287.08,
          "probability": 1.0
        },
        {
          "word": "目",
          "start": 287.08,
          "end": 287.29999999999995,
          "probability": 1.0
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.09083681895320576,
        "compression_ratio": 1.189922480620155,
        "no_speech_prob": 2.283331683450296e-11
      },
      "lid_status": "UNCERTAIN",
      "alternatives": [
        {
          "language": "zh",
          "text": "Windows和Mac上也是一样不仅要适配各种显卡和统一内存而且上了桌面端大家肯定会想要更强的模型而不只是两臂的量化小模型继续支持这些会给项目引入大量非必要的开发成本如果你想在本地部署模型其实有很多更优秀的专门项目用它"
        }
      ],
      "id": 119,
      "review_annotation": "",
      "row_id": "119",
      "bounded_row_id": 119,
      "parent_row_ids": [
        119
      ],
      "parent_row_id": 119,
      "packet_id": 10,
      "LID_evidence": [
        {
          "region_id": 57,
          "probabilities": {
            "zh": 0.997859126329422
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        57
      ],
      "language_confidence": 0.997859126329422,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 261.9,
      "packet_end": 288.40000000000003,
      "raw_text": "其实有很多更优秀的专门项目",
      "selected_text": "其实有很多更优秀的专门项目",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            119
          ],
          "packet_id": 10,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        10
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 119
    },
    {
      "start": 287.9,
      "end": 289.59999999999997,
      "text": "用它们在电脑上部署好",
      "raw_asr_text": "用它们在电脑上部署好",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 11,
      "words": [
        {
          "word": "用",
          "start": 287.9,
          "end": 288.21999999999997,
          "probability": 0.4853515625
        },
        {
          "word": "它",
          "start": 288.21999999999997,
          "end": 288.37999999999994,
          "probability": 0.8076171875
        },
        {
          "word": "们",
          "start": 288.37999999999994,
          "end": 288.49999999999994,
          "probability": 0.98193359375
        },
        {
          "word": "在",
          "start": 288.49999999999994,
          "end": 288.61999999999995,
          "probability": 0.9990234375
        },
        {
          "word": "电",
          "start": 288.61999999999995,
          "end": 288.78,
          "probability": 0.9970703125
        },
        {
          "word": "脑",
          "start": 288.78,
          "end": 288.93999999999994,
          "probability": 1.0
        },
        {
          "word": "上",
          "start": 288.93999999999994,
          "end": 289.08,
          "probability": 0.998046875
        },
        {
          "word": "部",
          "start": 289.08,
          "end": 289.23999999999995,
          "probability": 0.94091796875
        },
        {
          "word": "署",
          "start": 289.23999999999995,
          "end": 289.4,
          "probability": 0.9951171875
        },
        {
          "word": "好",
          "start": 289.4,
          "end": 289.59999999999997,
          "probability": 0.99853515625
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.09536371839807388,
        "compression_ratio": 1.224,
        "no_speech_prob": 7.165966431355031e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "用它们在电脑上部署好再通过本地API接入同传字幕专业的事交给专业的工具才更符合这个项目的定位这还带来了一个附带的好处以前为了照顾国内网络环境我做了很多模型下载失配现在只剩语音识别模型干脆直接内置进安装包把下载"
        }
      ],
      "id": 120,
      "review_annotation": "",
      "row_id": "120",
      "bounded_row_id": 120,
      "parent_row_ids": [
        120
      ],
      "parent_row_id": 120,
      "packet_id": 11,
      "LID_evidence": [
        {
          "region_id": 58,
          "probabilities": {
            "zh": 0.9979304944475492
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        58
      ],
      "language_confidence": 0.9979304944475492,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 287.09999999999997,
      "packet_end": 310.5,
      "raw_text": "用它们在电脑上部署好",
      "selected_text": "用它们在电脑上部署好",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            120
          ],
          "packet_id": 11,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        11
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 120
    },
    {
      "start": 289.59999999999997,
      "end": 292.29999999999995,
      "text": "再通过本地API接入同传字幕",
      "raw_asr_text": "再通过本地API接入同传字幕",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 11,
      "words": [
        {
          "word": "再",
          "start": 289.59999999999997,
          "end": 290.21999999999997,
          "probability": 0.8310546875
        },
        {
          "word": "通",
          "start": 290.21999999999997,
          "end": 290.37999999999994,
          "probability": 0.99755859375
        },
        {
          "word": "过",
          "start": 290.37999999999994,
          "end": 290.53999999999996,
          "probability": 0.9990234375
        },
        {
          "word": "本",
          "start": 290.53999999999996,
          "end": 290.7,
          "probability": 0.9970703125
        },
        {
          "word": "地",
          "start": 290.7,
          "end": 290.85999999999996,
          "probability": 0.99951171875
        },
        {
          "word": "AP",
          "start": 290.85999999999996,
          "end": 291.09999999999997,
          "probability": 0.94287109375
        },
        {
          "word": "I",
          "start": 291.09999999999997,
          "end": 291.32,
          "probability": 0.99951171875
        },
        {
          "word": "接",
          "start": 291.32,
          "end": 291.46,
          "probability": 0.8662109375
        },
        {
          "word": "入",
          "start": 291.46,
          "end": 291.61999999999995,
          "probability": 0.994140625
        },
        {
          "word": "同",
          "start": 291.61999999999995,
          "end": 291.78,
          "probability": 0.96044921875
        },
        {
          "word": "传",
          "start": 291.78,
          "end": 291.97999999999996,
          "probability": 0.99462890625
        },
        {
          "word": "字",
          "start": 291.97999999999996,
          "end": 292.11999999999995,
          "probability": 0.99658203125
        },
        {
          "word": "幕",
          "start": 292.11999999999995,
          "end": 292.29999999999995,
          "probability": 0.98681640625
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.09536371839807388,
        "compression_ratio": 1.224,
        "no_speech_prob": 7.165966431355031e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "用它们在电脑上部署好再通过本地API接入同传字幕专业的事交给专业的工具才更符合这个项目的定位这还带来了一个附带的好处以前为了照顾国内网络环境我做了很多模型下载失配现在只剩语音识别模型干脆直接内置进安装包把下载"
        }
      ],
      "id": 121,
      "review_annotation": "",
      "row_id": "121",
      "bounded_row_id": 121,
      "parent_row_ids": [
        121
      ],
      "parent_row_id": 121,
      "packet_id": 11,
      "LID_evidence": [
        {
          "region_id": 58,
          "probabilities": {
            "zh": 0.9979304944475492
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        58
      ],
      "language_confidence": 0.9979304944475492,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 287.09999999999997,
      "packet_end": 310.5,
      "raw_text": "再通过本地API接入同传字幕",
      "selected_text": "再通过本地API接入同传字幕",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            121
          ],
          "packet_id": 11,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        11
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 121
    },
    {
      "start": 292.29999999999995,
      "end": 294.58,
      "text": "专业的事交给专业的工具",
      "raw_asr_text": "专业的事交给专业的工具",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 11,
      "words": [
        {
          "word": "专",
          "start": 292.29999999999995,
          "end": 292.91999999999996,
          "probability": 0.990234375
        },
        {
          "word": "业",
          "start": 292.91999999999996,
          "end": 293.05999999999995,
          "probability": 0.9990234375
        },
        {
          "word": "的",
          "start": 293.05999999999995,
          "end": 293.15999999999997,
          "probability": 0.794921875
        },
        {
          "word": "事",
          "start": 293.15999999999997,
          "end": 293.32,
          "probability": 0.91748046875
        },
        {
          "word": "交",
          "start": 293.32,
          "end": 293.73999999999995,
          "probability": 0.95654296875
        },
        {
          "word": "给",
          "start": 293.73999999999995,
          "end": 293.87999999999994,
          "probability": 0.99951171875
        },
        {
          "word": "专",
          "start": 293.87999999999994,
          "end": 294.05999999999995,
          "probability": 1.0
        },
        {
          "word": "业",
          "start": 294.05999999999995,
          "end": 294.17999999999995,
          "probability": 1.0
        },
        {
          "word": "的",
          "start": 294.17999999999995,
          "end": 294.26,
          "probability": 0.99951171875
        },
        {
          "word": "工",
          "start": 294.26,
          "end": 294.43999999999994,
          "probability": 0.99951171875
        },
        {
          "word": "具",
          "start": 294.43999999999994,
          "end": 294.58,
          "probability": 0.99951171875
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.09536371839807388,
        "compression_ratio": 1.224,
        "no_speech_prob": 7.165966431355031e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "用它们在电脑上部署好再通过本地API接入同传字幕专业的事交给专业的工具才更符合这个项目的定位这还带来了一个附带的好处以前为了照顾国内网络环境我做了很多模型下载失配现在只剩语音识别模型干脆直接内置进安装包把下载"
        }
      ],
      "id": 122,
      "review_annotation": "",
      "row_id": "122",
      "bounded_row_id": 122,
      "parent_row_ids": [
        122
      ],
      "parent_row_id": 122,
      "packet_id": 11,
      "LID_evidence": [
        {
          "region_id": 58,
          "probabilities": {
            "zh": 0.9979304944475492
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        58
      ],
      "language_confidence": 0.9979304944475492,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 287.09999999999997,
      "packet_end": 310.5,
      "raw_text": "专业的事交给专业的工具",
      "selected_text": "专业的事交给专业的工具",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            122
          ],
          "packet_id": 11,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        11
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 122
    },
    {
      "start": 294.96,
      "end": 296.87999999999994,
      "text": "才更符合这个项目的定位",
      "raw_asr_text": "才更符合这个项目的定位",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 11,
      "words": [
        {
          "word": "才",
          "start": 294.96,
          "end": 295.28,
          "probability": 0.96923828125
        },
        {
          "word": "更",
          "start": 295.28,
          "end": 295.47999999999996,
          "probability": 0.9990234375
        },
        {
          "word": "符",
          "start": 295.47999999999996,
          "end": 295.65999999999997,
          "probability": 1.0
        },
        {
          "word": "合",
          "start": 295.65999999999997,
          "end": 295.82,
          "probability": 1.0
        },
        {
          "word": "这个",
          "start": 295.82,
          "end": 296.2,
          "probability": 0.99462890625
        },
        {
          "word": "项",
          "start": 296.2,
          "end": 296.37999999999994,
          "probability": 1.0
        },
        {
          "word": "目",
          "start": 296.37999999999994,
          "end": 296.47999999999996,
          "probability": 1.0
        },
        {
          "word": "的",
          "start": 296.47999999999996,
          "end": 296.55999999999995,
          "probability": 1.0
        },
        {
          "word": "定",
          "start": 296.55999999999995,
          "end": 296.67999999999995,
          "probability": 0.99951171875
        },
        {
          "word": "位",
          "start": 296.67999999999995,
          "end": 296.87999999999994,
          "probability": 1.0
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.09536371839807388,
        "compression_ratio": 1.224,
        "no_speech_prob": 7.165966431355031e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "用它们在电脑上部署好再通过本地API接入同传字幕专业的事交给专业的工具才更符合这个项目的定位这还带来了一个附带的好处以前为了照顾国内网络环境我做了很多模型下载失配现在只剩语音识别模型干脆直接内置进安装包把下载"
        }
      ],
      "id": 123,
      "review_annotation": "",
      "row_id": "123",
      "bounded_row_id": 123,
      "parent_row_ids": [
        123
      ],
      "parent_row_id": 123,
      "packet_id": 11,
      "LID_evidence": [
        {
          "region_id": 58,
          "probabilities": {
            "zh": 0.9979304944475492
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        58
      ],
      "language_confidence": 0.9979304944475492,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 287.09999999999997,
      "packet_end": 310.5,
      "raw_text": "才更符合这个项目的定位",
      "selected_text": "才更符合这个项目的定位",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            123
          ],
          "packet_id": 11,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        11
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 123
    },
    {
      "start": 297.81999999999994,
      "end": 299.76,
      "text": "这还带来了一个附带的好处",
      "raw_asr_text": "这还带来了一个附带的好处",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 11,
      "words": [
        {
          "word": "这",
          "start": 297.81999999999994,
          "end": 298.14,
          "probability": 0.9755859375
        },
        {
          "word": "还",
          "start": 298.14,
          "end": 298.28,
          "probability": 0.99755859375
        },
        {
          "word": "带",
          "start": 298.28,
          "end": 298.46,
          "probability": 1.0
        },
        {
          "word": "来了",
          "start": 298.46,
          "end": 298.65999999999997,
          "probability": 0.998046875
        },
        {
          "word": "一个",
          "start": 298.65999999999997,
          "end": 298.9,
          "probability": 0.9990234375
        },
        {
          "word": "附",
          "start": 298.9,
          "end": 299.11999999999995,
          "probability": 0.9794921875
        },
        {
          "word": "带",
          "start": 299.11999999999995,
          "end": 299.26,
          "probability": 0.9970703125
        },
        {
          "word": "的",
          "start": 299.26,
          "end": 299.37999999999994,
          "probability": 0.99951171875
        },
        {
          "word": "好",
          "start": 299.37999999999994,
          "end": 299.53999999999996,
          "probability": 1.0
        },
        {
          "word": "处",
          "start": 299.53999999999996,
          "end": 299.76,
          "probability": 0.99951171875
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.09536371839807388,
        "compression_ratio": 1.224,
        "no_speech_prob": 7.165966431355031e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "用它们在电脑上部署好再通过本地API接入同传字幕专业的事交给专业的工具才更符合这个项目的定位这还带来了一个附带的好处以前为了照顾国内网络环境我做了很多模型下载失配现在只剩语音识别模型干脆直接内置进安装包把下载"
        }
      ],
      "id": 124,
      "review_annotation": "",
      "row_id": "124",
      "bounded_row_id": 124,
      "parent_row_ids": [
        124
      ],
      "parent_row_id": 124,
      "packet_id": 11,
      "LID_evidence": [
        {
          "region_id": 59,
          "probabilities": {
            "zh": 0.9977846178743575
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        59
      ],
      "language_confidence": 0.9977846178743575,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 287.09999999999997,
      "packet_end": 310.5,
      "raw_text": "这还带来了一个附带的好处",
      "selected_text": "这还带来了一个附带的好处",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            124
          ],
          "packet_id": 11,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        11
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 124
    },
    {
      "start": 300.4,
      "end": 302.2,
      "text": "以前为了照顾国内网络环境",
      "raw_asr_text": "以前为了照顾国内网络环境",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 11,
      "words": [
        {
          "word": "以前",
          "start": 300.4,
          "end": 300.71999999999997,
          "probability": 0.998046875
        },
        {
          "word": "为",
          "start": 300.71999999999997,
          "end": 300.87999999999994,
          "probability": 0.99951171875
        },
        {
          "word": "了",
          "start": 300.87999999999994,
          "end": 300.97999999999996,
          "probability": 1.0
        },
        {
          "word": "照",
          "start": 300.97999999999996,
          "end": 301.11999999999995,
          "probability": 0.9990234375
        },
        {
          "word": "顾",
          "start": 301.11999999999995,
          "end": 301.26,
          "probability": 0.9990234375
        },
        {
          "word": "国",
          "start": 301.26,
          "end": 301.41999999999996,
          "probability": 0.99951171875
        },
        {
          "word": "内",
          "start": 301.41999999999996,
          "end": 301.55999999999995,
          "probability": 1.0
        },
        {
          "word": "网",
          "start": 301.55999999999995,
          "end": 301.71999999999997,
          "probability": 1.0
        },
        {
          "word": "络",
          "start": 301.71999999999997,
          "end": 301.82,
          "probability": 0.9970703125
        },
        {
          "word": "环",
          "start": 301.82,
          "end": 302.03999999999996,
          "probability": 1.0
        },
        {
          "word": "境",
          "start": 302.03999999999996,
          "end": 302.2,
          "probability": 1.0
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.09536371839807388,
        "compression_ratio": 1.224,
        "no_speech_prob": 7.165966431355031e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "用它们在电脑上部署好再通过本地API接入同传字幕专业的事交给专业的工具才更符合这个项目的定位这还带来了一个附带的好处以前为了照顾国内网络环境我做了很多模型下载失配现在只剩语音识别模型干脆直接内置进安装包把下载"
        }
      ],
      "id": 125,
      "review_annotation": "",
      "row_id": "125",
      "bounded_row_id": 125,
      "parent_row_ids": [
        125
      ],
      "parent_row_id": 125,
      "packet_id": 11,
      "LID_evidence": [
        {
          "region_id": 59,
          "probabilities": {
            "zh": 0.9977846178743575
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        59
      ],
      "language_confidence": 0.9977846178743575,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 287.09999999999997,
      "packet_end": 310.5,
      "raw_text": "以前为了照顾国内网络环境",
      "selected_text": "以前为了照顾国内网络环境",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            125
          ],
          "packet_id": 11,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        11
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 125
    },
    {
      "start": 302.2,
      "end": 304.47999999999996,
      "text": "模型下载失配",
      "raw_asr_text": "我做了很多模型下载失配",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 11,
      "words": [],
      "asr_metrics": {
        "avg_logprob": -0.09536371839807388,
        "compression_ratio": 1.224,
        "no_speech_prob": 7.165966431355031e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "用它们在电脑上部署好再通过本地API接入同传字幕专业的事交给专业的工具才更符合这个项目的定位这还带来了一个附带的好处以前为了照顾国内网络环境我做了很多模型下载失配现在只剩语音识别模型干脆直接内置进安装包把下载"
        }
      ],
      "id": 126,
      "review_annotation": "待確認",
      "row_id": "126",
      "bounded_row_id": 126,
      "parent_row_ids": [
        126
      ],
      "parent_row_id": 126,
      "packet_id": 11,
      "LID_evidence": [
        {
          "region_id": 59,
          "probabilities": {
            "zh": 0.9977846178743575
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        59
      ],
      "language_confidence": 0.9977846178743575,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 287.09999999999997,
      "packet_end": 310.5,
      "raw_text": "我做了很多模型下载失配",
      "selected_text": "模型下载失配",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            126
          ],
          "packet_id": 11,
          "engine": "mlx-whisper"
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "126",
          "before": "我做了很多模型下载失配",
          "after": "模型下载失配",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "失配疑似适配误识",
          "context_evidence": "上下文为国内网络环境，'适配'更符合逻辑",
          "independent_ASR": {
            "status": "REVIEW_ONLY_RESCUE_BUDGET"
          },
          "suppression_proposed": null
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "126",
          "before": "模型下载失配",
          "after": "模型下载失配",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "Nonsense: '失配' vs '失败' (failure)",
          "context_evidence": "Context: '模型下载失败'",
          "independent_ASR": {
            "status": "REVIEW_ONLY_RESCUE_BUDGET"
          },
          "suppression_proposed": null
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "126",
          "before": "模型下载失配",
          "after": "模型下载失配",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "Nonsense: '失配' vs '失败' (failure)",
          "context_evidence": "Context: '模型下载失败'",
          "independent_ASR": {
            "status": "REVIEW_ONLY_RESCUE_BUDGET"
          },
          "suppression_proposed": null
        }
      ],
      "language": "zh",
      "packet_ids": [
        11
      ],
      "classification": "BEST_SUPPORTED_WITH_REVIEW",
      "independent_ASR": {
        "status": "REVIEW_ONLY_RESCUE_BUDGET"
      },
      "raw_word_timestamps": [
        {
          "word": "我",
          "start": 302.2,
          "end": 302.61999999999995,
          "probability": 0.99658203125
        },
        {
          "word": "做",
          "start": 302.61999999999995,
          "end": 302.73999999999995,
          "probability": 0.99951171875
        },
        {
          "word": "了",
          "start": 302.73999999999995,
          "end": 302.84,
          "probability": 1.0
        },
        {
          "word": "很多",
          "start": 302.84,
          "end": 303.11999999999995,
          "probability": 1.0
        },
        {
          "word": "模",
          "start": 303.11999999999995,
          "end": 303.31999999999994,
          "probability": 0.99951171875
        },
        {
          "word": "型",
          "start": 303.31999999999994,
          "end": 303.59999999999997,
          "probability": 0.9990234375
        },
        {
          "word": "下",
          "start": 303.59999999999997,
          "end": 303.99999999999994,
          "probability": 0.9794921875
        },
        {
          "word": "载",
          "start": 303.99999999999994,
          "end": 304.15999999999997,
          "probability": 0.994140625
        },
        {
          "word": "失",
          "start": 304.15999999999997,
          "end": 304.31999999999994,
          "probability": 0.96044921875
        },
        {
          "word": "配",
          "start": 304.31999999999994,
          "end": 304.47999999999996,
          "probability": 0.7998046875
        }
      ],
      "word_timestamps": [],
      "source_text": "模型下载失配",
      "review_status": "BEST_SUPPORTED_WITH_REVIEW",
      "canonical_timing_evidence": "CUE_ANCHORED",
      "index": 126
    },
    {
      "start": 305.38,
      "end": 307.2,
      "text": "现在只剩语音识别模型",
      "raw_asr_text": "现在只剩语音识别模型",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 11,
      "words": [
        {
          "word": "现在",
          "start": 305.38,
          "end": 305.7,
          "probability": 0.99755859375
        },
        {
          "word": "只",
          "start": 305.7,
          "end": 305.91999999999996,
          "probability": 0.99072265625
        },
        {
          "word": "剩",
          "start": 305.91999999999996,
          "end": 306.09999999999997,
          "probability": 0.9912109375
        },
        {
          "word": "语",
          "start": 306.09999999999997,
          "end": 306.26,
          "probability": 0.99560546875
        },
        {
          "word": "音",
          "start": 306.26,
          "end": 306.46,
          "probability": 0.9990234375
        },
        {
          "word": "识",
          "start": 306.46,
          "end": 306.61999999999995,
          "probability": 0.99658203125
        },
        {
          "word": "别",
          "start": 306.61999999999995,
          "end": 306.76,
          "probability": 0.9990234375
        },
        {
          "word": "模",
          "start": 306.76,
          "end": 306.91999999999996,
          "probability": 0.99951171875
        },
        {
          "word": "型",
          "start": 306.91999999999996,
          "end": 307.2,
          "probability": 1.0
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.09536371839807388,
        "compression_ratio": 1.224,
        "no_speech_prob": 7.165966431355031e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "用它们在电脑上部署好再通过本地API接入同传字幕专业的事交给专业的工具才更符合这个项目的定位这还带来了一个附带的好处以前为了照顾国内网络环境我做了很多模型下载失配现在只剩语音识别模型干脆直接内置进安装包把下载"
        }
      ],
      "id": 127,
      "review_annotation": "",
      "row_id": "127",
      "bounded_row_id": 127,
      "parent_row_ids": [
        127
      ],
      "parent_row_id": 127,
      "packet_id": 11,
      "LID_evidence": [
        {
          "region_id": 60,
          "probabilities": {
            "zh": 0.9970128719623272
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        60
      ],
      "language_confidence": 0.9970128719623272,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 287.09999999999997,
      "packet_end": 310.5,
      "raw_text": "现在只剩语音识别模型",
      "selected_text": "现在只剩语音识别模型",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            127
          ],
          "packet_id": 11,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        11
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 127
    },
    {
      "start": 307.2,
      "end": 309.34,
      "text": "干脆直接内置进安装包",
      "raw_asr_text": "干脆直接内置进安装包",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 11,
      "words": [
        {
          "word": "干",
          "start": 307.2,
          "end": 307.67999999999995,
          "probability": 0.99560546875
        },
        {
          "word": "脆",
          "start": 307.67999999999995,
          "end": 307.93999999999994,
          "probability": 0.998046875
        },
        {
          "word": "直接",
          "start": 307.93999999999994,
          "end": 308.29999999999995,
          "probability": 0.99951171875
        },
        {
          "word": "内",
          "start": 308.29999999999995,
          "end": 308.47999999999996,
          "probability": 0.98388671875
        },
        {
          "word": "置",
          "start": 308.47999999999996,
          "end": 308.64,
          "probability": 0.99267578125
        },
        {
          "word": "进",
          "start": 308.64,
          "end": 308.78,
          "probability": 0.998046875
        },
        {
          "word": "安",
          "start": 308.78,
          "end": 308.96,
          "probability": 0.998046875
        },
        {
          "word": "装",
          "start": 308.96,
          "end": 309.11999999999995,
          "probability": 1.0
        },
        {
          "word": "包",
          "start": 309.11999999999995,
          "end": 309.34,
          "probability": 0.9951171875
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.09536371839807388,
        "compression_ratio": 1.224,
        "no_speech_prob": 7.165966431355031e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "用它们在电脑上部署好再通过本地API接入同传字幕专业的事交给专业的工具才更符合这个项目的定位这还带来了一个附带的好处以前为了照顾国内网络环境我做了很多模型下载失配现在只剩语音识别模型干脆直接内置进安装包把下载"
        }
      ],
      "id": 128,
      "review_annotation": "",
      "row_id": "128",
      "bounded_row_id": 128,
      "parent_row_ids": [
        128
      ],
      "parent_row_id": 128,
      "packet_id": 11,
      "LID_evidence": [
        {
          "region_id": 60,
          "probabilities": {
            "zh": 0.9970128719623272
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        60
      ],
      "language_confidence": 0.9970128719623272,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 287.09999999999997,
      "packet_end": 310.5,
      "raw_text": "干脆直接内置进安装包",
      "selected_text": "干脆直接内置进安装包",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            128
          ],
          "packet_id": 11,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        11
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 128
    },
    {
      "start": 309.8,
      "end": 311.8,
      "text": "把下载部分彻底剥离掉",
      "raw_asr_text": "把下载部分彻底剥离掉",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 12,
      "words": [
        {
          "word": "把",
          "start": 309.62,
          "end": 310.02,
          "probability": 0.0687255859375
        },
        {
          "word": "下",
          "start": 310.02,
          "end": 310.22,
          "probability": 0.998046875
        },
        {
          "word": "载",
          "start": 310.22,
          "end": 310.44,
          "probability": 0.9951171875
        },
        {
          "word": "部分",
          "start": 310.44,
          "end": 310.7,
          "probability": 0.99462890625
        },
        {
          "word": "彻",
          "start": 310.7,
          "end": 311.18,
          "probability": 0.99755859375
        },
        {
          "word": "底",
          "start": 311.18,
          "end": 311.34,
          "probability": 1.0
        },
        {
          "word": "剥",
          "start": 311.34,
          "end": 311.5,
          "probability": 0.958984375
        },
        {
          "word": "离",
          "start": 311.5,
          "end": 311.62,
          "probability": 0.986328125
        },
        {
          "word": "掉",
          "start": 311.62,
          "end": 311.8,
          "probability": 0.9970703125
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.16505358559744698,
        "compression_ratio": 0.9554140127388535,
        "no_speech_prob": 4.084320381902984e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "把下载部分彻底剥离掉也方便大家通过Titop以外的渠道分享传播只保留BYOK你选你信任的服务你来掌控成本和稳定性"
        }
      ],
      "id": 129,
      "review_annotation": "",
      "row_id": "129",
      "bounded_row_id": 129,
      "parent_row_ids": [
        129
      ],
      "parent_row_id": 129,
      "packet_id": 12,
      "LID_evidence": [
        {
          "region_id": 61,
          "probabilities": {
            "zh": 0.9963418946546667
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        61
      ],
      "language_confidence": 0.9963418946546667,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 309.0,
      "packet_end": 324.40000000000003,
      "raw_text": "把下载部分彻底剥离掉",
      "selected_text": "把下载部分彻底剥离掉",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            129
          ],
          "packet_id": 12,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        12
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 129
    },
    {
      "start": 311.8,
      "end": 316.2,
      "text": "也方便大家通过Titop以外的渠道分享传播",
      "raw_asr_text": "也方便大家通过Titop以外的渠道分享传播",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 12,
      "words": [
        {
          "word": "也",
          "start": 311.8,
          "end": 312.36,
          "probability": 0.8974609375
        },
        {
          "word": "方",
          "start": 312.36,
          "end": 312.56,
          "probability": 0.99853515625
        },
        {
          "word": "便",
          "start": 312.56,
          "end": 312.74,
          "probability": 1.0
        },
        {
          "word": "大家",
          "start": 312.74,
          "end": 313.2,
          "probability": 0.99853515625
        },
        {
          "word": "通",
          "start": 313.2,
          "end": 313.42,
          "probability": 0.9990234375
        },
        {
          "word": "过",
          "start": 313.42,
          "end": 313.64,
          "probability": 0.99951171875
        },
        {
          "word": "T",
          "start": 313.64,
          "end": 313.9,
          "probability": 0.70703125
        },
        {
          "word": "it",
          "start": 313.9,
          "end": 314.0,
          "probability": 0.16455078125
        },
        {
          "word": "op",
          "start": 314.0,
          "end": 314.2,
          "probability": 0.31884765625
        },
        {
          "word": "以",
          "start": 314.2,
          "end": 314.6,
          "probability": 0.99560546875
        },
        {
          "word": "外",
          "start": 314.6,
          "end": 314.78,
          "probability": 1.0
        },
        {
          "word": "的",
          "start": 314.78,
          "end": 314.9,
          "probability": 0.99951171875
        },
        {
          "word": "渠",
          "start": 314.9,
          "end": 315.0,
          "probability": 0.92626953125
        },
        {
          "word": "道",
          "start": 315.0,
          "end": 315.2,
          "probability": 0.998046875
        },
        {
          "word": "分享",
          "start": 315.2,
          "end": 315.72,
          "probability": 0.9990234375
        },
        {
          "word": "传",
          "start": 315.72,
          "end": 315.96,
          "probability": 0.9501953125
        },
        {
          "word": "播",
          "start": 315.96,
          "end": 316.2,
          "probability": 1.0
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.16505358559744698,
        "compression_ratio": 0.9554140127388535,
        "no_speech_prob": 4.084320381902984e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "把下载部分彻底剥离掉也方便大家通过Titop以外的渠道分享传播只保留BYOK你选你信任的服务你来掌控成本和稳定性"
        }
      ],
      "id": 130,
      "review_annotation": "待確認",
      "row_id": "130",
      "bounded_row_id": 130,
      "parent_row_ids": [
        130
      ],
      "parent_row_id": 130,
      "packet_id": 12,
      "LID_evidence": [
        {
          "region_id": 61,
          "probabilities": {
            "zh": 0.9963418946546667
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        61
      ],
      "language_confidence": 0.9963418946546667,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 309.0,
      "packet_end": 324.40000000000003,
      "raw_text": "也方便大家通过Titop以外的渠道分享传播",
      "selected_text": "也方便大家通过Titop以外的渠道分享传播",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            130
          ],
          "packet_id": 12,
          "engine": "mlx-whisper"
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "130",
          "before": "也方便大家通过Titop以外的渠道分享传播",
          "after": "也方便大家通过Titop以外的渠道分享传播",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "Titop疑似GitHub或特定平台误识",
          "context_evidence": "上下文为分享传播渠道，'GitHub'为常见平台",
          "independent_ASR": {
            "status": "REVIEW_ONLY_RESCUE_BUDGET"
          },
          "suppression_proposed": null
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "130",
          "before": "也方便大家通过Titop以外的渠道分享传播",
          "after": "也方便大家通过Titop以外的渠道分享传播",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "Name error: 'Titop' vs 'GitHub'",
          "context_evidence": "Context: '通过GitHub以外的渠道'",
          "independent_ASR": {
            "status": "REVIEW_ONLY_RESCUE_BUDGET"
          },
          "suppression_proposed": null
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "130",
          "before": "也方便大家通过Titop以外的渠道分享传播",
          "after": "也方便大家通过Titop以外的渠道分享传播",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "Name error: 'Titop' vs 'GitHub'",
          "context_evidence": "Context: '通过GitHub以外的渠道'",
          "independent_ASR": {
            "status": "REVIEW_ONLY_RESCUE_BUDGET"
          },
          "suppression_proposed": null
        }
      ],
      "language": "zh",
      "packet_ids": [
        12
      ],
      "classification": "BEST_SUPPORTED_WITH_REVIEW",
      "independent_ASR": {
        "status": "REVIEW_ONLY_RESCUE_BUDGET"
      },
      "source_text": "也方便大家通过Titop以外的渠道分享传播",
      "review_status": "BEST_SUPPORTED_WITH_REVIEW",
      "canonical_timing_evidence": "CUE_ANCHORED",
      "index": 130
    },
    {
      "start": 316.86,
      "end": 318.36,
      "text": "只保留BYOK",
      "raw_asr_text": "只保留BYOK",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 12,
      "words": [
        {
          "word": "只",
          "start": 316.86,
          "end": 317.26,
          "probability": 0.892578125
        },
        {
          "word": "保",
          "start": 317.26,
          "end": 317.4,
          "probability": 0.9970703125
        },
        {
          "word": "留",
          "start": 317.4,
          "end": 317.66,
          "probability": 0.9990234375
        },
        {
          "word": "BY",
          "start": 317.66,
          "end": 317.92,
          "probability": 0.8310546875
        },
        {
          "word": "OK",
          "start": 317.92,
          "end": 318.36,
          "probability": 0.91845703125
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.16505358559744698,
        "compression_ratio": 0.9554140127388535,
        "no_speech_prob": 4.084320381902984e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "把下载部分彻底剥离掉也方便大家通过Titop以外的渠道分享传播只保留BYOK你选你信任的服务你来掌控成本和稳定性"
        }
      ],
      "id": 131,
      "review_annotation": "",
      "row_id": "131",
      "bounded_row_id": 131,
      "parent_row_ids": [
        131
      ],
      "parent_row_id": 131,
      "packet_id": 12,
      "LID_evidence": [
        {
          "region_id": 62,
          "probabilities": {
            "zh": 0.9897701569965908
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        62
      ],
      "language_confidence": 0.9897701569965908,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 309.0,
      "packet_end": 324.40000000000003,
      "raw_text": "只保留BYOK",
      "selected_text": "只保留BYOK",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            131
          ],
          "packet_id": 12,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        12
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 131
    },
    {
      "start": 318.94,
      "end": 320.54,
      "text": "你选你信任的服务",
      "raw_asr_text": "你选你信任的服务",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 12,
      "words": [
        {
          "word": "你",
          "start": 318.94,
          "end": 319.34,
          "probability": 0.97900390625
        },
        {
          "word": "选",
          "start": 319.34,
          "end": 319.6,
          "probability": 0.94140625
        },
        {
          "word": "你",
          "start": 319.6,
          "end": 319.74,
          "probability": 0.9990234375
        },
        {
          "word": "信",
          "start": 319.74,
          "end": 319.92,
          "probability": 0.90478515625
        },
        {
          "word": "任",
          "start": 319.92,
          "end": 320.06,
          "probability": 0.9990234375
        },
        {
          "word": "的",
          "start": 320.06,
          "end": 320.18,
          "probability": 0.99951171875
        },
        {
          "word": "服",
          "start": 320.18,
          "end": 320.34,
          "probability": 1.0
        },
        {
          "word": "务",
          "start": 320.34,
          "end": 320.54,
          "probability": 1.0
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.16505358559744698,
        "compression_ratio": 0.9554140127388535,
        "no_speech_prob": 4.084320381902984e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "把下载部分彻底剥离掉也方便大家通过Titop以外的渠道分享传播只保留BYOK你选你信任的服务你来掌控成本和稳定性"
        }
      ],
      "id": 132,
      "review_annotation": "",
      "row_id": "132",
      "bounded_row_id": 132,
      "parent_row_ids": [
        132
      ],
      "parent_row_id": 132,
      "packet_id": 12,
      "LID_evidence": [
        {
          "region_id": 63,
          "probabilities": {
            "zh": 0.9929605552128383
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        63
      ],
      "language_confidence": 0.9929605552128383,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 309.0,
      "packet_end": 324.40000000000003,
      "raw_text": "你选你信任的服务",
      "selected_text": "你选你信任的服务",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            132
          ],
          "packet_id": 12,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        12
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 132
    },
    {
      "start": 320.54,
      "end": 323.28,
      "text": "你来掌控成本和稳定性",
      "raw_asr_text": "你来掌控成本和稳定性",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 12,
      "words": [
        {
          "word": "你",
          "start": 320.54,
          "end": 321.1,
          "probability": 0.98583984375
        },
        {
          "word": "来",
          "start": 321.1,
          "end": 321.3,
          "probability": 1.0
        },
        {
          "word": "掌",
          "start": 321.3,
          "end": 321.5,
          "probability": 0.9990234375
        },
        {
          "word": "控",
          "start": 321.5,
          "end": 321.66,
          "probability": 0.998046875
        },
        {
          "word": "成",
          "start": 321.66,
          "end": 321.86,
          "probability": 1.0
        },
        {
          "word": "本",
          "start": 321.86,
          "end": 322.12,
          "probability": 0.99951171875
        },
        {
          "word": "和",
          "start": 322.12,
          "end": 322.66,
          "probability": 0.99609375
        },
        {
          "word": "稳",
          "start": 322.66,
          "end": 322.8,
          "probability": 0.998046875
        },
        {
          "word": "定",
          "start": 322.8,
          "end": 323.02,
          "probability": 0.99951171875
        },
        {
          "word": "性",
          "start": 323.02,
          "end": 323.28,
          "probability": 0.99951171875
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.16505358559744698,
        "compression_ratio": 0.9554140127388535,
        "no_speech_prob": 4.084320381902984e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "把下载部分彻底剥离掉也方便大家通过Titop以外的渠道分享传播只保留BYOK你选你信任的服务你来掌控成本和稳定性"
        }
      ],
      "id": 133,
      "review_annotation": "",
      "row_id": "133",
      "bounded_row_id": 133,
      "parent_row_ids": [
        133
      ],
      "parent_row_id": 133,
      "packet_id": 12,
      "LID_evidence": [
        {
          "region_id": 63,
          "probabilities": {
            "zh": 0.9929605552128383
          },
          "status": "STABLE"
        },
        {
          "region_id": 64,
          "probabilities": {
            "zh": 0.9978526473045349
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        63,
        64
      ],
      "language_confidence": 0.9978526473045349,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 309.0,
      "packet_end": 324.40000000000003,
      "raw_text": "你来掌控成本和稳定性",
      "selected_text": "你来掌控成本和稳定性",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            133
          ],
          "packet_id": 12,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        12
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 133
    },
    {
      "start": 328.2,
      "end": 329.08,
      "text": "摇完了功能",
      "raw_asr_text": "摇完了功能",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 13,
      "words": [
        {
          "word": "摇",
          "start": 328.03999999999996,
          "end": 328.44,
          "probability": 0.84130859375
        },
        {
          "word": "完了",
          "start": 328.44,
          "end": 328.64,
          "probability": 0.9765625
        },
        {
          "word": "功",
          "start": 328.64,
          "end": 328.85999999999996,
          "probability": 0.9990234375
        },
        {
          "word": "能",
          "start": 328.85999999999996,
          "end": 329.08,
          "probability": 0.99951171875
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.19097018241882324,
        "compression_ratio": 1.2310756972111554,
        "no_speech_prob": 1.463243415322779e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "摇完了功能最后想跟大家说说做这个项目的初衷其实特别简单前阵子我看火雨直播的时候平时用的千万同传因为看的人太多突然炸掉了那天半夜直播结束越想越上头的我决定干脆自己动手搓一个一开始真的只是打算留着自己用的但后来"
        }
      ],
      "id": 134,
      "review_annotation": "待確認",
      "row_id": "134",
      "bounded_row_id": 134,
      "parent_row_ids": [
        134
      ],
      "parent_row_id": 134,
      "packet_id": 13,
      "LID_evidence": [
        {
          "region_id": 65,
          "probabilities": {
            "zh": 0.9960214495658875
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        65
      ],
      "language_confidence": 0.9960214495658875,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 327.4,
      "packet_end": 353.8,
      "raw_text": "摇完了功能",
      "selected_text": "摇完了功能",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            134
          ],
          "packet_id": 13,
          "engine": "mlx-whisper"
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "134",
          "before": "摇完了功能",
          "after": "摇完了功能",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "摇完疑似聊完或讲完误识",
          "context_evidence": "上下文为功能介绍结束，'聊完'或'讲完'更符合",
          "independent_ASR": {
            "status": "REVIEW_ONLY_RESCUE_BUDGET"
          },
          "suppression_proposed": null
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "134",
          "before": "摇完了功能",
          "after": "摇完了功能",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "Nonsense: '摇完' vs '聊完' (finished chatting)",
          "context_evidence": "Context: '聊完了功能'",
          "independent_ASR": {
            "status": "REVIEW_ONLY_RESCUE_BUDGET"
          },
          "suppression_proposed": null
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "134",
          "before": "摇完了功能",
          "after": "摇完了功能",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "Nonsense: '摇完' vs '聊完' (finished chatting)",
          "context_evidence": "Context: '聊完了功能'",
          "independent_ASR": {
            "status": "REVIEW_ONLY_RESCUE_BUDGET"
          },
          "suppression_proposed": null
        }
      ],
      "language": "zh",
      "packet_ids": [
        13
      ],
      "classification": "BEST_SUPPORTED_WITH_REVIEW",
      "independent_ASR": {
        "status": "REVIEW_ONLY_RESCUE_BUDGET"
      },
      "source_text": "摇完了功能",
      "review_status": "BEST_SUPPORTED_WITH_REVIEW",
      "canonical_timing_evidence": "CUE_ANCHORED",
      "index": 134
    },
    {
      "start": 329.08,
      "end": 331.4,
      "text": "最后想跟大家说说",
      "raw_asr_text": "最后想跟大家说说",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 13,
      "words": [
        {
          "word": "最",
          "start": 329.08,
          "end": 329.78,
          "probability": 0.92822265625
        },
        {
          "word": "后",
          "start": 329.78,
          "end": 330.09999999999997,
          "probability": 1.0
        },
        {
          "word": "想",
          "start": 330.09999999999997,
          "end": 330.56,
          "probability": 0.9912109375
        },
        {
          "word": "跟大家",
          "start": 330.56,
          "end": 330.91999999999996,
          "probability": 0.99951171875
        },
        {
          "word": "说",
          "start": 330.91999999999996,
          "end": 331.21999999999997,
          "probability": 1.0
        },
        {
          "word": "说",
          "start": 331.21999999999997,
          "end": 331.4,
          "probability": 0.99951171875
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.19097018241882324,
        "compression_ratio": 1.2310756972111554,
        "no_speech_prob": 1.463243415322779e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "摇完了功能最后想跟大家说说做这个项目的初衷其实特别简单前阵子我看火雨直播的时候平时用的千万同传因为看的人太多突然炸掉了那天半夜直播结束越想越上头的我决定干脆自己动手搓一个一开始真的只是打算留着自己用的但后来"
        }
      ],
      "id": 135,
      "review_annotation": "",
      "row_id": "135",
      "bounded_row_id": 135,
      "parent_row_ids": [
        135
      ],
      "parent_row_id": 135,
      "packet_id": 13,
      "LID_evidence": [
        {
          "region_id": 65,
          "probabilities": {
            "zh": 0.9960214495658875
          },
          "status": "STABLE"
        },
        {
          "region_id": 66,
          "probabilities": {
            "zh": 0.9974024772644043
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        65,
        66
      ],
      "language_confidence": 0.9974024772644043,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 327.4,
      "packet_end": 353.8,
      "raw_text": "最后想跟大家说说",
      "selected_text": "最后想跟大家说说",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            135
          ],
          "packet_id": 13,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        13
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 135
    },
    {
      "start": 331.4,
      "end": 332.78,
      "text": "做这个项目的初衷",
      "raw_asr_text": "做这个项目的初衷",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 13,
      "words": [
        {
          "word": "做",
          "start": 331.4,
          "end": 331.7,
          "probability": 0.99658203125
        },
        {
          "word": "这个",
          "start": 331.7,
          "end": 331.9,
          "probability": 1.0
        },
        {
          "word": "项",
          "start": 331.9,
          "end": 332.09999999999997,
          "probability": 0.9990234375
        },
        {
          "word": "目",
          "start": 332.09999999999997,
          "end": 332.23999999999995,
          "probability": 1.0
        },
        {
          "word": "的",
          "start": 332.23999999999995,
          "end": 332.35999999999996,
          "probability": 1.0
        },
        {
          "word": "初",
          "start": 332.35999999999996,
          "end": 332.53999999999996,
          "probability": 0.9580078125
        },
        {
          "word": "衷",
          "start": 332.53999999999996,
          "end": 332.78,
          "probability": 0.99267578125
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.19097018241882324,
        "compression_ratio": 1.2310756972111554,
        "no_speech_prob": 1.463243415322779e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "摇完了功能最后想跟大家说说做这个项目的初衷其实特别简单前阵子我看火雨直播的时候平时用的千万同传因为看的人太多突然炸掉了那天半夜直播结束越想越上头的我决定干脆自己动手搓一个一开始真的只是打算留着自己用的但后来"
        }
      ],
      "id": 136,
      "review_annotation": "",
      "row_id": "136",
      "bounded_row_id": 136,
      "parent_row_ids": [
        136
      ],
      "parent_row_id": 136,
      "packet_id": 13,
      "LID_evidence": [
        {
          "region_id": 66,
          "probabilities": {
            "zh": 0.9974024772644043
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        66
      ],
      "language_confidence": 0.9974024772644043,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 327.4,
      "packet_end": 353.8,
      "raw_text": "做这个项目的初衷",
      "selected_text": "做这个项目的初衷",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            136
          ],
          "packet_id": 13,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        13
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 136
    },
    {
      "start": 333.44,
      "end": 334.56,
      "text": "其实特别简单",
      "raw_asr_text": "其实特别简单",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 13,
      "words": [
        {
          "word": "其实",
          "start": 333.44,
          "end": 333.84,
          "probability": 0.9990234375
        },
        {
          "word": "特",
          "start": 333.84,
          "end": 334.06,
          "probability": 1.0
        },
        {
          "word": "别",
          "start": 334.06,
          "end": 334.2,
          "probability": 1.0
        },
        {
          "word": "简",
          "start": 334.2,
          "end": 334.38,
          "probability": 1.0
        },
        {
          "word": "单",
          "start": 334.38,
          "end": 334.56,
          "probability": 1.0
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.19097018241882324,
        "compression_ratio": 1.2310756972111554,
        "no_speech_prob": 1.463243415322779e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "摇完了功能最后想跟大家说说做这个项目的初衷其实特别简单前阵子我看火雨直播的时候平时用的千万同传因为看的人太多突然炸掉了那天半夜直播结束越想越上头的我决定干脆自己动手搓一个一开始真的只是打算留着自己用的但后来"
        }
      ],
      "id": 137,
      "review_annotation": "",
      "row_id": "137",
      "bounded_row_id": 137,
      "parent_row_ids": [
        137
      ],
      "parent_row_id": 137,
      "packet_id": 13,
      "LID_evidence": [
        {
          "region_id": 67,
          "probabilities": {
            "zh": 0.9982185278620038
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        67
      ],
      "language_confidence": 0.9982185278620038,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 327.4,
      "packet_end": 353.8,
      "raw_text": "其实特别简单",
      "selected_text": "其实特别简单",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            137
          ],
          "packet_id": 13,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        13
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 137
    },
    {
      "start": 335.08,
      "end": 337.46,
      "text": "火雨直播",
      "raw_asr_text": "前阵子我看火雨直播的时候",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 13,
      "words": [],
      "asr_metrics": {
        "avg_logprob": -0.19097018241882324,
        "compression_ratio": 1.2310756972111554,
        "no_speech_prob": 1.463243415322779e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "摇完了功能最后想跟大家说说做这个项目的初衷其实特别简单前阵子我看火雨直播的时候平时用的千万同传因为看的人太多突然炸掉了那天半夜直播结束越想越上头的我决定干脆自己动手搓一个一开始真的只是打算留着自己用的但后来"
        }
      ],
      "id": 138,
      "review_annotation": "待確認",
      "row_id": "138",
      "bounded_row_id": 138,
      "parent_row_ids": [
        138
      ],
      "parent_row_id": 138,
      "packet_id": 13,
      "LID_evidence": [
        {
          "region_id": 68,
          "probabilities": {
            "zh": 0.998153539498647
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        68
      ],
      "language_confidence": 0.998153539498647,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 327.4,
      "packet_end": 353.8,
      "raw_text": "前阵子我看火雨直播的时候",
      "selected_text": "火雨直播",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            138
          ],
          "packet_id": 13,
          "engine": "mlx-whisper"
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "138",
          "before": "前阵子我看火雨直播的时候",
          "after": "火雨直播",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "火雨疑似虎牙或特定直播误识",
          "context_evidence": "上下文为直播平台，'虎牙'为常见平台",
          "independent_ASR": {
            "status": "REVIEW_ONLY_RESCUE_BUDGET"
          },
          "suppression_proposed": null
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "138",
          "before": "火雨直播",
          "after": "火雨直播",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "Name error: '火雨' vs 'B站' (Bilibili)",
          "context_evidence": "Context: 'B站直播'",
          "independent_ASR": {
            "status": "REVIEW_ONLY_RESCUE_BUDGET"
          },
          "suppression_proposed": null
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "138",
          "before": "火雨直播",
          "after": "火雨直播",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "Name error: '火雨' vs 'B站' (Bilibili)",
          "context_evidence": "Context: 'B站直播'",
          "independent_ASR": {
            "status": "REVIEW_ONLY_RESCUE_BUDGET"
          },
          "suppression_proposed": null
        }
      ],
      "language": "zh",
      "packet_ids": [
        13
      ],
      "classification": "BEST_SUPPORTED_WITH_REVIEW",
      "independent_ASR": {
        "status": "REVIEW_ONLY_RESCUE_BUDGET"
      },
      "raw_word_timestamps": [
        {
          "word": "前",
          "start": 335.08,
          "end": 335.47999999999996,
          "probability": 0.99951171875
        },
        {
          "word": "阵",
          "start": 335.47999999999996,
          "end": 335.7,
          "probability": 0.998046875
        },
        {
          "word": "子",
          "start": 335.7,
          "end": 335.88,
          "probability": 1.0
        },
        {
          "word": "我",
          "start": 335.88,
          "end": 336.35999999999996,
          "probability": 0.99658203125
        },
        {
          "word": "看",
          "start": 336.35999999999996,
          "end": 336.5,
          "probability": 1.0
        },
        {
          "word": "火",
          "start": 336.5,
          "end": 336.67999999999995,
          "probability": 0.87451171875
        },
        {
          "word": "雨",
          "start": 336.67999999999995,
          "end": 336.82,
          "probability": 0.7841796875
        },
        {
          "word": "直",
          "start": 336.82,
          "end": 336.97999999999996,
          "probability": 1.0
        },
        {
          "word": "播",
          "start": 336.97999999999996,
          "end": 337.09999999999997,
          "probability": 1.0
        },
        {
          "word": "的时候",
          "start": 337.09999999999997,
          "end": 337.46,
          "probability": 1.0
        }
      ],
      "word_timestamps": [],
      "source_text": "火雨直播",
      "review_status": "BEST_SUPPORTED_WITH_REVIEW",
      "canonical_timing_evidence": "CUE_ANCHORED",
      "index": 138
    },
    {
      "start": 337.46,
      "end": 339.15999999999997,
      "text": "千万同传",
      "raw_asr_text": "平时用的千万同传",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 13,
      "words": [],
      "asr_metrics": {
        "avg_logprob": -0.19097018241882324,
        "compression_ratio": 1.2310756972111554,
        "no_speech_prob": 1.463243415322779e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "摇完了功能最后想跟大家说说做这个项目的初衷其实特别简单前阵子我看火雨直播的时候平时用的千万同传因为看的人太多突然炸掉了那天半夜直播结束越想越上头的我决定干脆自己动手搓一个一开始真的只是打算留着自己用的但后来"
        }
      ],
      "id": 139,
      "review_annotation": "待確認",
      "row_id": "139",
      "bounded_row_id": 139,
      "parent_row_ids": [
        139
      ],
      "parent_row_id": 139,
      "packet_id": 13,
      "LID_evidence": [
        {
          "region_id": 68,
          "probabilities": {
            "zh": 0.998153539498647
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        68
      ],
      "language_confidence": 0.998153539498647,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 327.4,
      "packet_end": 353.8,
      "raw_text": "平时用的千万同传",
      "selected_text": "千万同传",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            139
          ],
          "packet_id": 13,
          "engine": "mlx-whisper"
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "139",
          "before": "平时用的千万同传",
          "after": "千万同传",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "千万同传疑似某款同传软件误识",
          "context_evidence": "上下文为同传工具，'千万'可能为软件名",
          "independent_ASR": {
            "status": "REVIEW_ONLY_RESCUE_BUDGET"
          },
          "suppression_proposed": null
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "139",
          "before": "千万同传",
          "after": "千万同传",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "Nonsense: '千万同传' vs '千万同传' (millions simultaneous)",
          "context_evidence": "Context: '千万同传'",
          "independent_ASR": {
            "status": "REVIEW_ONLY_RESCUE_BUDGET"
          },
          "suppression_proposed": null
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "139",
          "before": "千万同传",
          "after": "千万同传",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "Nonsense: '千万同传' vs '千万同传' (millions simultaneous)",
          "context_evidence": "Context: '千万同传'",
          "independent_ASR": {
            "status": "REVIEW_ONLY_RESCUE_BUDGET"
          },
          "suppression_proposed": null
        }
      ],
      "language": "zh",
      "packet_ids": [
        13
      ],
      "classification": "BEST_SUPPORTED_WITH_REVIEW",
      "independent_ASR": {
        "status": "REVIEW_ONLY_RESCUE_BUDGET"
      },
      "raw_word_timestamps": [
        {
          "word": "平",
          "start": 337.46,
          "end": 337.96,
          "probability": 0.87060546875
        },
        {
          "word": "时",
          "start": 337.96,
          "end": 338.14,
          "probability": 0.99560546875
        },
        {
          "word": "用",
          "start": 338.14,
          "end": 338.28,
          "probability": 1.0
        },
        {
          "word": "的",
          "start": 338.28,
          "end": 338.4,
          "probability": 0.9638671875
        },
        {
          "word": "千",
          "start": 338.4,
          "end": 338.56,
          "probability": 0.98388671875
        },
        {
          "word": "万",
          "start": 338.56,
          "end": 338.73999999999995,
          "probability": 0.9951171875
        },
        {
          "word": "同",
          "start": 338.73999999999995,
          "end": 338.9,
          "probability": 0.94140625
        },
        {
          "word": "传",
          "start": 338.9,
          "end": 339.15999999999997,
          "probability": 0.9072265625
        }
      ],
      "word_timestamps": [],
      "source_text": "千万同传",
      "review_status": "BEST_SUPPORTED_WITH_REVIEW",
      "canonical_timing_evidence": "CUE_ANCHORED",
      "index": 139
    },
    {
      "start": 339.15999999999997,
      "end": 340.34,
      "text": "因为看的人太多",
      "raw_asr_text": "因为看的人太多",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 13,
      "words": [
        {
          "word": "因为",
          "start": 339.15999999999997,
          "end": 339.52,
          "probability": 0.99609375
        },
        {
          "word": "看",
          "start": 339.52,
          "end": 339.71999999999997,
          "probability": 0.99853515625
        },
        {
          "word": "的人",
          "start": 339.71999999999997,
          "end": 339.88,
          "probability": 0.998046875
        },
        {
          "word": "太",
          "start": 339.88,
          "end": 340.12,
          "probability": 1.0
        },
        {
          "word": "多",
          "start": 340.12,
          "end": 340.34,
          "probability": 0.99853515625
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.19097018241882324,
        "compression_ratio": 1.2310756972111554,
        "no_speech_prob": 1.463243415322779e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "摇完了功能最后想跟大家说说做这个项目的初衷其实特别简单前阵子我看火雨直播的时候平时用的千万同传因为看的人太多突然炸掉了那天半夜直播结束越想越上头的我决定干脆自己动手搓一个一开始真的只是打算留着自己用的但后来"
        }
      ],
      "id": 140,
      "review_annotation": "",
      "row_id": "140",
      "bounded_row_id": 140,
      "parent_row_ids": [
        140
      ],
      "parent_row_id": 140,
      "packet_id": 13,
      "LID_evidence": [
        {
          "region_id": 68,
          "probabilities": {
            "zh": 0.998153539498647
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        68
      ],
      "language_confidence": 0.998153539498647,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 327.4,
      "packet_end": 353.8,
      "raw_text": "因为看的人太多",
      "selected_text": "因为看的人太多",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            140
          ],
          "packet_id": 13,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        13
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 140
    },
    {
      "start": 340.34,
      "end": 341.78,
      "text": "突然炸掉了",
      "raw_asr_text": "突然炸掉了",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 13,
      "words": [
        {
          "word": "突",
          "start": 340.34,
          "end": 341.09999999999997,
          "probability": 0.99755859375
        },
        {
          "word": "然",
          "start": 341.09999999999997,
          "end": 341.21999999999997,
          "probability": 0.99951171875
        },
        {
          "word": "炸",
          "start": 341.21999999999997,
          "end": 341.4,
          "probability": 0.9912109375
        },
        {
          "word": "掉",
          "start": 341.4,
          "end": 341.58,
          "probability": 1.0
        },
        {
          "word": "了",
          "start": 341.58,
          "end": 341.78,
          "probability": 1.0
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.19097018241882324,
        "compression_ratio": 1.2310756972111554,
        "no_speech_prob": 1.463243415322779e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "摇完了功能最后想跟大家说说做这个项目的初衷其实特别简单前阵子我看火雨直播的时候平时用的千万同传因为看的人太多突然炸掉了那天半夜直播结束越想越上头的我决定干脆自己动手搓一个一开始真的只是打算留着自己用的但后来"
        }
      ],
      "id": 141,
      "review_annotation": "",
      "row_id": "141",
      "bounded_row_id": 141,
      "parent_row_ids": [
        141
      ],
      "parent_row_id": 141,
      "packet_id": 13,
      "LID_evidence": [
        {
          "region_id": 68,
          "probabilities": {
            "zh": 0.998153539498647
          },
          "status": "STABLE"
        },
        {
          "region_id": 69,
          "probabilities": {
            "zh": 0.9982088406880697
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        68,
        69
      ],
      "language_confidence": 0.9982088406880697,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 327.4,
      "packet_end": 353.8,
      "raw_text": "突然炸掉了",
      "selected_text": "突然炸掉了",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            141
          ],
          "packet_id": 13,
          "engine": "mlx-whisper"
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "141",
          "before": "突然炸掉了",
          "after": "突然炸掉了",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "VALID_SPOKEN_LANGUAGE",
          "reason": "炸掉了口语化表达，语义通顺",
          "context_evidence": "上下文为服务崩溃，'炸掉'为常见口语",
          "independent_ASR": {},
          "suppression_proposed": null
        }
      ],
      "language": "zh",
      "packet_ids": [
        13
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "source_text": "突然炸掉了",
      "review_status": "VALID_SPOKEN_LANGUAGE",
      "canonical_timing_evidence": "CUE_ANCHORED",
      "index": 141
    },
    {
      "start": 342.29999999999995,
      "end": 343.97999999999996,
      "text": "那天半夜直播结束",
      "raw_asr_text": "那天半夜直播结束",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 13,
      "words": [
        {
          "word": "那",
          "start": 342.29999999999995,
          "end": 342.7,
          "probability": 0.99951171875
        },
        {
          "word": "天",
          "start": 342.7,
          "end": 342.91999999999996,
          "probability": 0.99951171875
        },
        {
          "word": "半",
          "start": 342.91999999999996,
          "end": 343.09999999999997,
          "probability": 0.9990234375
        },
        {
          "word": "夜",
          "start": 343.09999999999997,
          "end": 343.23999999999995,
          "probability": 0.9990234375
        },
        {
          "word": "直",
          "start": 343.23999999999995,
          "end": 343.44,
          "probability": 1.0
        },
        {
          "word": "播",
          "start": 343.44,
          "end": 343.53999999999996,
          "probability": 1.0
        },
        {
          "word": "结",
          "start": 343.53999999999996,
          "end": 343.71999999999997,
          "probability": 0.99951171875
        },
        {
          "word": "束",
          "start": 343.71999999999997,
          "end": 343.97999999999996,
          "probability": 1.0
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.19097018241882324,
        "compression_ratio": 1.2310756972111554,
        "no_speech_prob": 1.463243415322779e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "摇完了功能最后想跟大家说说做这个项目的初衷其实特别简单前阵子我看火雨直播的时候平时用的千万同传因为看的人太多突然炸掉了那天半夜直播结束越想越上头的我决定干脆自己动手搓一个一开始真的只是打算留着自己用的但后来"
        }
      ],
      "id": 142,
      "review_annotation": "",
      "row_id": "142",
      "bounded_row_id": 142,
      "parent_row_ids": [
        142
      ],
      "parent_row_id": 142,
      "packet_id": 13,
      "LID_evidence": [
        {
          "region_id": 70,
          "probabilities": {
            "zh": 0.9979465723037719
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        70
      ],
      "language_confidence": 0.9979465723037719,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 327.4,
      "packet_end": 353.8,
      "raw_text": "那天半夜直播结束",
      "selected_text": "那天半夜直播结束",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            142
          ],
          "packet_id": 13,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        13
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 142
    },
    {
      "start": 343.97999999999996,
      "end": 345.58,
      "text": "越想越上头的我",
      "raw_asr_text": "越想越上头的我",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 13,
      "words": [
        {
          "word": "越",
          "start": 343.97999999999996,
          "end": 344.62,
          "probability": 0.99560546875
        },
        {
          "word": "想",
          "start": 344.62,
          "end": 344.82,
          "probability": 0.998046875
        },
        {
          "word": "越",
          "start": 344.82,
          "end": 344.96,
          "probability": 1.0
        },
        {
          "word": "上",
          "start": 344.96,
          "end": 345.12,
          "probability": 1.0
        },
        {
          "word": "头",
          "start": 345.12,
          "end": 345.28,
          "probability": 0.9970703125
        },
        {
          "word": "的",
          "start": 345.28,
          "end": 345.38,
          "probability": 1.0
        },
        {
          "word": "我",
          "start": 345.38,
          "end": 345.58,
          "probability": 1.0
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.19097018241882324,
        "compression_ratio": 1.2310756972111554,
        "no_speech_prob": 1.463243415322779e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "摇完了功能最后想跟大家说说做这个项目的初衷其实特别简单前阵子我看火雨直播的时候平时用的千万同传因为看的人太多突然炸掉了那天半夜直播结束越想越上头的我决定干脆自己动手搓一个一开始真的只是打算留着自己用的但后来"
        }
      ],
      "id": 143,
      "review_annotation": "",
      "row_id": "143",
      "bounded_row_id": 143,
      "parent_row_ids": [
        143
      ],
      "parent_row_id": 143,
      "packet_id": 13,
      "LID_evidence": [
        {
          "region_id": 70,
          "probabilities": {
            "zh": 0.9979465723037719
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        70
      ],
      "language_confidence": 0.9979465723037719,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 327.4,
      "packet_end": 353.8,
      "raw_text": "越想越上头的我",
      "selected_text": "越想越上头的我",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            143
          ],
          "packet_id": 13,
          "engine": "mlx-whisper"
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "143",
          "before": "越想越上头的我",
          "after": "越想越上头的我",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "VALID_SPOKEN_LANGUAGE",
          "reason": "上头口语化表达，语义通顺",
          "context_evidence": "上下文为情绪描述，'上头'为常见口语",
          "independent_ASR": {},
          "suppression_proposed": null
        }
      ],
      "language": "zh",
      "packet_ids": [
        13
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "source_text": "越想越上头的我",
      "review_status": "VALID_SPOKEN_LANGUAGE",
      "canonical_timing_evidence": "CUE_ANCHORED",
      "index": 143
    },
    {
      "start": 345.58,
      "end": 347.84,
      "text": "决定干脆自己动手搓一个",
      "raw_asr_text": "决定干脆自己动手搓一个",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 13,
      "words": [
        {
          "word": "决",
          "start": 345.58,
          "end": 346.06,
          "probability": 0.99951171875
        },
        {
          "word": "定",
          "start": 346.06,
          "end": 346.2,
          "probability": 1.0
        },
        {
          "word": "干",
          "start": 346.2,
          "end": 346.34,
          "probability": 0.99658203125
        },
        {
          "word": "脆",
          "start": 346.34,
          "end": 346.56,
          "probability": 0.99609375
        },
        {
          "word": "自己",
          "start": 346.56,
          "end": 347.12,
          "probability": 0.99951171875
        },
        {
          "word": "动",
          "start": 347.12,
          "end": 347.28,
          "probability": 0.99951171875
        },
        {
          "word": "手",
          "start": 347.28,
          "end": 347.44,
          "probability": 1.0
        },
        {
          "word": "搓",
          "start": 347.44,
          "end": 347.59999999999997,
          "probability": 0.92578125
        },
        {
          "word": "一个",
          "start": 347.59999999999997,
          "end": 347.84,
          "probability": 0.99951171875
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.19097018241882324,
        "compression_ratio": 1.2310756972111554,
        "no_speech_prob": 1.463243415322779e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "摇完了功能最后想跟大家说说做这个项目的初衷其实特别简单前阵子我看火雨直播的时候平时用的千万同传因为看的人太多突然炸掉了那天半夜直播结束越想越上头的我决定干脆自己动手搓一个一开始真的只是打算留着自己用的但后来"
        }
      ],
      "id": 144,
      "review_annotation": "",
      "row_id": "144",
      "bounded_row_id": 144,
      "parent_row_ids": [
        144
      ],
      "parent_row_id": 144,
      "packet_id": 13,
      "LID_evidence": [
        {
          "region_id": 70,
          "probabilities": {
            "zh": 0.9979465723037719
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        70
      ],
      "language_confidence": 0.9979465723037719,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 327.4,
      "packet_end": 353.8,
      "raw_text": "决定干脆自己动手搓一个",
      "selected_text": "决定干脆自己动手搓一个",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            144
          ],
          "packet_id": 13,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        13
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 144
    },
    {
      "start": 348.58,
      "end": 349.38,
      "text": "一开始",
      "raw_asr_text": "一开始",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 13,
      "words": [
        {
          "word": "一",
          "start": 348.58,
          "end": 348.97999999999996,
          "probability": 0.9990234375
        },
        {
          "word": "开始",
          "start": 348.97999999999996,
          "end": 349.38,
          "probability": 1.0
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.19097018241882324,
        "compression_ratio": 1.2310756972111554,
        "no_speech_prob": 1.463243415322779e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "摇完了功能最后想跟大家说说做这个项目的初衷其实特别简单前阵子我看火雨直播的时候平时用的千万同传因为看的人太多突然炸掉了那天半夜直播结束越想越上头的我决定干脆自己动手搓一个一开始真的只是打算留着自己用的但后来"
        }
      ],
      "id": 145,
      "review_annotation": "",
      "row_id": "145",
      "bounded_row_id": 145,
      "parent_row_ids": [
        145
      ],
      "parent_row_id": 145,
      "packet_id": 13,
      "LID_evidence": [
        {
          "region_id": 71,
          "probabilities": {
            "zh": 0.9972564663205828
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        71
      ],
      "language_confidence": 0.9972564663205828,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 327.4,
      "packet_end": 353.8,
      "raw_text": "一开始",
      "selected_text": "一开始",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            145
          ],
          "packet_id": 13,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        13
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 145
    },
    {
      "start": 350.05999999999995,
      "end": 352.71999999999997,
      "text": "真的只是打算留着自己用的",
      "raw_asr_text": "真的只是打算留着自己用的",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 13,
      "words": [
        {
          "word": "真的",
          "start": 350.05999999999995,
          "end": 350.46,
          "probability": 0.998046875
        },
        {
          "word": "只是",
          "start": 350.46,
          "end": 350.73999999999995,
          "probability": 0.99462890625
        },
        {
          "word": "打",
          "start": 350.73999999999995,
          "end": 350.91999999999996,
          "probability": 1.0
        },
        {
          "word": "算",
          "start": 350.91999999999996,
          "end": 351.15999999999997,
          "probability": 1.0
        },
        {
          "word": "留",
          "start": 351.15999999999997,
          "end": 351.73999999999995,
          "probability": 0.998046875
        },
        {
          "word": "着",
          "start": 351.73999999999995,
          "end": 351.9,
          "probability": 0.99853515625
        },
        {
          "word": "自己",
          "start": 351.9,
          "end": 352.21999999999997,
          "probability": 1.0
        },
        {
          "word": "用",
          "start": 352.21999999999997,
          "end": 352.5,
          "probability": 1.0
        },
        {
          "word": "的",
          "start": 352.5,
          "end": 352.71999999999997,
          "probability": 1.0
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.19097018241882324,
        "compression_ratio": 1.2310756972111554,
        "no_speech_prob": 1.463243415322779e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "摇完了功能最后想跟大家说说做这个项目的初衷其实特别简单前阵子我看火雨直播的时候平时用的千万同传因为看的人太多突然炸掉了那天半夜直播结束越想越上头的我决定干脆自己动手搓一个一开始真的只是打算留着自己用的但后来"
        }
      ],
      "id": 146,
      "review_annotation": "",
      "row_id": "146",
      "bounded_row_id": 146,
      "parent_row_ids": [
        146
      ],
      "parent_row_id": 146,
      "packet_id": 13,
      "LID_evidence": [
        {
          "region_id": 72,
          "probabilities": {
            "zh": 0.9761556148529053
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        72
      ],
      "language_confidence": 0.9761556148529053,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 327.4,
      "packet_end": 353.8,
      "raw_text": "真的只是打算留着自己用的",
      "selected_text": "真的只是打算留着自己用的",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            146
          ],
          "packet_id": 13,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        13
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 146
    },
    {
      "start": 353.4,
      "end": 357.23999999999995,
      "text": "直播镜里",
      "raw_asr_text": "但后来在直播镜里感受到大家都面临着一样的问题",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 14,
      "words": [],
      "asr_metrics": {
        "avg_logprob": -0.11080598831176758,
        "compression_ratio": 1.1842105263157894,
        "no_speech_prob": 1.5514787371495586e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "但后来在直播镜里感受到大家都面临着一样的问题看到不少人被迫去买付费翻译软件一边用还一边收起抱怨我就想既然自己用着看直播挺好的不如分享出来给大家一起用吧"
        }
      ],
      "id": 147,
      "review_annotation": "待確認",
      "row_id": "147",
      "bounded_row_id": 147,
      "parent_row_ids": [
        147
      ],
      "parent_row_id": 147,
      "packet_id": 14,
      "LID_evidence": [
        {
          "region_id": 73,
          "probabilities": {
            "zh": 0.9871415562099881
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        73
      ],
      "language_confidence": 0.9871415562099881,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 352.59999999999997,
      "packet_end": 371.3,
      "raw_text": "但后来在直播镜里感受到大家都面临着一样的问题",
      "selected_text": "直播镜里",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            147
          ],
          "packet_id": 14,
          "engine": "mlx-whisper"
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "147",
          "before": "但后来在直播镜里感受到大家都面临着一样的问题",
          "after": "直播镜里",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "直播镜里疑似直播间里误识",
          "context_evidence": "上下文为直播场景，'直播间'更符合逻辑",
          "independent_ASR": {
            "status": "REVIEW_ONLY_RESCUE_BUDGET"
          },
          "suppression_proposed": null
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "147",
          "before": "直播镜里",
          "after": "直播镜里",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "Nonsense: '直播镜里' vs '直播镜头' (live stream camera)",
          "context_evidence": "Context: '直播镜头'",
          "independent_ASR": {
            "status": "REVIEW_ONLY_RESCUE_BUDGET"
          },
          "suppression_proposed": null
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "147",
          "before": "直播镜里",
          "after": "直播镜里",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "Nonsense: '直播镜里' vs '直播镜头' (live stream camera)",
          "context_evidence": "Context: '直播镜头'",
          "independent_ASR": {
            "status": "REVIEW_ONLY_RESCUE_BUDGET"
          },
          "suppression_proposed": null
        }
      ],
      "language": "zh",
      "packet_ids": [
        14
      ],
      "classification": "BEST_SUPPORTED_WITH_REVIEW",
      "independent_ASR": {
        "status": "REVIEW_ONLY_RESCUE_BUDGET"
      },
      "raw_word_timestamps": [
        {
          "word": "但",
          "start": 353.23999999999995,
          "end": 353.59999999999997,
          "probability": 0.8837890625
        },
        {
          "word": "后",
          "start": 353.59999999999997,
          "end": 353.73999999999995,
          "probability": 0.9990234375
        },
        {
          "word": "来",
          "start": 353.73999999999995,
          "end": 353.96,
          "probability": 1.0
        },
        {
          "word": "在",
          "start": 353.96,
          "end": 354.29999999999995,
          "probability": 0.953125
        },
        {
          "word": "直",
          "start": 354.29999999999995,
          "end": 354.49999999999994,
          "probability": 0.99951171875
        },
        {
          "word": "播",
          "start": 354.49999999999994,
          "end": 354.64,
          "probability": 1.0
        },
        {
          "word": "镜",
          "start": 354.64,
          "end": 354.84,
          "probability": 0.7353515625
        },
        {
          "word": "里",
          "start": 354.84,
          "end": 354.99999999999994,
          "probability": 0.998046875
        },
        {
          "word": "感",
          "start": 354.99999999999994,
          "end": 355.34,
          "probability": 0.99609375
        },
        {
          "word": "受",
          "start": 355.34,
          "end": 355.49999999999994,
          "probability": 0.9990234375
        },
        {
          "word": "到",
          "start": 355.49999999999994,
          "end": 355.61999999999995,
          "probability": 1.0
        },
        {
          "word": "大家都",
          "start": 355.61999999999995,
          "end": 355.96,
          "probability": 0.9970703125
        },
        {
          "word": "面",
          "start": 355.96,
          "end": 356.23999999999995,
          "probability": 1.0
        },
        {
          "word": "临",
          "start": 356.23999999999995,
          "end": 356.4,
          "probability": 1.0
        },
        {
          "word": "着",
          "start": 356.4,
          "end": 356.52,
          "probability": 0.99951171875
        },
        {
          "word": "一",
          "start": 356.52,
          "end": 356.71999999999997,
          "probability": 0.99951171875
        },
        {
          "word": "样",
          "start": 356.71999999999997,
          "end": 356.79999999999995,
          "probability": 1.0
        },
        {
          "word": "的",
          "start": 356.79999999999995,
          "end": 356.91999999999996,
          "probability": 1.0
        },
        {
          "word": "问题",
          "start": 356.91999999999996,
          "end": 357.23999999999995,
          "probability": 1.0
        }
      ],
      "word_timestamps": [],
      "source_text": "直播镜里",
      "review_status": "BEST_SUPPORTED_WITH_REVIEW",
      "canonical_timing_evidence": "CUE_ANCHORED",
      "index": 147
    },
    {
      "start": 357.23999999999995,
      "end": 360.17999999999995,
      "text": "看到不少人被迫去买付费翻译软件",
      "raw_asr_text": "看到不少人被迫去买付费翻译软件",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 14,
      "words": [
        {
          "word": "看到",
          "start": 357.23999999999995,
          "end": 357.84,
          "probability": 0.962890625
        },
        {
          "word": "不",
          "start": 357.84,
          "end": 358.05999999999995,
          "probability": 1.0
        },
        {
          "word": "少",
          "start": 358.05999999999995,
          "end": 358.21999999999997,
          "probability": 0.99951171875
        },
        {
          "word": "人",
          "start": 358.21999999999997,
          "end": 358.37999999999994,
          "probability": 1.0
        },
        {
          "word": "被",
          "start": 358.37999999999994,
          "end": 358.73999999999995,
          "probability": 0.99951171875
        },
        {
          "word": "迫",
          "start": 358.73999999999995,
          "end": 358.9,
          "probability": 0.99951171875
        },
        {
          "word": "去",
          "start": 358.9,
          "end": 359.03999999999996,
          "probability": 0.99951171875
        },
        {
          "word": "买",
          "start": 359.03999999999996,
          "end": 359.2,
          "probability": 1.0
        },
        {
          "word": "付",
          "start": 359.2,
          "end": 359.34,
          "probability": 0.69677734375
        },
        {
          "word": "费",
          "start": 359.34,
          "end": 359.52,
          "probability": 0.802734375
        },
        {
          "word": "翻",
          "start": 359.52,
          "end": 359.65999999999997,
          "probability": 0.9677734375
        },
        {
          "word": "译",
          "start": 359.65999999999997,
          "end": 359.79999999999995,
          "probability": 0.990234375
        },
        {
          "word": "软",
          "start": 359.79999999999995,
          "end": 359.97999999999996,
          "probability": 0.9990234375
        },
        {
          "word": "件",
          "start": 359.97999999999996,
          "end": 360.17999999999995,
          "probability": 0.99951171875
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.11080598831176758,
        "compression_ratio": 1.1842105263157894,
        "no_speech_prob": 1.5514787371495586e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "但后来在直播镜里感受到大家都面临着一样的问题看到不少人被迫去买付费翻译软件一边用还一边收起抱怨我就想既然自己用着看直播挺好的不如分享出来给大家一起用吧"
        }
      ],
      "id": 148,
      "review_annotation": "",
      "row_id": "148",
      "bounded_row_id": 148,
      "parent_row_ids": [
        148
      ],
      "parent_row_id": 148,
      "packet_id": 14,
      "LID_evidence": [
        {
          "region_id": 73,
          "probabilities": {
            "zh": 0.9871415562099881
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        73
      ],
      "language_confidence": 0.9871415562099881,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 352.59999999999997,
      "packet_end": 371.3,
      "raw_text": "看到不少人被迫去买付费翻译软件",
      "selected_text": "看到不少人被迫去买付费翻译软件",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            148
          ],
          "packet_id": 14,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        14
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 148
    },
    {
      "start": 360.17999999999995,
      "end": 362.52,
      "text": "收起抱怨",
      "raw_asr_text": "一边用还一边收起抱怨",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 14,
      "words": [],
      "asr_metrics": {
        "avg_logprob": -0.11080598831176758,
        "compression_ratio": 1.1842105263157894,
        "no_speech_prob": 1.5514787371495586e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "但后来在直播镜里感受到大家都面临着一样的问题看到不少人被迫去买付费翻译软件一边用还一边收起抱怨我就想既然自己用着看直播挺好的不如分享出来给大家一起用吧"
        }
      ],
      "id": 149,
      "review_annotation": "待確認",
      "row_id": "149",
      "bounded_row_id": 149,
      "parent_row_ids": [
        149
      ],
      "parent_row_id": 149,
      "packet_id": 14,
      "LID_evidence": [
        {
          "region_id": 73,
          "probabilities": {
            "zh": 0.9871415562099881
          },
          "status": "STABLE"
        },
        {
          "region_id": 74,
          "probabilities": {
            "zh": 0.9980296691258749
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        73,
        74
      ],
      "language_confidence": 0.9980296691258749,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 352.59999999999997,
      "packet_end": 371.3,
      "raw_text": "一边用还一边收起抱怨",
      "selected_text": "收起抱怨",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            149
          ],
          "packet_id": 14,
          "engine": "mlx-whisper"
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "149",
          "before": "一边用还一边收起抱怨",
          "after": "收起抱怨",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "收起疑似忍受或发出误识",
          "context_evidence": "上下文为付费软件体验，'忍受'或'发出'更符合",
          "independent_ASR": {
            "status": "REVIEW_ONLY_RESCUE_BUDGET"
          },
          "suppression_proposed": null
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "149",
          "before": "收起抱怨",
          "after": "收起抱怨",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "Nonsense: '收起抱怨' vs '看到抱怨' (saw complaints)",
          "context_evidence": "Context: '看到不少人被迫去买付费翻译软件'",
          "independent_ASR": {
            "status": "REVIEW_ONLY_RESCUE_BUDGET"
          },
          "suppression_proposed": null
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "149",
          "before": "收起抱怨",
          "after": "收起抱怨",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "Nonsense: '收起抱怨' vs '看到抱怨' (saw complaints)",
          "context_evidence": "Context: '看到不少人被迫去买付费翻译软件'",
          "independent_ASR": {
            "status": "REVIEW_ONLY_RESCUE_BUDGET"
          },
          "suppression_proposed": null
        }
      ],
      "language": "zh",
      "packet_ids": [
        14
      ],
      "classification": "BEST_SUPPORTED_WITH_REVIEW",
      "independent_ASR": {
        "status": "REVIEW_ONLY_RESCUE_BUDGET"
      },
      "raw_word_timestamps": [
        {
          "word": "一",
          "start": 360.17999999999995,
          "end": 360.85999999999996,
          "probability": 0.9921875
        },
        {
          "word": "边",
          "start": 360.85999999999996,
          "end": 361.02,
          "probability": 0.99853515625
        },
        {
          "word": "用",
          "start": 361.02,
          "end": 361.15999999999997,
          "probability": 0.8408203125
        },
        {
          "word": "还",
          "start": 361.15999999999997,
          "end": 361.35999999999996,
          "probability": 0.71142578125
        },
        {
          "word": "一",
          "start": 361.35999999999996,
          "end": 361.71999999999997,
          "probability": 0.98095703125
        },
        {
          "word": "边",
          "start": 361.71999999999997,
          "end": 361.82,
          "probability": 0.99951171875
        },
        {
          "word": "收",
          "start": 361.82,
          "end": 361.97999999999996,
          "probability": 0.76220703125
        },
        {
          "word": "起",
          "start": 361.97999999999996,
          "end": 362.11999999999995,
          "probability": 0.73486328125
        },
        {
          "word": "抱",
          "start": 362.11999999999995,
          "end": 362.34,
          "probability": 0.96142578125
        },
        {
          "word": "怨",
          "start": 362.34,
          "end": 362.52,
          "probability": 0.9951171875
        }
      ],
      "word_timestamps": [],
      "source_text": "收起抱怨",
      "review_status": "BEST_SUPPORTED_WITH_REVIEW",
      "canonical_timing_evidence": "CUE_ANCHORED",
      "index": 149
    },
    {
      "start": 365.17999999999995,
      "end": 367.76,
      "text": "我就想既然自己用着看直播挺好的",
      "raw_asr_text": "我就想既然自己用着看直播挺好的",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 14,
      "words": [
        {
          "word": "我就",
          "start": 365.17999999999995,
          "end": 365.53999999999996,
          "probability": 0.99609375
        },
        {
          "word": "想",
          "start": 365.53999999999996,
          "end": 365.82,
          "probability": 0.99951171875
        },
        {
          "word": "既",
          "start": 365.82,
          "end": 366.23999999999995,
          "probability": 0.9765625
        },
        {
          "word": "然",
          "start": 366.23999999999995,
          "end": 366.32,
          "probability": 1.0
        },
        {
          "word": "自己",
          "start": 366.32,
          "end": 366.58,
          "probability": 1.0
        },
        {
          "word": "用",
          "start": 366.58,
          "end": 366.76,
          "probability": 0.9990234375
        },
        {
          "word": "着",
          "start": 366.76,
          "end": 366.85999999999996,
          "probability": 0.99853515625
        },
        {
          "word": "看",
          "start": 366.85999999999996,
          "end": 366.99999999999994,
          "probability": 0.9990234375
        },
        {
          "word": "直",
          "start": 366.99999999999994,
          "end": 367.14,
          "probability": 1.0
        },
        {
          "word": "播",
          "start": 367.14,
          "end": 367.26,
          "probability": 1.0
        },
        {
          "word": "挺",
          "start": 367.26,
          "end": 367.41999999999996,
          "probability": 0.99853515625
        },
        {
          "word": "好的",
          "start": 367.41999999999996,
          "end": 367.76,
          "probability": 1.0
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.11080598831176758,
        "compression_ratio": 1.1842105263157894,
        "no_speech_prob": 1.5514787371495586e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "但后来在直播镜里感受到大家都面临着一样的问题看到不少人被迫去买付费翻译软件一边用还一边收起抱怨我就想既然自己用着看直播挺好的不如分享出来给大家一起用吧"
        }
      ],
      "id": 150,
      "review_annotation": "",
      "row_id": "150",
      "bounded_row_id": 150,
      "parent_row_ids": [
        150
      ],
      "parent_row_id": 150,
      "packet_id": 14,
      "LID_evidence": [
        {
          "region_id": 75,
          "probabilities": {
            "zh": 0.9595849243077365
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        75
      ],
      "language_confidence": 0.9595849243077365,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 352.59999999999997,
      "packet_end": 371.3,
      "raw_text": "我就想既然自己用着看直播挺好的",
      "selected_text": "我就想既然自己用着看直播挺好的",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            150
          ],
          "packet_id": 14,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        14
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 150
    },
    {
      "start": 367.76,
      "end": 370.15999999999997,
      "text": "不如分享出来给大家一起用吧",
      "raw_asr_text": "不如分享出来给大家一起用吧",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 14,
      "words": [
        {
          "word": "不",
          "start": 367.76,
          "end": 368.23999999999995,
          "probability": 0.9970703125
        },
        {
          "word": "如",
          "start": 368.23999999999995,
          "end": 368.34,
          "probability": 0.99951171875
        },
        {
          "word": "分享",
          "start": 368.34,
          "end": 368.61999999999995,
          "probability": 1.0
        },
        {
          "word": "出来",
          "start": 368.61999999999995,
          "end": 368.91999999999996,
          "probability": 0.9990234375
        },
        {
          "word": "给",
          "start": 368.91999999999996,
          "end": 369.23999999999995,
          "probability": 0.99853515625
        },
        {
          "word": "大家",
          "start": 369.23999999999995,
          "end": 369.47999999999996,
          "probability": 1.0
        },
        {
          "word": "一起",
          "start": 369.47999999999996,
          "end": 369.78,
          "probability": 1.0
        },
        {
          "word": "用",
          "start": 369.78,
          "end": 369.97999999999996,
          "probability": 1.0
        },
        {
          "word": "吧",
          "start": 369.97999999999996,
          "end": 370.15999999999997,
          "probability": 1.0
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.11080598831176758,
        "compression_ratio": 1.1842105263157894,
        "no_speech_prob": 1.5514787371495586e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "但后来在直播镜里感受到大家都面临着一样的问题看到不少人被迫去买付费翻译软件一边用还一边收起抱怨我就想既然自己用着看直播挺好的不如分享出来给大家一起用吧"
        }
      ],
      "id": 151,
      "review_annotation": "",
      "row_id": "151",
      "bounded_row_id": 151,
      "parent_row_ids": [
        151
      ],
      "parent_row_id": 151,
      "packet_id": 14,
      "LID_evidence": [
        {
          "region_id": 75,
          "probabilities": {
            "zh": 0.9595849243077365
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        75
      ],
      "language_confidence": 0.9595849243077365,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 352.59999999999997,
      "packet_end": 371.3,
      "raw_text": "不如分享出来给大家一起用吧",
      "selected_text": "不如分享出来给大家一起用吧",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            151
          ],
          "packet_id": 14,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        14
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 151
    },
    {
      "start": 371.1,
      "end": 375.34000000000003,
      "text": "两个切片和一个卢播",
      "raw_asr_text": "最开始我只是在两个切片和一个卢播的评论区里发了一下",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 15,
      "words": [],
      "asr_metrics": {
        "avg_logprob": -0.10487588451754662,
        "compression_ratio": 1.3598615916955017,
        "no_speech_prob": 1.723270484643713e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "最开始我只是在两个切片和一个卢播的评论区里发了一下后来又做了个特别简陋现在回头看甚至有点垃圾的小视频哈哈即使当时只是想有个能放下载地址接受大家反馈的地方没想到一下子收到了各种各样的反馈和功能诉求这几天我几乎把所有能想到的功能开发和三个平台的适配全都干出来了还有了"
        }
      ],
      "id": 152,
      "review_annotation": "待確認",
      "row_id": "152",
      "bounded_row_id": 152,
      "parent_row_ids": [
        152
      ],
      "parent_row_id": 152,
      "packet_id": 15,
      "LID_evidence": [
        {
          "region_id": 76,
          "probabilities": {
            "zh": 0.9834216052088244
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        76
      ],
      "language_confidence": 0.9834216052088244,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 370.3,
      "packet_end": 398.90000000000003,
      "raw_text": "最开始我只是在两个切片和一个卢播的评论区里发了一下",
      "selected_text": "两个切片和一个卢播",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            152
          ],
          "packet_id": 15,
          "engine": "mlx-whisper"
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "152",
          "before": "最开始我只是在两个切片和一个卢播的评论区里发了一下",
          "after": "两个切片和一个卢播",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "切片和卢播疑似特定平台或UP主误识",
          "context_evidence": "上下文为发布渠道，'切片'和'卢播'可能为专有名词",
          "independent_ASR": {
            "status": "REVIEW_ONLY_RESCUE_BUDGET"
          },
          "suppression_proposed": null
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "152",
          "before": "两个切片和一个卢播",
          "after": "两个切片和一个卢播",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "Nonsense: '卢播' vs '直播' (live stream)",
          "context_evidence": "Context: '两个切片和一个直播'",
          "independent_ASR": {
            "status": "REVIEW_ONLY_RESCUE_BUDGET"
          },
          "suppression_proposed": null
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "152",
          "before": "两个切片和一个卢播",
          "after": "两个切片和一个卢播",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "Nonsense: '卢播' vs '直播' (live stream)",
          "context_evidence": "Context: '两个切片和一个直播'",
          "independent_ASR": {
            "status": "REVIEW_ONLY_RESCUE_BUDGET"
          },
          "suppression_proposed": null
        }
      ],
      "language": "zh",
      "packet_ids": [
        15
      ],
      "classification": "BEST_SUPPORTED_WITH_REVIEW",
      "independent_ASR": {
        "status": "REVIEW_ONLY_RESCUE_BUDGET"
      },
      "raw_word_timestamps": [
        {
          "word": "最",
          "start": 371.08,
          "end": 371.40000000000003,
          "probability": 0.99267578125
        },
        {
          "word": "开始",
          "start": 371.40000000000003,
          "end": 371.72,
          "probability": 0.99951171875
        },
        {
          "word": "我",
          "start": 371.72,
          "end": 372.2,
          "probability": 0.9638671875
        },
        {
          "word": "只是",
          "start": 372.2,
          "end": 372.40000000000003,
          "probability": 1.0
        },
        {
          "word": "在",
          "start": 372.40000000000003,
          "end": 372.58,
          "probability": 1.0
        },
        {
          "word": "两",
          "start": 372.58,
          "end": 372.72,
          "probability": 1.0
        },
        {
          "word": "个",
          "start": 372.72,
          "end": 372.86,
          "probability": 1.0
        },
        {
          "word": "切",
          "start": 372.86,
          "end": 373.04,
          "probability": 0.98486328125
        },
        {
          "word": "片",
          "start": 373.04,
          "end": 373.22,
          "probability": 0.99951171875
        },
        {
          "word": "和",
          "start": 373.22,
          "end": 373.54,
          "probability": 0.99951171875
        },
        {
          "word": "一个",
          "start": 373.54,
          "end": 373.76,
          "probability": 0.99951171875
        },
        {
          "word": "卢",
          "start": 373.76,
          "end": 373.92,
          "probability": 0.91796875
        },
        {
          "word": "播",
          "start": 373.92,
          "end": 374.04,
          "probability": 0.59765625
        },
        {
          "word": "的",
          "start": 374.04,
          "end": 374.18,
          "probability": 0.99951171875
        },
        {
          "word": "评",
          "start": 374.18,
          "end": 374.34000000000003,
          "probability": 0.99951171875
        },
        {
          "word": "论",
          "start": 374.34000000000003,
          "end": 374.5,
          "probability": 1.0
        },
        {
          "word": "区",
          "start": 374.5,
          "end": 374.66,
          "probability": 0.9990234375
        },
        {
          "word": "里",
          "start": 374.66,
          "end": 374.76,
          "probability": 0.99951171875
        },
        {
          "word": "发",
          "start": 374.76,
          "end": 374.94,
          "probability": 0.99951171875
        },
        {
          "word": "了",
          "start": 374.94,
          "end": 375.1,
          "probability": 1.0
        },
        {
          "word": "一下",
          "start": 375.1,
          "end": 375.34000000000003,
          "probability": 1.0
        }
      ],
      "word_timestamps": [],
      "source_text": "两个切片和一个卢播",
      "review_status": "BEST_SUPPORTED_WITH_REVIEW",
      "canonical_timing_evidence": "CUE_ANCHORED",
      "index": 152
    },
    {
      "start": 375.34000000000003,
      "end": 377.2,
      "text": "后来又做了个特别简陋",
      "raw_asr_text": "后来又做了个特别简陋",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 15,
      "words": [
        {
          "word": "后",
          "start": 375.34000000000003,
          "end": 375.92,
          "probability": 0.9951171875
        },
        {
          "word": "来",
          "start": 375.92,
          "end": 376.06,
          "probability": 0.99951171875
        },
        {
          "word": "又",
          "start": 376.06,
          "end": 376.18,
          "probability": 0.98486328125
        },
        {
          "word": "做",
          "start": 376.18,
          "end": 376.32,
          "probability": 0.9990234375
        },
        {
          "word": "了",
          "start": 376.32,
          "end": 376.40000000000003,
          "probability": 1.0
        },
        {
          "word": "个",
          "start": 376.40000000000003,
          "end": 376.5,
          "probability": 0.99755859375
        },
        {
          "word": "特",
          "start": 376.5,
          "end": 376.68,
          "probability": 0.99951171875
        },
        {
          "word": "别",
          "start": 376.68,
          "end": 376.86,
          "probability": 1.0
        },
        {
          "word": "简",
          "start": 376.86,
          "end": 377.02000000000004,
          "probability": 0.9990234375
        },
        {
          "word": "陋",
          "start": 377.02000000000004,
          "end": 377.2,
          "probability": 0.98828125
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.10487588451754662,
        "compression_ratio": 1.3598615916955017,
        "no_speech_prob": 1.723270484643713e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "最开始我只是在两个切片和一个卢播的评论区里发了一下后来又做了个特别简陋现在回头看甚至有点垃圾的小视频哈哈即使当时只是想有个能放下载地址接受大家反馈的地方没想到一下子收到了各种各样的反馈和功能诉求这几天我几乎把所有能想到的功能开发和三个平台的适配全都干出来了还有了"
        }
      ],
      "id": 153,
      "review_annotation": "",
      "row_id": "153",
      "bounded_row_id": 153,
      "parent_row_ids": [
        153
      ],
      "parent_row_id": 153,
      "packet_id": 15,
      "LID_evidence": [
        {
          "region_id": 76,
          "probabilities": {
            "zh": 0.9834216052088244
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        76
      ],
      "language_confidence": 0.9834216052088244,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 370.3,
      "packet_end": 398.90000000000003,
      "raw_text": "后来又做了个特别简陋",
      "selected_text": "后来又做了个特别简陋",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            153
          ],
          "packet_id": 15,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        15
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 153
    },
    {
      "start": 377.2,
      "end": 380.18,
      "text": "垃圾的小视频",
      "raw_asr_text": "现在回头看甚至有点垃圾的小视频",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 15,
      "words": [],
      "asr_metrics": {
        "avg_logprob": -0.10487588451754662,
        "compression_ratio": 1.3598615916955017,
        "no_speech_prob": 1.723270484643713e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "最开始我只是在两个切片和一个卢播的评论区里发了一下后来又做了个特别简陋现在回头看甚至有点垃圾的小视频哈哈即使当时只是想有个能放下载地址接受大家反馈的地方没想到一下子收到了各种各样的反馈和功能诉求这几天我几乎把所有能想到的功能开发和三个平台的适配全都干出来了还有了"
        }
      ],
      "id": 154,
      "review_annotation": "",
      "row_id": "154",
      "bounded_row_id": 154,
      "parent_row_ids": [
        154
      ],
      "parent_row_id": 154,
      "packet_id": 15,
      "LID_evidence": [
        {
          "region_id": 76,
          "probabilities": {
            "zh": 0.9834216052088244
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        76
      ],
      "language_confidence": 0.9834216052088244,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 370.3,
      "packet_end": 398.90000000000003,
      "raw_text": "现在回头看甚至有点垃圾的小视频",
      "selected_text": "垃圾的小视频",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            154
          ],
          "packet_id": 15,
          "engine": "mlx-whisper"
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "154",
          "before": "现在回头看甚至有点垃圾的小视频",
          "after": "垃圾的小视频",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "VALID_SPOKEN_LANGUAGE",
          "reason": "垃圾口语化表达，语义通顺",
          "context_evidence": "上下文为自嘲，'垃圾'为常见口语",
          "independent_ASR": {},
          "suppression_proposed": null
        }
      ],
      "language": "zh",
      "packet_ids": [
        15
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "raw_word_timestamps": [
        {
          "word": "现在",
          "start": 377.2,
          "end": 377.82,
          "probability": 0.99755859375
        },
        {
          "word": "回",
          "start": 377.82,
          "end": 378.04,
          "probability": 1.0
        },
        {
          "word": "头",
          "start": 378.04,
          "end": 378.18,
          "probability": 1.0
        },
        {
          "word": "看",
          "start": 378.18,
          "end": 378.38,
          "probability": 1.0
        },
        {
          "word": "甚",
          "start": 378.38,
          "end": 378.8,
          "probability": 0.99169921875
        },
        {
          "word": "至",
          "start": 378.8,
          "end": 378.92,
          "probability": 1.0
        },
        {
          "word": "有",
          "start": 378.92,
          "end": 379.06,
          "probability": 0.99169921875
        },
        {
          "word": "点",
          "start": 379.06,
          "end": 379.16,
          "probability": 1.0
        },
        {
          "word": "垃",
          "start": 379.16,
          "end": 379.36,
          "probability": 0.9951171875
        },
        {
          "word": "圾",
          "start": 379.36,
          "end": 379.46000000000004,
          "probability": 0.99853515625
        },
        {
          "word": "的",
          "start": 379.46000000000004,
          "end": 379.58,
          "probability": 1.0
        },
        {
          "word": "小",
          "start": 379.58,
          "end": 379.74,
          "probability": 1.0
        },
        {
          "word": "视",
          "start": 379.74,
          "end": 379.90000000000003,
          "probability": 1.0
        },
        {
          "word": "频",
          "start": 379.90000000000003,
          "end": 380.18,
          "probability": 1.0
        }
      ],
      "word_timestamps": [],
      "source_text": "垃圾的小视频",
      "review_status": "VALID_SPOKEN_LANGUAGE",
      "canonical_timing_evidence": "CUE_ANCHORED",
      "index": 154
    },
    {
      "start": 380.72,
      "end": 383.40000000000003,
      "text": "哈哈即使当时只是想有个能放下载地址",
      "raw_asr_text": "哈哈即使当时只是想有个能放下载地址",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 15,
      "words": [
        {
          "word": "哈哈",
          "start": 380.72,
          "end": 381.04,
          "probability": 0.54248046875
        },
        {
          "word": "即",
          "start": 381.04,
          "end": 381.36,
          "probability": 0.74267578125
        },
        {
          "word": "使",
          "start": 381.36,
          "end": 381.48,
          "probability": 0.9755859375
        },
        {
          "word": "当",
          "start": 381.48,
          "end": 381.64,
          "probability": 1.0
        },
        {
          "word": "时",
          "start": 381.64,
          "end": 381.78000000000003,
          "probability": 0.99951171875
        },
        {
          "word": "只是",
          "start": 381.78000000000003,
          "end": 381.94,
          "probability": 0.99853515625
        },
        {
          "word": "想",
          "start": 381.94,
          "end": 382.16,
          "probability": 0.99462890625
        },
        {
          "word": "有",
          "start": 382.16,
          "end": 382.3,
          "probability": 0.99658203125
        },
        {
          "word": "个",
          "start": 382.3,
          "end": 382.42,
          "probability": 0.994140625
        },
        {
          "word": "能",
          "start": 382.42,
          "end": 382.52000000000004,
          "probability": 0.998046875
        },
        {
          "word": "放",
          "start": 382.52000000000004,
          "end": 382.74,
          "probability": 0.99951171875
        },
        {
          "word": "下",
          "start": 382.74,
          "end": 382.90000000000003,
          "probability": 0.99853515625
        },
        {
          "word": "载",
          "start": 382.90000000000003,
          "end": 383.02000000000004,
          "probability": 0.98046875
        },
        {
          "word": "地",
          "start": 383.02000000000004,
          "end": 383.2,
          "probability": 0.9990234375
        },
        {
          "word": "址",
          "start": 383.2,
          "end": 383.40000000000003,
          "probability": 0.9970703125
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.10487588451754662,
        "compression_ratio": 1.3598615916955017,
        "no_speech_prob": 1.723270484643713e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "最开始我只是在两个切片和一个卢播的评论区里发了一下后来又做了个特别简陋现在回头看甚至有点垃圾的小视频哈哈即使当时只是想有个能放下载地址接受大家反馈的地方没想到一下子收到了各种各样的反馈和功能诉求这几天我几乎把所有能想到的功能开发和三个平台的适配全都干出来了还有了"
        }
      ],
      "id": 155,
      "review_annotation": "",
      "row_id": "155",
      "bounded_row_id": 155,
      "parent_row_ids": [
        155
      ],
      "parent_row_id": 155,
      "packet_id": 15,
      "LID_evidence": [
        {
          "region_id": 76,
          "probabilities": {
            "zh": 0.9834216052088244
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        76
      ],
      "language_confidence": 0.9834216052088244,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 370.3,
      "packet_end": 398.90000000000003,
      "raw_text": "哈哈即使当时只是想有个能放下载地址",
      "selected_text": "哈哈即使当时只是想有个能放下载地址",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            155
          ],
          "packet_id": 15,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        15
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 155
    },
    {
      "start": 383.40000000000003,
      "end": 385.32,
      "text": "接受大家反馈的地方",
      "raw_asr_text": "接受大家反馈的地方",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 15,
      "words": [
        {
          "word": "接",
          "start": 383.40000000000003,
          "end": 383.98,
          "probability": 0.99560546875
        },
        {
          "word": "受",
          "start": 383.98,
          "end": 384.16,
          "probability": 0.80712890625
        },
        {
          "word": "大家",
          "start": 384.16,
          "end": 384.42,
          "probability": 1.0
        },
        {
          "word": "反",
          "start": 384.42,
          "end": 384.66,
          "probability": 0.9990234375
        },
        {
          "word": "馈",
          "start": 384.66,
          "end": 384.90000000000003,
          "probability": 0.9990234375
        },
        {
          "word": "的",
          "start": 384.90000000000003,
          "end": 385.02000000000004,
          "probability": 1.0
        },
        {
          "word": "地方",
          "start": 385.02000000000004,
          "end": 385.32,
          "probability": 1.0
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.10487588451754662,
        "compression_ratio": 1.3598615916955017,
        "no_speech_prob": 1.723270484643713e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "最开始我只是在两个切片和一个卢播的评论区里发了一下后来又做了个特别简陋现在回头看甚至有点垃圾的小视频哈哈即使当时只是想有个能放下载地址接受大家反馈的地方没想到一下子收到了各种各样的反馈和功能诉求这几天我几乎把所有能想到的功能开发和三个平台的适配全都干出来了还有了"
        }
      ],
      "id": 156,
      "review_annotation": "",
      "row_id": "156",
      "bounded_row_id": 156,
      "parent_row_ids": [
        156
      ],
      "parent_row_id": 156,
      "packet_id": 15,
      "LID_evidence": [
        {
          "region_id": 76,
          "probabilities": {
            "zh": 0.9834216052088244
          },
          "status": "STABLE"
        },
        {
          "region_id": 77,
          "probabilities": {
            "zh": 0.9977233558893204
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        76,
        77
      ],
      "language_confidence": 0.9977233558893204,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 370.3,
      "packet_end": 398.90000000000003,
      "raw_text": "接受大家反馈的地方",
      "selected_text": "接受大家反馈的地方",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            156
          ],
          "packet_id": 15,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        15
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 156
    },
    {
      "start": 386.32,
      "end": 387.06,
      "text": "没想到",
      "raw_asr_text": "没想到",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 15,
      "words": [
        {
          "word": "没",
          "start": 386.32,
          "end": 386.64,
          "probability": 0.99951171875
        },
        {
          "word": "想",
          "start": 386.64,
          "end": 386.84000000000003,
          "probability": 0.99951171875
        },
        {
          "word": "到",
          "start": 386.84000000000003,
          "end": 387.06,
          "probability": 0.9990234375
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.10487588451754662,
        "compression_ratio": 1.3598615916955017,
        "no_speech_prob": 1.723270484643713e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "最开始我只是在两个切片和一个卢播的评论区里发了一下后来又做了个特别简陋现在回头看甚至有点垃圾的小视频哈哈即使当时只是想有个能放下载地址接受大家反馈的地方没想到一下子收到了各种各样的反馈和功能诉求这几天我几乎把所有能想到的功能开发和三个平台的适配全都干出来了还有了"
        }
      ],
      "id": 157,
      "review_annotation": "",
      "row_id": "157",
      "bounded_row_id": 157,
      "parent_row_ids": [
        157
      ],
      "parent_row_id": 157,
      "packet_id": 15,
      "LID_evidence": [
        {
          "region_id": 78,
          "probabilities": {
            "zh": 0.9945706923802694
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        78
      ],
      "language_confidence": 0.9945706923802694,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 370.3,
      "packet_end": 398.90000000000003,
      "raw_text": "没想到",
      "selected_text": "没想到",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            157
          ],
          "packet_id": 15,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        15
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 157
    },
    {
      "start": 387.96000000000004,
      "end": 391.62,
      "text": "一下子收到了各种各样的反馈和功能诉求",
      "raw_asr_text": "一下子收到了各种各样的反馈和功能诉求",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 15,
      "words": [
        {
          "word": "一下",
          "start": 387.96000000000004,
          "end": 388.28000000000003,
          "probability": 0.95654296875
        },
        {
          "word": "子",
          "start": 388.28000000000003,
          "end": 388.6,
          "probability": 1.0
        },
        {
          "word": "收",
          "start": 388.6,
          "end": 389.06,
          "probability": 0.94091796875
        },
        {
          "word": "到了",
          "start": 389.06,
          "end": 389.3,
          "probability": 1.0
        },
        {
          "word": "各",
          "start": 389.3,
          "end": 389.52,
          "probability": 1.0
        },
        {
          "word": "种",
          "start": 389.52,
          "end": 389.74,
          "probability": 1.0
        },
        {
          "word": "各",
          "start": 389.74,
          "end": 389.92,
          "probability": 1.0
        },
        {
          "word": "样",
          "start": 389.92,
          "end": 390.04,
          "probability": 1.0
        },
        {
          "word": "的",
          "start": 390.04,
          "end": 390.14,
          "probability": 1.0
        },
        {
          "word": "反",
          "start": 390.14,
          "end": 390.32,
          "probability": 1.0
        },
        {
          "word": "馈",
          "start": 390.32,
          "end": 390.58000000000004,
          "probability": 1.0
        },
        {
          "word": "和",
          "start": 390.58000000000004,
          "end": 390.88,
          "probability": 0.99951171875
        },
        {
          "word": "功",
          "start": 390.88,
          "end": 391.06,
          "probability": 0.9990234375
        },
        {
          "word": "能",
          "start": 391.06,
          "end": 391.22,
          "probability": 1.0
        },
        {
          "word": "诉",
          "start": 391.22,
          "end": 391.38,
          "probability": 0.998046875
        },
        {
          "word": "求",
          "start": 391.38,
          "end": 391.62,
          "probability": 0.99951171875
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.10487588451754662,
        "compression_ratio": 1.3598615916955017,
        "no_speech_prob": 1.723270484643713e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "最开始我只是在两个切片和一个卢播的评论区里发了一下后来又做了个特别简陋现在回头看甚至有点垃圾的小视频哈哈即使当时只是想有个能放下载地址接受大家反馈的地方没想到一下子收到了各种各样的反馈和功能诉求这几天我几乎把所有能想到的功能开发和三个平台的适配全都干出来了还有了"
        }
      ],
      "id": 158,
      "review_annotation": "",
      "row_id": "158",
      "bounded_row_id": 158,
      "parent_row_ids": [
        158
      ],
      "parent_row_id": 158,
      "packet_id": 15,
      "LID_evidence": [
        {
          "region_id": 79,
          "probabilities": {
            "zh": 0.996545652548472
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        79
      ],
      "language_confidence": 0.996545652548472,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 370.3,
      "packet_end": 398.90000000000003,
      "raw_text": "一下子收到了各种各样的反馈和功能诉求",
      "selected_text": "一下子收到了各种各样的反馈和功能诉求",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            158
          ],
          "packet_id": 15,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        15
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 158
    },
    {
      "start": 392.34000000000003,
      "end": 394.94,
      "text": "这几天我几乎把所有能想到的功能",
      "raw_asr_text": "这几天我几乎把所有能想到的功能",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 15,
      "words": [
        {
          "word": "这",
          "start": 392.34000000000003,
          "end": 392.66,
          "probability": 0.9990234375
        },
        {
          "word": "几",
          "start": 392.66,
          "end": 392.78000000000003,
          "probability": 1.0
        },
        {
          "word": "天",
          "start": 392.78000000000003,
          "end": 392.94,
          "probability": 1.0
        },
        {
          "word": "我",
          "start": 392.94,
          "end": 393.3,
          "probability": 0.99951171875
        },
        {
          "word": "几",
          "start": 393.3,
          "end": 393.48,
          "probability": 0.9990234375
        },
        {
          "word": "乎",
          "start": 393.48,
          "end": 393.58000000000004,
          "probability": 1.0
        },
        {
          "word": "把",
          "start": 393.58000000000004,
          "end": 393.7,
          "probability": 0.99951171875
        },
        {
          "word": "所有",
          "start": 393.7,
          "end": 394.08000000000004,
          "probability": 0.99951171875
        },
        {
          "word": "能",
          "start": 394.08000000000004,
          "end": 394.24,
          "probability": 1.0
        },
        {
          "word": "想",
          "start": 394.24,
          "end": 394.40000000000003,
          "probability": 0.99951171875
        },
        {
          "word": "到",
          "start": 394.40000000000003,
          "end": 394.54,
          "probability": 1.0
        },
        {
          "word": "的",
          "start": 394.54,
          "end": 394.64,
          "probability": 1.0
        },
        {
          "word": "功",
          "start": 394.64,
          "end": 394.8,
          "probability": 0.9990234375
        },
        {
          "word": "能",
          "start": 394.8,
          "end": 394.94,
          "probability": 1.0
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.10487588451754662,
        "compression_ratio": 1.3598615916955017,
        "no_speech_prob": 1.723270484643713e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "最开始我只是在两个切片和一个卢播的评论区里发了一下后来又做了个特别简陋现在回头看甚至有点垃圾的小视频哈哈即使当时只是想有个能放下载地址接受大家反馈的地方没想到一下子收到了各种各样的反馈和功能诉求这几天我几乎把所有能想到的功能开发和三个平台的适配全都干出来了还有了"
        }
      ],
      "id": 159,
      "review_annotation": "",
      "row_id": "159",
      "bounded_row_id": 159,
      "parent_row_ids": [
        159
      ],
      "parent_row_id": 159,
      "packet_id": 15,
      "LID_evidence": [
        {
          "region_id": 80,
          "probabilities": {
            "zh": 0.9981539527575175
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        80
      ],
      "language_confidence": 0.9981539527575175,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 370.3,
      "packet_end": 398.90000000000003,
      "raw_text": "这几天我几乎把所有能想到的功能",
      "selected_text": "这几天我几乎把所有能想到的功能",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            159
          ],
          "packet_id": 15,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        15
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 159
    },
    {
      "start": 394.94,
      "end": 397.78000000000003,
      "text": "开发和三个平台的适配全都干出来了",
      "raw_asr_text": "开发和三个平台的适配全都干出来了",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 15,
      "words": [
        {
          "word": "开",
          "start": 394.94,
          "end": 395.3,
          "probability": 0.9931640625
        },
        {
          "word": "发",
          "start": 395.3,
          "end": 395.54,
          "probability": 0.99658203125
        },
        {
          "word": "和",
          "start": 395.54,
          "end": 395.7,
          "probability": 0.99951171875
        },
        {
          "word": "三",
          "start": 395.7,
          "end": 395.86,
          "probability": 0.9990234375
        },
        {
          "word": "个",
          "start": 395.86,
          "end": 395.98,
          "probability": 1.0
        },
        {
          "word": "平",
          "start": 395.98,
          "end": 396.1,
          "probability": 1.0
        },
        {
          "word": "台",
          "start": 396.1,
          "end": 396.24,
          "probability": 0.99658203125
        },
        {
          "word": "的",
          "start": 396.24,
          "end": 396.34000000000003,
          "probability": 1.0
        },
        {
          "word": "适",
          "start": 396.34000000000003,
          "end": 396.5,
          "probability": 0.9951171875
        },
        {
          "word": "配",
          "start": 396.5,
          "end": 396.66,
          "probability": 0.9990234375
        },
        {
          "word": "全",
          "start": 396.66,
          "end": 397.06,
          "probability": 0.9970703125
        },
        {
          "word": "都",
          "start": 397.06,
          "end": 397.24,
          "probability": 1.0
        },
        {
          "word": "干",
          "start": 397.24,
          "end": 397.38,
          "probability": 0.900390625
        },
        {
          "word": "出",
          "start": 397.38,
          "end": 397.52,
          "probability": 0.99951171875
        },
        {
          "word": "来了",
          "start": 397.52,
          "end": 397.78000000000003,
          "probability": 1.0
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.10487588451754662,
        "compression_ratio": 1.3598615916955017,
        "no_speech_prob": 1.723270484643713e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "最开始我只是在两个切片和一个卢播的评论区里发了一下后来又做了个特别简陋现在回头看甚至有点垃圾的小视频哈哈即使当时只是想有个能放下载地址接受大家反馈的地方没想到一下子收到了各种各样的反馈和功能诉求这几天我几乎把所有能想到的功能开发和三个平台的适配全都干出来了还有了"
        }
      ],
      "id": 160,
      "review_annotation": "",
      "row_id": "160",
      "bounded_row_id": 160,
      "parent_row_ids": [
        160
      ],
      "parent_row_id": 160,
      "packet_id": 15,
      "LID_evidence": [
        {
          "region_id": 80,
          "probabilities": {
            "zh": 0.9981539527575175
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        80
      ],
      "language_confidence": 0.9981539527575175,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 370.3,
      "packet_end": 398.90000000000003,
      "raw_text": "开发和三个平台的适配全都干出来了",
      "selected_text": "开发和三个平台的适配全都干出来了",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            160
          ],
          "packet_id": 15,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        15
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 160
    },
    {
      "start": 398.4,
      "end": 401.03999999999996,
      "text": "他有了今天正式和大家见面的1.0",
      "raw_asr_text": "他有了今天正式和大家见面的1.0",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 16,
      "words": [
        {
          "word": "他",
          "start": 398.28,
          "end": 398.64,
          "probability": 0.03790283203125
        },
        {
          "word": "有",
          "start": 398.64,
          "end": 398.82,
          "probability": 0.9326171875
        },
        {
          "word": "了",
          "start": 398.82,
          "end": 398.97999999999996,
          "probability": 0.99951171875
        },
        {
          "word": "今天",
          "start": 398.97999999999996,
          "end": 399.35999999999996,
          "probability": 0.99462890625
        },
        {
          "word": "正",
          "start": 399.35999999999996,
          "end": 399.64,
          "probability": 0.99755859375
        },
        {
          "word": "式",
          "start": 399.64,
          "end": 399.78,
          "probability": 0.93603515625
        },
        {
          "word": "和",
          "start": 399.78,
          "end": 399.9,
          "probability": 0.99951171875
        },
        {
          "word": "大家",
          "start": 399.9,
          "end": 400.15999999999997,
          "probability": 1.0
        },
        {
          "word": "见",
          "start": 400.15999999999997,
          "end": 400.32,
          "probability": 0.9970703125
        },
        {
          "word": "面",
          "start": 400.32,
          "end": 400.46,
          "probability": 1.0
        },
        {
          "word": "的",
          "start": 400.46,
          "end": 400.55999999999995,
          "probability": 1.0
        },
        {
          "word": "1.",
          "start": 400.55999999999995,
          "end": 400.71999999999997,
          "probability": 0.96142578125
        },
        {
          "word": "0",
          "start": 400.82,
          "end": 401.03999999999996,
          "probability": 0.99951171875
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.13858545237574085,
        "compression_ratio": 1.231939163498099,
        "no_speech_prob": 3.8521175832473276e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "他有了今天正式和大家见面的1.0接下来我还会持续优化他的性能表现和翻译效果比如现在电脑端用的还是和手机端一样的语音识别模型手机上是实在没得选但在桌面端未来其实还有很大的升级和选择空间真心感谢这段时间每一位提出反馈帮我测试"
        }
      ],
      "id": 161,
      "review_annotation": "",
      "row_id": "161",
      "bounded_row_id": 161,
      "parent_row_ids": [
        161
      ],
      "parent_row_id": 161,
      "packet_id": 16,
      "LID_evidence": [
        {
          "region_id": 81,
          "probabilities": {
            "zh": 0.9970018804073334
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        81
      ],
      "language_confidence": 0.9970018804073334,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 397.59999999999997,
      "packet_end": 424.6,
      "raw_text": "他有了今天正式和大家见面的1.0",
      "selected_text": "他有了今天正式和大家见面的1.0",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            161
          ],
          "packet_id": 16,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        16
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 161
    },
    {
      "start": 402.32,
      "end": 403.09999999999997,
      "text": "接下来",
      "raw_asr_text": "接下来",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 16,
      "words": [
        {
          "word": "接",
          "start": 402.32,
          "end": 402.67999999999995,
          "probability": 0.96826171875
        },
        {
          "word": "下",
          "start": 402.67999999999995,
          "end": 402.85999999999996,
          "probability": 1.0
        },
        {
          "word": "来",
          "start": 402.85999999999996,
          "end": 403.09999999999997,
          "probability": 1.0
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.13858545237574085,
        "compression_ratio": 1.231939163498099,
        "no_speech_prob": 3.8521175832473276e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "他有了今天正式和大家见面的1.0接下来我还会持续优化他的性能表现和翻译效果比如现在电脑端用的还是和手机端一样的语音识别模型手机上是实在没得选但在桌面端未来其实还有很大的升级和选择空间真心感谢这段时间每一位提出反馈帮我测试"
        }
      ],
      "id": 162,
      "review_annotation": "",
      "row_id": "162",
      "bounded_row_id": 162,
      "parent_row_ids": [
        162
      ],
      "parent_row_id": 162,
      "packet_id": 16,
      "LID_evidence": [
        {
          "region_id": 82,
          "probabilities": {
            "zh": 0.9944639461381095
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        82
      ],
      "language_confidence": 0.9944639461381095,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 397.59999999999997,
      "packet_end": 424.6,
      "raw_text": "接下来",
      "selected_text": "接下来",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            162
          ],
          "packet_id": 16,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        16
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 162
    },
    {
      "start": 403.09999999999997,
      "end": 405.73999999999995,
      "text": "我还会持续优化他的性能表现",
      "raw_asr_text": "我还会持续优化他的性能表现",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 16,
      "words": [
        {
          "word": "我",
          "start": 403.09999999999997,
          "end": 403.67999999999995,
          "probability": 0.91748046875
        },
        {
          "word": "还",
          "start": 403.67999999999995,
          "end": 403.84,
          "probability": 0.99951171875
        },
        {
          "word": "会",
          "start": 403.84,
          "end": 404.03999999999996,
          "probability": 0.99951171875
        },
        {
          "word": "持",
          "start": 404.03999999999996,
          "end": 404.2,
          "probability": 0.99755859375
        },
        {
          "word": "续",
          "start": 404.2,
          "end": 404.37999999999994,
          "probability": 0.99951171875
        },
        {
          "word": "优",
          "start": 404.37999999999994,
          "end": 404.52,
          "probability": 0.9990234375
        },
        {
          "word": "化",
          "start": 404.52,
          "end": 404.67999999999995,
          "probability": 0.9990234375
        },
        {
          "word": "他的",
          "start": 404.67999999999995,
          "end": 404.91999999999996,
          "probability": 0.96826171875
        },
        {
          "word": "性",
          "start": 404.91999999999996,
          "end": 405.11999999999995,
          "probability": 0.99755859375
        },
        {
          "word": "能",
          "start": 405.11999999999995,
          "end": 405.29999999999995,
          "probability": 1.0
        },
        {
          "word": "表",
          "start": 405.29999999999995,
          "end": 405.46,
          "probability": 1.0
        },
        {
          "word": "现",
          "start": 405.46,
          "end": 405.73999999999995,
          "probability": 0.99951171875
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.13858545237574085,
        "compression_ratio": 1.231939163498099,
        "no_speech_prob": 3.8521175832473276e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "他有了今天正式和大家见面的1.0接下来我还会持续优化他的性能表现和翻译效果比如现在电脑端用的还是和手机端一样的语音识别模型手机上是实在没得选但在桌面端未来其实还有很大的升级和选择空间真心感谢这段时间每一位提出反馈帮我测试"
        }
      ],
      "id": 163,
      "review_annotation": "",
      "row_id": "163",
      "bounded_row_id": 163,
      "parent_row_ids": [
        163
      ],
      "parent_row_id": 163,
      "packet_id": 16,
      "LID_evidence": [
        {
          "region_id": 82,
          "probabilities": {
            "zh": 0.9944639461381095
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        82
      ],
      "language_confidence": 0.9944639461381095,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 397.59999999999997,
      "packet_end": 424.6,
      "raw_text": "我还会持续优化他的性能表现",
      "selected_text": "我还会持续优化他的性能表现",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            163
          ],
          "packet_id": 16,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        16
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 163
    },
    {
      "start": 405.73999999999995,
      "end": 407.93999999999994,
      "text": "和翻译效果",
      "raw_asr_text": "和翻译效果",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 16,
      "words": [
        {
          "word": "和",
          "start": 405.73999999999995,
          "end": 406.41999999999996,
          "probability": 0.9990234375
        },
        {
          "word": "翻",
          "start": 406.41999999999996,
          "end": 407.29999999999995,
          "probability": 0.99658203125
        },
        {
          "word": "译",
          "start": 407.29999999999995,
          "end": 407.49999999999994,
          "probability": 0.998046875
        },
        {
          "word": "效",
          "start": 407.49999999999994,
          "end": 407.67999999999995,
          "probability": 0.99951171875
        },
        {
          "word": "果",
          "start": 407.67999999999995,
          "end": 407.93999999999994,
          "probability": 1.0
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.13858545237574085,
        "compression_ratio": 1.231939163498099,
        "no_speech_prob": 3.8521175832473276e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "他有了今天正式和大家见面的1.0接下来我还会持续优化他的性能表现和翻译效果比如现在电脑端用的还是和手机端一样的语音识别模型手机上是实在没得选但在桌面端未来其实还有很大的升级和选择空间真心感谢这段时间每一位提出反馈帮我测试"
        }
      ],
      "id": 164,
      "review_annotation": "",
      "row_id": "164",
      "bounded_row_id": 164,
      "parent_row_ids": [
        164
      ],
      "parent_row_id": 164,
      "packet_id": 16,
      "LID_evidence": [
        {
          "region_id": 82,
          "probabilities": {
            "zh": 0.9944639461381095
          },
          "status": "STABLE"
        },
        {
          "region_id": 83,
          "probabilities": {
            "zh": 0.9926074913569859
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        82,
        83
      ],
      "language_confidence": 0.9944639461381095,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 397.59999999999997,
      "packet_end": 424.6,
      "raw_text": "和翻译效果",
      "selected_text": "和翻译效果",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            164
          ],
          "packet_id": 16,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        16
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 164
    },
    {
      "start": 408.61999999999995,
      "end": 409.34,
      "text": "比如现在",
      "raw_asr_text": "比如现在",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 16,
      "words": [
        {
          "word": "比如",
          "start": 408.61999999999995,
          "end": 408.97999999999996,
          "probability": 0.99853515625
        },
        {
          "word": "现在",
          "start": 408.97999999999996,
          "end": 409.34,
          "probability": 0.99951171875
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.13858545237574085,
        "compression_ratio": 1.231939163498099,
        "no_speech_prob": 3.8521175832473276e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "他有了今天正式和大家见面的1.0接下来我还会持续优化他的性能表现和翻译效果比如现在电脑端用的还是和手机端一样的语音识别模型手机上是实在没得选但在桌面端未来其实还有很大的升级和选择空间真心感谢这段时间每一位提出反馈帮我测试"
        }
      ],
      "id": 165,
      "review_annotation": "",
      "row_id": "165",
      "bounded_row_id": 165,
      "parent_row_ids": [
        165
      ],
      "parent_row_id": 165,
      "packet_id": 16,
      "LID_evidence": [
        {
          "region_id": 84,
          "probabilities": {
            "zh": 0.996095210313797
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        84
      ],
      "language_confidence": 0.996095210313797,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 397.59999999999997,
      "packet_end": 424.6,
      "raw_text": "比如现在",
      "selected_text": "比如现在",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            165
          ],
          "packet_id": 16,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        16
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 165
    },
    {
      "start": 409.34,
      "end": 411.73999999999995,
      "text": "电脑端用的还是和手机端",
      "raw_asr_text": "电脑端用的还是和手机端",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 16,
      "words": [
        {
          "word": "电",
          "start": 409.34,
          "end": 409.91999999999996,
          "probability": 0.98681640625
        },
        {
          "word": "脑",
          "start": 409.91999999999996,
          "end": 410.08,
          "probability": 1.0
        },
        {
          "word": "端",
          "start": 410.08,
          "end": 410.29999999999995,
          "probability": 1.0
        },
        {
          "word": "用",
          "start": 410.29999999999995,
          "end": 410.71999999999997,
          "probability": 0.99951171875
        },
        {
          "word": "的",
          "start": 410.71999999999997,
          "end": 410.82,
          "probability": 0.99267578125
        },
        {
          "word": "还是",
          "start": 410.82,
          "end": 411.05999999999995,
          "probability": 0.994140625
        },
        {
          "word": "和",
          "start": 411.05999999999995,
          "end": 411.2,
          "probability": 0.9990234375
        },
        {
          "word": "手",
          "start": 411.2,
          "end": 411.37999999999994,
          "probability": 0.99951171875
        },
        {
          "word": "机",
          "start": 411.37999999999994,
          "end": 411.53999999999996,
          "probability": 0.99951171875
        },
        {
          "word": "端",
          "start": 411.53999999999996,
          "end": 411.73999999999995,
          "probability": 1.0
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.13858545237574085,
        "compression_ratio": 1.231939163498099,
        "no_speech_prob": 3.8521175832473276e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "他有了今天正式和大家见面的1.0接下来我还会持续优化他的性能表现和翻译效果比如现在电脑端用的还是和手机端一样的语音识别模型手机上是实在没得选但在桌面端未来其实还有很大的升级和选择空间真心感谢这段时间每一位提出反馈帮我测试"
        }
      ],
      "id": 166,
      "review_annotation": "",
      "row_id": "166",
      "bounded_row_id": 166,
      "parent_row_ids": [
        166
      ],
      "parent_row_id": 166,
      "packet_id": 16,
      "LID_evidence": [
        {
          "region_id": 84,
          "probabilities": {
            "zh": 0.996095210313797
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        84
      ],
      "language_confidence": 0.996095210313797,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 397.59999999999997,
      "packet_end": 424.6,
      "raw_text": "电脑端用的还是和手机端",
      "selected_text": "电脑端用的还是和手机端",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            166
          ],
          "packet_id": 16,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        16
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 166
    },
    {
      "start": 411.73999999999995,
      "end": 413.35999999999996,
      "text": "一样的语音识别模型",
      "raw_asr_text": "一样的语音识别模型",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 16,
      "words": [
        {
          "word": "一",
          "start": 411.73999999999995,
          "end": 412.11999999999995,
          "probability": 0.9990234375
        },
        {
          "word": "样",
          "start": 412.11999999999995,
          "end": 412.21999999999997,
          "probability": 1.0
        },
        {
          "word": "的",
          "start": 412.21999999999997,
          "end": 412.34,
          "probability": 1.0
        },
        {
          "word": "语",
          "start": 412.34,
          "end": 412.47999999999996,
          "probability": 1.0
        },
        {
          "word": "音",
          "start": 412.47999999999996,
          "end": 412.65999999999997,
          "probability": 0.99951171875
        },
        {
          "word": "识",
          "start": 412.65999999999997,
          "end": 412.82,
          "probability": 0.9892578125
        },
        {
          "word": "别",
          "start": 412.82,
          "end": 412.93999999999994,
          "probability": 0.994140625
        },
        {
          "word": "模",
          "start": 412.93999999999994,
          "end": 413.09999999999997,
          "probability": 0.99951171875
        },
        {
          "word": "型",
          "start": 413.09999999999997,
          "end": 413.35999999999996,
          "probability": 0.99853515625
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.13858545237574085,
        "compression_ratio": 1.231939163498099,
        "no_speech_prob": 3.8521175832473276e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "他有了今天正式和大家见面的1.0接下来我还会持续优化他的性能表现和翻译效果比如现在电脑端用的还是和手机端一样的语音识别模型手机上是实在没得选但在桌面端未来其实还有很大的升级和选择空间真心感谢这段时间每一位提出反馈帮我测试"
        }
      ],
      "id": 167,
      "review_annotation": "",
      "row_id": "167",
      "bounded_row_id": 167,
      "parent_row_ids": [
        167
      ],
      "parent_row_id": 167,
      "packet_id": 16,
      "LID_evidence": [
        {
          "region_id": 84,
          "probabilities": {
            "zh": 0.996095210313797
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        84
      ],
      "language_confidence": 0.996095210313797,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 397.59999999999997,
      "packet_end": 424.6,
      "raw_text": "一样的语音识别模型",
      "selected_text": "一样的语音识别模型",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            167
          ],
          "packet_id": 16,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        16
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 167
    },
    {
      "start": 413.35999999999996,
      "end": 415.58,
      "text": "手机上是实在没得选",
      "raw_asr_text": "手机上是实在没得选",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 16,
      "words": [
        {
          "word": "手",
          "start": 413.35999999999996,
          "end": 413.91999999999996,
          "probability": 0.9970703125
        },
        {
          "word": "机",
          "start": 413.91999999999996,
          "end": 414.09999999999997,
          "probability": 1.0
        },
        {
          "word": "上",
          "start": 414.09999999999997,
          "end": 414.29999999999995,
          "probability": 1.0
        },
        {
          "word": "是",
          "start": 414.29999999999995,
          "end": 414.73999999999995,
          "probability": 0.99853515625
        },
        {
          "word": "实",
          "start": 414.73999999999995,
          "end": 414.91999999999996,
          "probability": 0.99951171875
        },
        {
          "word": "在",
          "start": 414.91999999999996,
          "end": 415.08,
          "probability": 0.99951171875
        },
        {
          "word": "没",
          "start": 415.08,
          "end": 415.23999999999995,
          "probability": 0.99658203125
        },
        {
          "word": "得",
          "start": 415.23999999999995,
          "end": 415.34,
          "probability": 0.99755859375
        },
        {
          "word": "选",
          "start": 415.34,
          "end": 415.58,
          "probability": 1.0
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.13858545237574085,
        "compression_ratio": 1.231939163498099,
        "no_speech_prob": 3.8521175832473276e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "他有了今天正式和大家见面的1.0接下来我还会持续优化他的性能表现和翻译效果比如现在电脑端用的还是和手机端一样的语音识别模型手机上是实在没得选但在桌面端未来其实还有很大的升级和选择空间真心感谢这段时间每一位提出反馈帮我测试"
        }
      ],
      "id": 168,
      "review_annotation": "",
      "row_id": "168",
      "bounded_row_id": 168,
      "parent_row_ids": [
        168
      ],
      "parent_row_id": 168,
      "packet_id": 16,
      "LID_evidence": [
        {
          "region_id": 84,
          "probabilities": {
            "zh": 0.996095210313797
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        84
      ],
      "language_confidence": 0.996095210313797,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 397.59999999999997,
      "packet_end": 424.6,
      "raw_text": "手机上是实在没得选",
      "selected_text": "手机上是实在没得选",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            168
          ],
          "packet_id": 16,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        16
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 168
    },
    {
      "start": 415.58,
      "end": 416.78,
      "text": "但在桌面端",
      "raw_asr_text": "但在桌面端",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 16,
      "words": [
        {
          "word": "但",
          "start": 415.58,
          "end": 416.03999999999996,
          "probability": 0.9873046875
        },
        {
          "word": "在",
          "start": 416.03999999999996,
          "end": 416.2,
          "probability": 1.0
        },
        {
          "word": "桌",
          "start": 416.2,
          "end": 416.34,
          "probability": 0.9990234375
        },
        {
          "word": "面",
          "start": 416.34,
          "end": 416.49999999999994,
          "probability": 1.0
        },
        {
          "word": "端",
          "start": 416.49999999999994,
          "end": 416.78,
          "probability": 1.0
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.13858545237574085,
        "compression_ratio": 1.231939163498099,
        "no_speech_prob": 3.8521175832473276e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "他有了今天正式和大家见面的1.0接下来我还会持续优化他的性能表现和翻译效果比如现在电脑端用的还是和手机端一样的语音识别模型手机上是实在没得选但在桌面端未来其实还有很大的升级和选择空间真心感谢这段时间每一位提出反馈帮我测试"
        }
      ],
      "id": 169,
      "review_annotation": "",
      "row_id": "169",
      "bounded_row_id": 169,
      "parent_row_ids": [
        169
      ],
      "parent_row_id": 169,
      "packet_id": 16,
      "LID_evidence": [
        {
          "region_id": 84,
          "probabilities": {
            "zh": 0.996095210313797
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        84
      ],
      "language_confidence": 0.996095210313797,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 397.59999999999997,
      "packet_end": 424.6,
      "raw_text": "但在桌面端",
      "selected_text": "但在桌面端",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            169
          ],
          "packet_id": 16,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        16
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 169
    },
    {
      "start": 416.78,
      "end": 419.79999999999995,
      "text": "未来其实还有很大的升级和选择空间",
      "raw_asr_text": "未来其实还有很大的升级和选择空间",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 16,
      "words": [
        {
          "word": "未",
          "start": 416.78,
          "end": 417.14,
          "probability": 0.9873046875
        },
        {
          "word": "来",
          "start": 417.14,
          "end": 417.35999999999996,
          "probability": 1.0
        },
        {
          "word": "其实",
          "start": 417.35999999999996,
          "end": 417.79999999999995,
          "probability": 0.998046875
        },
        {
          "word": "还有",
          "start": 417.79999999999995,
          "end": 418.05999999999995,
          "probability": 0.9990234375
        },
        {
          "word": "很",
          "start": 418.05999999999995,
          "end": 418.23999999999995,
          "probability": 0.99951171875
        },
        {
          "word": "大的",
          "start": 418.23999999999995,
          "end": 418.47999999999996,
          "probability": 0.9990234375
        },
        {
          "word": "升",
          "start": 418.47999999999996,
          "end": 418.64,
          "probability": 0.96240234375
        },
        {
          "word": "级",
          "start": 418.64,
          "end": 418.85999999999996,
          "probability": 0.994140625
        },
        {
          "word": "和",
          "start": 418.85999999999996,
          "end": 419.14,
          "probability": 0.9990234375
        },
        {
          "word": "选",
          "start": 419.14,
          "end": 419.26,
          "probability": 1.0
        },
        {
          "word": "择",
          "start": 419.26,
          "end": 419.4,
          "probability": 1.0
        },
        {
          "word": "空",
          "start": 419.4,
          "end": 419.55999999999995,
          "probability": 1.0
        },
        {
          "word": "间",
          "start": 419.55999999999995,
          "end": 419.79999999999995,
          "probability": 0.99951171875
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.13858545237574085,
        "compression_ratio": 1.231939163498099,
        "no_speech_prob": 3.8521175832473276e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "他有了今天正式和大家见面的1.0接下来我还会持续优化他的性能表现和翻译效果比如现在电脑端用的还是和手机端一样的语音识别模型手机上是实在没得选但在桌面端未来其实还有很大的升级和选择空间真心感谢这段时间每一位提出反馈帮我测试"
        }
      ],
      "id": 170,
      "review_annotation": "",
      "row_id": "170",
      "bounded_row_id": 170,
      "parent_row_ids": [
        170
      ],
      "parent_row_id": 170,
      "packet_id": 16,
      "LID_evidence": [
        {
          "region_id": 84,
          "probabilities": {
            "zh": 0.996095210313797
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        84
      ],
      "language_confidence": 0.996095210313797,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 397.59999999999997,
      "packet_end": 424.6,
      "raw_text": "未来其实还有很大的升级和选择空间",
      "selected_text": "未来其实还有很大的升级和选择空间",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            170
          ],
          "packet_id": 16,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        16
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 170
    },
    {
      "start": 420.65999999999997,
      "end": 422.03999999999996,
      "text": "真心感谢这段时间",
      "raw_asr_text": "真心感谢这段时间",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 16,
      "words": [
        {
          "word": "真",
          "start": 420.65999999999997,
          "end": 421.02,
          "probability": 0.9892578125
        },
        {
          "word": "心",
          "start": 421.02,
          "end": 421.2,
          "probability": 0.9990234375
        },
        {
          "word": "感",
          "start": 421.2,
          "end": 421.34,
          "probability": 0.994140625
        },
        {
          "word": "谢",
          "start": 421.34,
          "end": 421.47999999999996,
          "probability": 0.998046875
        },
        {
          "word": "这",
          "start": 421.47999999999996,
          "end": 421.61999999999995,
          "probability": 0.99951171875
        },
        {
          "word": "段",
          "start": 421.61999999999995,
          "end": 421.71999999999997,
          "probability": 1.0
        },
        {
          "word": "时间",
          "start": 421.71999999999997,
          "end": 422.03999999999996,
          "probability": 1.0
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.13858545237574085,
        "compression_ratio": 1.231939163498099,
        "no_speech_prob": 3.8521175832473276e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "他有了今天正式和大家见面的1.0接下来我还会持续优化他的性能表现和翻译效果比如现在电脑端用的还是和手机端一样的语音识别模型手机上是实在没得选但在桌面端未来其实还有很大的升级和选择空间真心感谢这段时间每一位提出反馈帮我测试"
        }
      ],
      "id": 171,
      "review_annotation": "",
      "row_id": "171",
      "bounded_row_id": 171,
      "parent_row_ids": [
        171
      ],
      "parent_row_id": 171,
      "packet_id": 16,
      "LID_evidence": [
        {
          "region_id": 85,
          "probabilities": {
            "zh": 0.9939846813678741
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        85
      ],
      "language_confidence": 0.9939846813678741,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 397.59999999999997,
      "packet_end": 424.6,
      "raw_text": "真心感谢这段时间",
      "selected_text": "真心感谢这段时间",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            171
          ],
          "packet_id": 16,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        16
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 171
    },
    {
      "start": 422.03999999999996,
      "end": 423.47999999999996,
      "text": "每一位提出反馈",
      "raw_asr_text": "每一位提出反馈",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 16,
      "words": [
        {
          "word": "每",
          "start": 422.03999999999996,
          "end": 422.55999999999995,
          "probability": 0.97412109375
        },
        {
          "word": "一",
          "start": 422.55999999999995,
          "end": 422.7,
          "probability": 0.99951171875
        },
        {
          "word": "位",
          "start": 422.7,
          "end": 422.79999999999995,
          "probability": 0.99951171875
        },
        {
          "word": "提",
          "start": 422.79999999999995,
          "end": 422.96,
          "probability": 0.9990234375
        },
        {
          "word": "出",
          "start": 422.96,
          "end": 423.09999999999997,
          "probability": 0.99951171875
        },
        {
          "word": "反",
          "start": 423.09999999999997,
          "end": 423.23999999999995,
          "probability": 0.99462890625
        },
        {
          "word": "馈",
          "start": 423.23999999999995,
          "end": 423.47999999999996,
          "probability": 0.99560546875
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.13858545237574085,
        "compression_ratio": 1.231939163498099,
        "no_speech_prob": 3.8521175832473276e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "他有了今天正式和大家见面的1.0接下来我还会持续优化他的性能表现和翻译效果比如现在电脑端用的还是和手机端一样的语音识别模型手机上是实在没得选但在桌面端未来其实还有很大的升级和选择空间真心感谢这段时间每一位提出反馈帮我测试"
        }
      ],
      "id": 172,
      "review_annotation": "",
      "row_id": "172",
      "bounded_row_id": 172,
      "parent_row_ids": [
        172
      ],
      "parent_row_id": 172,
      "packet_id": 16,
      "LID_evidence": [
        {
          "region_id": 85,
          "probabilities": {
            "zh": 0.9939846813678741
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        85
      ],
      "language_confidence": 0.9939846813678741,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 397.59999999999997,
      "packet_end": 424.6,
      "raw_text": "每一位提出反馈",
      "selected_text": "每一位提出反馈",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            172
          ],
          "packet_id": 16,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        16
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 172
    },
    {
      "start": 424.0,
      "end": 426.06,
      "text": "帮我测试捉虫的朋友",
      "raw_asr_text": "帮我测试捉虫的朋友",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 17,
      "words": [
        {
          "word": "帮",
          "start": 423.92,
          "end": 424.28,
          "probability": 0.92431640625
        },
        {
          "word": "我",
          "start": 424.28,
          "end": 424.46,
          "probability": 0.9990234375
        },
        {
          "word": "测",
          "start": 424.46,
          "end": 424.68,
          "probability": 0.9990234375
        },
        {
          "word": "试",
          "start": 424.68,
          "end": 424.96,
          "probability": 1.0
        },
        {
          "word": "捉",
          "start": 424.96,
          "end": 425.52,
          "probability": 0.9130859375
        },
        {
          "word": "虫",
          "start": 425.52,
          "end": 425.7,
          "probability": 0.9990234375
        },
        {
          "word": "的",
          "start": 425.7,
          "end": 425.82,
          "probability": 0.99951171875
        },
        {
          "word": "朋友",
          "start": 425.82,
          "end": 426.06,
          "probability": 0.99951171875
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.14325368629311616,
        "compression_ratio": 0.9917355371900827,
        "no_speech_prob": 3.911510698673126e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "帮我测试捉虫的朋友希望这个深夜上头搓出来的小工具也能陪你度过每一个开心的看台夜晚"
        }
      ],
      "id": 173,
      "review_annotation": "",
      "row_id": "173",
      "bounded_row_id": 173,
      "parent_row_ids": [
        173
      ],
      "parent_row_id": 173,
      "packet_id": 17,
      "LID_evidence": [
        {
          "region_id": 86,
          "probabilities": {
            "zh": 0.9977762500445048
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        86
      ],
      "language_confidence": 0.9977762500445048,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 423.2,
      "packet_end": 433.7,
      "raw_text": "帮我测试捉虫的朋友",
      "selected_text": "帮我测试捉虫的朋友",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            173
          ],
          "packet_id": 17,
          "engine": "mlx-whisper"
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "173",
          "before": "帮我测试捉虫的朋友",
          "after": "帮我测试捉虫的朋友",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "VALID_SPOKEN_LANGUAGE",
          "reason": "捉虫口语化表达，语义通顺",
          "context_evidence": "上下文为测试，'捉虫'为常见口语",
          "independent_ASR": {},
          "suppression_proposed": null
        }
      ],
      "language": "zh",
      "packet_ids": [
        17
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "source_text": "帮我测试捉虫的朋友",
      "review_status": "VALID_SPOKEN_LANGUAGE",
      "canonical_timing_evidence": "CUE_ANCHORED",
      "index": 173
    },
    {
      "start": 426.7,
      "end": 428.06,
      "text": "希望这个深夜上头",
      "raw_asr_text": "希望这个深夜上头",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 17,
      "words": [
        {
          "word": "希望",
          "start": 426.7,
          "end": 427.06,
          "probability": 0.9794921875
        },
        {
          "word": "这个",
          "start": 427.06,
          "end": 427.3,
          "probability": 0.99853515625
        },
        {
          "word": "深",
          "start": 427.3,
          "end": 427.47999999999996,
          "probability": 0.99560546875
        },
        {
          "word": "夜",
          "start": 427.47999999999996,
          "end": 427.65999999999997,
          "probability": 0.99951171875
        },
        {
          "word": "上",
          "start": 427.65999999999997,
          "end": 427.84,
          "probability": 0.98828125
        },
        {
          "word": "头",
          "start": 427.84,
          "end": 428.06,
          "probability": 0.9072265625
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.14325368629311616,
        "compression_ratio": 0.9917355371900827,
        "no_speech_prob": 3.911510698673126e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "帮我测试捉虫的朋友希望这个深夜上头搓出来的小工具也能陪你度过每一个开心的看台夜晚"
        }
      ],
      "id": 174,
      "review_annotation": "",
      "row_id": "174",
      "bounded_row_id": 174,
      "parent_row_ids": [
        174
      ],
      "parent_row_id": 174,
      "packet_id": 17,
      "LID_evidence": [
        {
          "region_id": 87,
          "probabilities": {
            "zh": 0.9979599316914877
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        87
      ],
      "language_confidence": 0.9979599316914877,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 423.2,
      "packet_end": 433.7,
      "raw_text": "希望这个深夜上头",
      "selected_text": "希望这个深夜上头",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            174
          ],
          "packet_id": 17,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        17
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 174
    },
    {
      "start": 428.06,
      "end": 429.24,
      "text": "搓出来的小工具",
      "raw_asr_text": "搓出来的小工具",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 17,
      "words": [
        {
          "word": "搓",
          "start": 428.06,
          "end": 428.47999999999996,
          "probability": 0.85791015625
        },
        {
          "word": "出来",
          "start": 428.47999999999996,
          "end": 428.65999999999997,
          "probability": 0.99609375
        },
        {
          "word": "的",
          "start": 428.65999999999997,
          "end": 428.78,
          "probability": 1.0
        },
        {
          "word": "小",
          "start": 428.78,
          "end": 428.92,
          "probability": 1.0
        },
        {
          "word": "工",
          "start": 428.92,
          "end": 429.09999999999997,
          "probability": 0.99951171875
        },
        {
          "word": "具",
          "start": 429.09999999999997,
          "end": 429.24,
          "probability": 1.0
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.14325368629311616,
        "compression_ratio": 0.9917355371900827,
        "no_speech_prob": 3.911510698673126e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "帮我测试捉虫的朋友希望这个深夜上头搓出来的小工具也能陪你度过每一个开心的看台夜晚"
        }
      ],
      "id": 175,
      "review_annotation": "",
      "row_id": "175",
      "bounded_row_id": 175,
      "parent_row_ids": [
        175
      ],
      "parent_row_id": 175,
      "packet_id": 17,
      "LID_evidence": [
        {
          "region_id": 87,
          "probabilities": {
            "zh": 0.9979599316914877
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        87
      ],
      "language_confidence": 0.9979599316914877,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 423.2,
      "packet_end": 433.7,
      "raw_text": "搓出来的小工具",
      "selected_text": "搓出来的小工具",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            175
          ],
          "packet_id": 17,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        17
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 175
    },
    {
      "start": 429.24,
      "end": 430.7,
      "text": "也能陪你度过",
      "raw_asr_text": "也能陪你度过",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 17,
      "words": [
        {
          "word": "也",
          "start": 429.24,
          "end": 429.92,
          "probability": 0.99853515625
        },
        {
          "word": "能",
          "start": 429.92,
          "end": 430.06,
          "probability": 1.0
        },
        {
          "word": "陪",
          "start": 430.06,
          "end": 430.21999999999997,
          "probability": 0.99951171875
        },
        {
          "word": "你",
          "start": 430.21999999999997,
          "end": 430.38,
          "probability": 0.99951171875
        },
        {
          "word": "度",
          "start": 430.38,
          "end": 430.5,
          "probability": 0.96240234375
        },
        {
          "word": "过",
          "start": 430.5,
          "end": 430.7,
          "probability": 0.99853515625
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.14325368629311616,
        "compression_ratio": 0.9917355371900827,
        "no_speech_prob": 3.911510698673126e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "帮我测试捉虫的朋友希望这个深夜上头搓出来的小工具也能陪你度过每一个开心的看台夜晚"
        }
      ],
      "id": 176,
      "review_annotation": "",
      "row_id": "176",
      "bounded_row_id": 176,
      "parent_row_ids": [
        176
      ],
      "parent_row_id": 176,
      "packet_id": 17,
      "LID_evidence": [
        {
          "region_id": 87,
          "probabilities": {
            "zh": 0.9979599316914877
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        87
      ],
      "language_confidence": 0.9979599316914877,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 423.2,
      "packet_end": 433.7,
      "raw_text": "也能陪你度过",
      "selected_text": "也能陪你度过",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            176
          ],
          "packet_id": 17,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        17
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 176
    },
    {
      "start": 430.7,
      "end": 432.59999999999997,
      "text": "看台夜晚",
      "raw_asr_text": "每一个开心的看台夜晚",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 17,
      "words": [],
      "asr_metrics": {
        "avg_logprob": -0.14325368629311616,
        "compression_ratio": 0.9917355371900827,
        "no_speech_prob": 3.911510698673126e-11
      },
      "lid_status": "STABLE",
      "alternatives": [
        {
          "language": "zh",
          "text": "帮我测试捉虫的朋友希望这个深夜上头搓出来的小工具也能陪你度过每一个开心的看台夜晚"
        }
      ],
      "id": 177,
      "review_annotation": "待確認",
      "row_id": "177",
      "bounded_row_id": 177,
      "parent_row_ids": [
        177
      ],
      "parent_row_id": 177,
      "packet_id": 17,
      "LID_evidence": [
        {
          "region_id": 87,
          "probabilities": {
            "zh": 0.9979599316914877
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        87
      ],
      "language_confidence": 0.9979599316914877,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 423.2,
      "packet_end": 433.7,
      "raw_text": "每一个开心的看台夜晚",
      "selected_text": "看台夜晚",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            177
          ],
          "packet_id": 17,
          "engine": "mlx-whisper"
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "177",
          "before": "每一个开心的看台夜晚",
          "after": "看台夜晚",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "看台疑似看直播或看台误识",
          "context_evidence": "上下文为直播场景，'看直播'更符合逻辑",
          "independent_ASR": {
            "status": "REVIEW_ONLY_RESCUE_BUDGET"
          },
          "suppression_proposed": null
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "177",
          "before": "看台夜晚",
          "after": "看台夜晚",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "Nonsense: '看台夜晚' vs '看台夜晚' (watching late night)",
          "context_evidence": "Context: '看台夜晚'",
          "independent_ASR": {
            "status": "REVIEW_ONLY_RESCUE_BUDGET"
          },
          "suppression_proposed": null
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "177",
          "before": "看台夜晚",
          "after": "看台夜晚",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "Nonsense: '看台夜晚' vs '看台夜晚' (watching late night)",
          "context_evidence": "Context: '看台夜晚'",
          "independent_ASR": {
            "status": "REVIEW_ONLY_RESCUE_BUDGET"
          },
          "suppression_proposed": null
        }
      ],
      "language": "zh",
      "packet_ids": [
        17
      ],
      "classification": "BEST_SUPPORTED_WITH_REVIEW",
      "independent_ASR": {
        "status": "REVIEW_ONLY_RESCUE_BUDGET"
      },
      "raw_word_timestamps": [
        {
          "word": "每",
          "start": 430.7,
          "end": 431.06,
          "probability": 0.99951171875
        },
        {
          "word": "一个",
          "start": 431.06,
          "end": 431.3,
          "probability": 0.99853515625
        },
        {
          "word": "开",
          "start": 431.3,
          "end": 431.5,
          "probability": 0.99853515625
        },
        {
          "word": "心",
          "start": 431.5,
          "end": 431.65999999999997,
          "probability": 0.9990234375
        },
        {
          "word": "的",
          "start": 431.65999999999997,
          "end": 431.8,
          "probability": 0.998046875
        },
        {
          "word": "看",
          "start": 431.8,
          "end": 432.08,
          "probability": 0.5634765625
        },
        {
          "word": "台",
          "start": 432.08,
          "end": 432.26,
          "probability": 0.8935546875
        },
        {
          "word": "夜",
          "start": 432.26,
          "end": 432.42,
          "probability": 0.99755859375
        },
        {
          "word": "晚",
          "start": 432.42,
          "end": 432.59999999999997,
          "probability": 0.99755859375
        }
      ],
      "word_timestamps": [],
      "source_text": "看台夜晚",
      "review_status": "BEST_SUPPORTED_WITH_REVIEW",
      "canonical_timing_evidence": "CUE_ANCHORED",
      "index": 177
    },
    {
      "start": 435.3,
      "end": 439.04,
      "text": "项目代码和三个平台的安装包都在GitHub上",
      "raw_asr_text": "项目代码和三个平台的安装包都在GitHub上",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 18,
      "words": [
        {
          "word": "项",
          "start": 435.2,
          "end": 435.58,
          "probability": 0.97216796875
        },
        {
          "word": "目",
          "start": 435.58,
          "end": 435.74,
          "probability": 0.9990234375
        },
        {
          "word": "代",
          "start": 435.74,
          "end": 435.86,
          "probability": 0.99609375
        },
        {
          "word": "码",
          "start": 435.86,
          "end": 436.14,
          "probability": 0.99609375
        },
        {
          "word": "和",
          "start": 436.14,
          "end": 436.52,
          "probability": 0.9990234375
        },
        {
          "word": "三",
          "start": 436.52,
          "end": 436.7,
          "probability": 0.99267578125
        },
        {
          "word": "个",
          "start": 436.7,
          "end": 436.84,
          "probability": 0.99853515625
        },
        {
          "word": "平",
          "start": 436.84,
          "end": 436.96,
          "probability": 0.99951171875
        },
        {
          "word": "台",
          "start": 436.96,
          "end": 437.14,
          "probability": 0.99951171875
        },
        {
          "word": "的",
          "start": 437.14,
          "end": 437.24,
          "probability": 0.99951171875
        },
        {
          "word": "安",
          "start": 437.24,
          "end": 437.4,
          "probability": 0.99951171875
        },
        {
          "word": "装",
          "start": 437.4,
          "end": 437.56,
          "probability": 1.0
        },
        {
          "word": "包",
          "start": 437.56,
          "end": 437.76,
          "probability": 0.9990234375
        },
        {
          "word": "都",
          "start": 437.76,
          "end": 438.2,
          "probability": 0.99755859375
        },
        {
          "word": "在",
          "start": 438.2,
          "end": 438.4,
          "probability": 1.0
        },
        {
          "word": "G",
          "start": 438.4,
          "end": 438.56,
          "probability": 0.91650390625
        },
        {
          "word": "it",
          "start": 438.56,
          "end": 438.66,
          "probability": 0.72998046875
        },
        {
          "word": "Hub",
          "start": 438.66,
          "end": 438.8,
          "probability": 0.9599609375
        },
        {
          "word": "上",
          "start": 438.8,
          "end": 439.04,
          "probability": 0.99853515625
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.1683034896850586,
        "compression_ratio": 0.9833333333333333,
        "no_speech_prob": 1.4493324075603997e-11
      },
      "lid_status": "UNCERTAIN",
      "alternatives": [
        {
          "language": "zh",
          "text": "项目代码和三个平台的安装包都在GitHub上搜索CMU Subtitle就能找到如果它也帮到了你别忘了点赞投币收藏支持一下呀我们下个版本见拜拜"
        }
      ],
      "id": 178,
      "review_annotation": "",
      "row_id": "178",
      "bounded_row_id": 178,
      "parent_row_ids": [
        178
      ],
      "parent_row_id": 178,
      "packet_id": 18,
      "LID_evidence": [
        {
          "region_id": 88,
          "probabilities": {
            "zh": 0.830851229528586
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        88
      ],
      "language_confidence": 0.830851229528586,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 434.5,
      "packet_end": 450.5173125,
      "raw_text": "项目代码和三个平台的安装包都在GitHub上",
      "selected_text": "项目代码和三个平台的安装包都在GitHub上",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            178
          ],
          "packet_id": 18,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        18
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 178
    },
    {
      "start": 439.46,
      "end": 442.44,
      "text": "搜索CMU Subtitle就能找到",
      "raw_asr_text": "搜索CMU Subtitle就能找到",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 18,
      "words": [
        {
          "word": "搜",
          "start": 439.46,
          "end": 439.84,
          "probability": 0.998046875
        },
        {
          "word": "索",
          "start": 439.84,
          "end": 440.06,
          "probability": 0.99951171875
        },
        {
          "word": "C",
          "start": 440.06,
          "end": 440.54,
          "probability": 0.2958984375
        },
        {
          "word": "M",
          "start": 440.54,
          "end": 440.68,
          "probability": 0.4794921875
        },
        {
          "word": "U",
          "start": 440.68,
          "end": 440.74,
          "probability": 0.469482421875
        },
        {
          "word": " Sub",
          "start": 440.74,
          "end": 440.92,
          "probability": 0.5244140625
        },
        {
          "word": "tit",
          "start": 440.92,
          "end": 441.14,
          "probability": 0.994140625
        },
        {
          "word": "le",
          "start": 441.14,
          "end": 441.38,
          "probability": 0.9970703125
        },
        {
          "word": "就",
          "start": 441.38,
          "end": 441.88,
          "probability": 0.9912109375
        },
        {
          "word": "能",
          "start": 441.88,
          "end": 442.04,
          "probability": 1.0
        },
        {
          "word": "找",
          "start": 442.04,
          "end": 442.22,
          "probability": 1.0
        },
        {
          "word": "到",
          "start": 442.22,
          "end": 442.44,
          "probability": 1.0
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.1683034896850586,
        "compression_ratio": 0.9833333333333333,
        "no_speech_prob": 1.4493324075603997e-11
      },
      "lid_status": "UNCERTAIN",
      "alternatives": [
        {
          "language": "zh",
          "text": "项目代码和三个平台的安装包都在GitHub上搜索CMU Subtitle就能找到如果它也帮到了你别忘了点赞投币收藏支持一下呀我们下个版本见拜拜"
        }
      ],
      "id": 179,
      "review_annotation": "待確認",
      "row_id": "179",
      "bounded_row_id": 179,
      "parent_row_ids": [
        179
      ],
      "parent_row_id": 179,
      "packet_id": 18,
      "LID_evidence": [
        {
          "region_id": 89,
          "probabilities": {
            "zh": 0.9033115889877081
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        89
      ],
      "language_confidence": 0.9033115889877081,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 434.5,
      "packet_end": 450.5173125,
      "raw_text": "搜索CMU Subtitle就能找到",
      "selected_text": "搜索CMU Subtitle就能找到",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            179
          ],
          "packet_id": 18,
          "engine": "mlx-whisper"
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "179",
          "before": "搜索CMU Subtitle就能找到",
          "after": "搜索CMU Subtitle就能找到",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "CMU疑似同传或特定项目名误识",
          "context_evidence": "上下文为项目名称，'同传'或'CMU'可能为专有名词",
          "independent_ASR": {
            "engine": "whisper-large-v3",
            "model_path": "/Users/sagawa/AI_WORK/cache/huggingface/hub/models--mlx-community--whisper-large-v3",
            "language": "zh",
            "audio_interval": [
              437.46,
              444.44
            ],
            "target_interval": [
              439.46,
              442.44
            ],
            "text": "索SymuSubtitle就能找到",
            "words": [
              {
                "word": "索",
                "start": 439.76,
                "end": 440.06,
                "probability": 1.0
              },
              {
                "word": "S",
                "start": 440.06,
                "end": 440.38,
                "probability": 0.42529296875
              },
              {
                "word": "ym",
                "start": 440.38,
                "end": 440.52,
                "probability": 0.23681640625
              },
              {
                "word": "u",
                "start": 440.52,
                "end": 440.62,
                "probability": 0.289306640625
              },
              {
                "word": " Sub",
                "start": 440.62,
                "end": 440.84,
                "probability": 0.58740234375
              },
              {
                "word": "tit",
                "start": 440.84,
                "end": 440.96,
                "probability": 0.99951171875
              },
              {
                "word": "le",
                "start": 440.96,
                "end": 441.21999999999997,
                "probability": 0.986328125
              },
              {
                "word": "就",
                "start": 441.21999999999997,
                "end": 441.79999999999995,
                "probability": 0.9912109375
              },
              {
                "word": "能",
                "start": 441.79999999999995,
                "end": 441.94,
                "probability": 1.0
              },
              {
                "word": "找",
                "start": 441.94,
                "end": 442.15999999999997,
                "probability": 1.0
              },
              {
                "word": "到",
                "start": 442.15999999999997,
                "end": 442.35999999999996,
                "probability": 1.0
              }
            ],
            "raw_result": {
              "text": "都在Github上搜索Symu Subtitle就能找到如果他也帮到了你",
              "segments": [
                {
                  "id": 0,
                  "seek": 0,
                  "start": 0.0,
                  "end": 1.48,
                  "text": "都在Github上",
                  "tokens": [
                    50365,
                    7182,
                    3581,
                    38,
                    355,
                    836,
                    5708,
                    50446
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.3237885634104411,
                  "compression_ratio": 0.8641975308641975,
                  "no_speech_prob": 0.8838439583778381,
                  "words": [
                    {
                      "word": "都",
                      "start": 0.0,
                      "end": 0.36,
                      "probability": 0.044677734375
                    },
                    {
                      "word": "在",
                      "start": 0.36,
                      "end": 0.86,
                      "probability": 0.986328125
                    },
                    {
                      "word": "G",
                      "start": 0.86,
                      "end": 1.0,
                      "probability": 0.71142578125
                    },
                    {
                      "word": "ith",
                      "start": 1.0,
                      "end": 1.12,
                      "probability": 0.86376953125
                    },
                    {
                      "word": "ub",
                      "start": 1.12,
                      "end": 1.3,
                      "probability": 1.0
                    },
                    {
                      "word": "上",
                      "start": 1.3,
                      "end": 1.48,
                      "probability": 0.998046875
                    }
                  ]
                },
                {
                  "id": 1,
                  "seek": 0,
                  "start": 1.48,
                  "end": 3.76,
                  "text": "搜索Symu Subtitle",
                  "tokens": [
                    50446,
                    162,
                    14987,
                    7732,
                    95,
                    50,
                    4199,
                    84,
                    8511,
                    27689,
                    306,
                    50563
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.3237885634104411,
                  "compression_ratio": 0.8641975308641975,
                  "no_speech_prob": 0.8838439583778381,
                  "words": [
                    {
                      "word": "搜",
                      "start": 1.48,
                      "end": 2.3,
                      "probability": 0.9443359375
                    },
                    {
                      "word": "索",
                      "start": 2.3,
                      "end": 2.6,
                      "probability": 1.0
                    },
                    {
                      "word": "S",
                      "start": 2.6,
                      "end": 2.92,
                      "probability": 0.42529296875
                    },
                    {
                      "word": "ym",
                      "start": 2.92,
                      "end": 3.06,
                      "probability": 0.23681640625
                    },
                    {
                      "word": "u",
                      "start": 3.06,
                      "end": 3.16,
                      "probability": 0.289306640625
                    },
                    {
                      "word": " Sub",
                      "start": 3.16,
                      "end": 3.38,
                      "probability": 0.58740234375
                    },
                    {
                      "word": "tit",
                      "start": 3.38,
                      "end": 3.5,
                      "probability": 0.99951171875
                    },
                    {
                      "word": "le",
                      "start": 3.5,
                      "end": 3.76,
                      "probability": 0.986328125
                    }
                  ]
                },
                {
                  "id": 2,
                  "seek": 0,
                  "start": 3.76,
                  "end": 4.9,
                  "text": "就能找到",
                  "tokens": [
                    50563,
                    3111,
                    8225,
                    25085,
                    4511,
                    50618
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.3237885634104411,
                  "compression_ratio": 0.8641975308641975,
                  "no_speech_prob": 0.8838439583778381,
                  "words": [
                    {
                      "word": "就",
                      "start": 3.76,
                      "end": 4.34,
                      "probability": 0.9912109375
                    },
                    {
                      "word": "能",
                      "start": 4.34,
                      "end": 4.48,
                      "probability": 1.0
                    },
                    {
                      "word": "找",
                      "start": 4.48,
                      "end": 4.7,
                      "probability": 1.0
                    },
                    {
                      "word": "到",
                      "start": 4.7,
                      "end": 4.9,
                      "probability": 1.0
                    }
                  ]
                },
                {
                  "id": 3,
                  "seek": 0,
                  "start": 5.46,
                  "end": 6.84,
                  "text": "如果他也帮到了你",
                  "tokens": [
                    50618,
                    13119,
                    5000,
                    6404,
                    4845,
                    106,
                    21381,
                    2166,
                    50712
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.3237885634104411,
                  "compression_ratio": 0.8641975308641975,
                  "no_speech_prob": 0.8838439583778381,
                  "words": [
                    {
                      "word": "如果",
                      "start": 5.46,
                      "end": 5.88,
                      "probability": 0.97705078125
                    },
                    {
                      "word": "他",
                      "start": 5.88,
                      "end": 6.12,
                      "probability": 0.51123046875
                    },
                    {
                      "word": "也",
                      "start": 6.12,
                      "end": 6.28,
                      "probability": 0.998046875
                    },
                    {
                      "word": "帮",
                      "start": 6.28,
                      "end": 6.48,
                      "probability": 1.0
                    },
                    {
                      "word": "到了",
                      "start": 6.48,
                      "end": 6.58,
                      "probability": 0.9990234375
                    },
                    {
                      "word": "你",
                      "start": 6.58,
                      "end": 6.84,
                      "probability": 0.99951171875
                    }
                  ]
                }
              ],
              "language": "zh",
              "duration": 6.98
            },
            "status": "COMPLETE",
            "requested_language": "zh"
          },
          "suppression_proposed": null
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "179",
          "before": "搜索CMU Subtitle就能找到",
          "after": "搜索CMU Subtitle就能找到",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "Name error: 'CMU' vs '同传' (simultaneous interpretation)",
          "context_evidence": "Context: '搜索同传字幕'",
          "independent_ASR": {
            "engine": "whisper-large-v3",
            "model_path": "/Users/sagawa/AI_WORK/cache/huggingface/hub/models--mlx-community--whisper-large-v3",
            "language": "zh",
            "audio_interval": [
              437.46,
              444.44
            ],
            "target_interval": [
              439.46,
              442.44
            ],
            "text": "索SymuSubtitle就能找到",
            "words": [
              {
                "word": "索",
                "start": 439.76,
                "end": 440.06,
                "probability": 1.0
              },
              {
                "word": "S",
                "start": 440.06,
                "end": 440.38,
                "probability": 0.42529296875
              },
              {
                "word": "ym",
                "start": 440.38,
                "end": 440.52,
                "probability": 0.23681640625
              },
              {
                "word": "u",
                "start": 440.52,
                "end": 440.62,
                "probability": 0.289306640625
              },
              {
                "word": " Sub",
                "start": 440.62,
                "end": 440.84,
                "probability": 0.58740234375
              },
              {
                "word": "tit",
                "start": 440.84,
                "end": 440.96,
                "probability": 0.99951171875
              },
              {
                "word": "le",
                "start": 440.96,
                "end": 441.21999999999997,
                "probability": 0.986328125
              },
              {
                "word": "就",
                "start": 441.21999999999997,
                "end": 441.79999999999995,
                "probability": 0.9912109375
              },
              {
                "word": "能",
                "start": 441.79999999999995,
                "end": 441.94,
                "probability": 1.0
              },
              {
                "word": "找",
                "start": 441.94,
                "end": 442.15999999999997,
                "probability": 1.0
              },
              {
                "word": "到",
                "start": 442.15999999999997,
                "end": 442.35999999999996,
                "probability": 1.0
              }
            ],
            "raw_result": {
              "text": "都在Github上搜索Symu Subtitle就能找到如果他也帮到了你",
              "segments": [
                {
                  "id": 0,
                  "seek": 0,
                  "start": 0.0,
                  "end": 1.48,
                  "text": "都在Github上",
                  "tokens": [
                    50365,
                    7182,
                    3581,
                    38,
                    355,
                    836,
                    5708,
                    50446
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.3237885634104411,
                  "compression_ratio": 0.8641975308641975,
                  "no_speech_prob": 0.8838439583778381,
                  "words": [
                    {
                      "word": "都",
                      "start": 0.0,
                      "end": 0.36,
                      "probability": 0.044677734375
                    },
                    {
                      "word": "在",
                      "start": 0.36,
                      "end": 0.86,
                      "probability": 0.986328125
                    },
                    {
                      "word": "G",
                      "start": 0.86,
                      "end": 1.0,
                      "probability": 0.71142578125
                    },
                    {
                      "word": "ith",
                      "start": 1.0,
                      "end": 1.12,
                      "probability": 0.86376953125
                    },
                    {
                      "word": "ub",
                      "start": 1.12,
                      "end": 1.3,
                      "probability": 1.0
                    },
                    {
                      "word": "上",
                      "start": 1.3,
                      "end": 1.48,
                      "probability": 0.998046875
                    }
                  ]
                },
                {
                  "id": 1,
                  "seek": 0,
                  "start": 1.48,
                  "end": 3.76,
                  "text": "搜索Symu Subtitle",
                  "tokens": [
                    50446,
                    162,
                    14987,
                    7732,
                    95,
                    50,
                    4199,
                    84,
                    8511,
                    27689,
                    306,
                    50563
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.3237885634104411,
                  "compression_ratio": 0.8641975308641975,
                  "no_speech_prob": 0.8838439583778381,
                  "words": [
                    {
                      "word": "搜",
                      "start": 1.48,
                      "end": 2.3,
                      "probability": 0.9443359375
                    },
                    {
                      "word": "索",
                      "start": 2.3,
                      "end": 2.6,
                      "probability": 1.0
                    },
                    {
                      "word": "S",
                      "start": 2.6,
                      "end": 2.92,
                      "probability": 0.42529296875
                    },
                    {
                      "word": "ym",
                      "start": 2.92,
                      "end": 3.06,
                      "probability": 0.23681640625
                    },
                    {
                      "word": "u",
                      "start": 3.06,
                      "end": 3.16,
                      "probability": 0.289306640625
                    },
                    {
                      "word": " Sub",
                      "start": 3.16,
                      "end": 3.38,
                      "probability": 0.58740234375
                    },
                    {
                      "word": "tit",
                      "start": 3.38,
                      "end": 3.5,
                      "probability": 0.99951171875
                    },
                    {
                      "word": "le",
                      "start": 3.5,
                      "end": 3.76,
                      "probability": 0.986328125
                    }
                  ]
                },
                {
                  "id": 2,
                  "seek": 0,
                  "start": 3.76,
                  "end": 4.9,
                  "text": "就能找到",
                  "tokens": [
                    50563,
                    3111,
                    8225,
                    25085,
                    4511,
                    50618
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.3237885634104411,
                  "compression_ratio": 0.8641975308641975,
                  "no_speech_prob": 0.8838439583778381,
                  "words": [
                    {
                      "word": "就",
                      "start": 3.76,
                      "end": 4.34,
                      "probability": 0.9912109375
                    },
                    {
                      "word": "能",
                      "start": 4.34,
                      "end": 4.48,
                      "probability": 1.0
                    },
                    {
                      "word": "找",
                      "start": 4.48,
                      "end": 4.7,
                      "probability": 1.0
                    },
                    {
                      "word": "到",
                      "start": 4.7,
                      "end": 4.9,
                      "probability": 1.0
                    }
                  ]
                },
                {
                  "id": 3,
                  "seek": 0,
                  "start": 5.46,
                  "end": 6.84,
                  "text": "如果他也帮到了你",
                  "tokens": [
                    50618,
                    13119,
                    5000,
                    6404,
                    4845,
                    106,
                    21381,
                    2166,
                    50712
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.3237885634104411,
                  "compression_ratio": 0.8641975308641975,
                  "no_speech_prob": 0.8838439583778381,
                  "words": [
                    {
                      "word": "如果",
                      "start": 5.46,
                      "end": 5.88,
                      "probability": 0.97705078125
                    },
                    {
                      "word": "他",
                      "start": 5.88,
                      "end": 6.12,
                      "probability": 0.51123046875
                    },
                    {
                      "word": "也",
                      "start": 6.12,
                      "end": 6.28,
                      "probability": 0.998046875
                    },
                    {
                      "word": "帮",
                      "start": 6.28,
                      "end": 6.48,
                      "probability": 1.0
                    },
                    {
                      "word": "到了",
                      "start": 6.48,
                      "end": 6.58,
                      "probability": 0.9990234375
                    },
                    {
                      "word": "你",
                      "start": 6.58,
                      "end": 6.84,
                      "probability": 0.99951171875
                    }
                  ]
                }
              ],
              "language": "zh",
              "duration": 6.98
            },
            "status": "COMPLETE",
            "requested_language": "zh"
          },
          "suppression_proposed": null
        },
        {
          "stage": "full_transcript_canonical_gate",
          "row_id": "179",
          "before": "搜索CMU Subtitle就能找到",
          "after": "搜索CMU Subtitle就能找到",
          "language_before": "zh",
          "language_after": "zh",
          "classification": "BEST_SUPPORTED_WITH_REVIEW",
          "reason": "Name error: 'CMU' vs '同传' (simultaneous interpretation)",
          "context_evidence": "Context: '搜索同传字幕'",
          "independent_ASR": {
            "engine": "whisper-large-v3",
            "model_path": "/Users/sagawa/AI_WORK/cache/huggingface/hub/models--mlx-community--whisper-large-v3",
            "language": "zh",
            "audio_interval": [
              437.46,
              444.44
            ],
            "target_interval": [
              439.46,
              442.44
            ],
            "text": "索SymuSubtitle就能找到",
            "words": [
              {
                "word": "索",
                "start": 439.76,
                "end": 440.06,
                "probability": 1.0
              },
              {
                "word": "S",
                "start": 440.06,
                "end": 440.38,
                "probability": 0.42529296875
              },
              {
                "word": "ym",
                "start": 440.38,
                "end": 440.52,
                "probability": 0.23681640625
              },
              {
                "word": "u",
                "start": 440.52,
                "end": 440.62,
                "probability": 0.289306640625
              },
              {
                "word": " Sub",
                "start": 440.62,
                "end": 440.84,
                "probability": 0.58740234375
              },
              {
                "word": "tit",
                "start": 440.84,
                "end": 440.96,
                "probability": 0.99951171875
              },
              {
                "word": "le",
                "start": 440.96,
                "end": 441.21999999999997,
                "probability": 0.986328125
              },
              {
                "word": "就",
                "start": 441.21999999999997,
                "end": 441.79999999999995,
                "probability": 0.9912109375
              },
              {
                "word": "能",
                "start": 441.79999999999995,
                "end": 441.94,
                "probability": 1.0
              },
              {
                "word": "找",
                "start": 441.94,
                "end": 442.15999999999997,
                "probability": 1.0
              },
              {
                "word": "到",
                "start": 442.15999999999997,
                "end": 442.35999999999996,
                "probability": 1.0
              }
            ],
            "raw_result": {
              "text": "都在Github上搜索Symu Subtitle就能找到如果他也帮到了你",
              "segments": [
                {
                  "id": 0,
                  "seek": 0,
                  "start": 0.0,
                  "end": 1.48,
                  "text": "都在Github上",
                  "tokens": [
                    50365,
                    7182,
                    3581,
                    38,
                    355,
                    836,
                    5708,
                    50446
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.3237885634104411,
                  "compression_ratio": 0.8641975308641975,
                  "no_speech_prob": 0.8838439583778381,
                  "words": [
                    {
                      "word": "都",
                      "start": 0.0,
                      "end": 0.36,
                      "probability": 0.044677734375
                    },
                    {
                      "word": "在",
                      "start": 0.36,
                      "end": 0.86,
                      "probability": 0.986328125
                    },
                    {
                      "word": "G",
                      "start": 0.86,
                      "end": 1.0,
                      "probability": 0.71142578125
                    },
                    {
                      "word": "ith",
                      "start": 1.0,
                      "end": 1.12,
                      "probability": 0.86376953125
                    },
                    {
                      "word": "ub",
                      "start": 1.12,
                      "end": 1.3,
                      "probability": 1.0
                    },
                    {
                      "word": "上",
                      "start": 1.3,
                      "end": 1.48,
                      "probability": 0.998046875
                    }
                  ]
                },
                {
                  "id": 1,
                  "seek": 0,
                  "start": 1.48,
                  "end": 3.76,
                  "text": "搜索Symu Subtitle",
                  "tokens": [
                    50446,
                    162,
                    14987,
                    7732,
                    95,
                    50,
                    4199,
                    84,
                    8511,
                    27689,
                    306,
                    50563
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.3237885634104411,
                  "compression_ratio": 0.8641975308641975,
                  "no_speech_prob": 0.8838439583778381,
                  "words": [
                    {
                      "word": "搜",
                      "start": 1.48,
                      "end": 2.3,
                      "probability": 0.9443359375
                    },
                    {
                      "word": "索",
                      "start": 2.3,
                      "end": 2.6,
                      "probability": 1.0
                    },
                    {
                      "word": "S",
                      "start": 2.6,
                      "end": 2.92,
                      "probability": 0.42529296875
                    },
                    {
                      "word": "ym",
                      "start": 2.92,
                      "end": 3.06,
                      "probability": 0.23681640625
                    },
                    {
                      "word": "u",
                      "start": 3.06,
                      "end": 3.16,
                      "probability": 0.289306640625
                    },
                    {
                      "word": " Sub",
                      "start": 3.16,
                      "end": 3.38,
                      "probability": 0.58740234375
                    },
                    {
                      "word": "tit",
                      "start": 3.38,
                      "end": 3.5,
                      "probability": 0.99951171875
                    },
                    {
                      "word": "le",
                      "start": 3.5,
                      "end": 3.76,
                      "probability": 0.986328125
                    }
                  ]
                },
                {
                  "id": 2,
                  "seek": 0,
                  "start": 3.76,
                  "end": 4.9,
                  "text": "就能找到",
                  "tokens": [
                    50563,
                    3111,
                    8225,
                    25085,
                    4511,
                    50618
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.3237885634104411,
                  "compression_ratio": 0.8641975308641975,
                  "no_speech_prob": 0.8838439583778381,
                  "words": [
                    {
                      "word": "就",
                      "start": 3.76,
                      "end": 4.34,
                      "probability": 0.9912109375
                    },
                    {
                      "word": "能",
                      "start": 4.34,
                      "end": 4.48,
                      "probability": 1.0
                    },
                    {
                      "word": "找",
                      "start": 4.48,
                      "end": 4.7,
                      "probability": 1.0
                    },
                    {
                      "word": "到",
                      "start": 4.7,
                      "end": 4.9,
                      "probability": 1.0
                    }
                  ]
                },
                {
                  "id": 3,
                  "seek": 0,
                  "start": 5.46,
                  "end": 6.84,
                  "text": "如果他也帮到了你",
                  "tokens": [
                    50618,
                    13119,
                    5000,
                    6404,
                    4845,
                    106,
                    21381,
                    2166,
                    50712
                  ],
                  "temperature": 0.0,
                  "avg_logprob": -0.3237885634104411,
                  "compression_ratio": 0.8641975308641975,
                  "no_speech_prob": 0.8838439583778381,
                  "words": [
                    {
                      "word": "如果",
                      "start": 5.46,
                      "end": 5.88,
                      "probability": 0.97705078125
                    },
                    {
                      "word": "他",
                      "start": 5.88,
                      "end": 6.12,
                      "probability": 0.51123046875
                    },
                    {
                      "word": "也",
                      "start": 6.12,
                      "end": 6.28,
                      "probability": 0.998046875
                    },
                    {
                      "word": "帮",
                      "start": 6.28,
                      "end": 6.48,
                      "probability": 1.0
                    },
                    {
                      "word": "到了",
                      "start": 6.48,
                      "end": 6.58,
                      "probability": 0.9990234375
                    },
                    {
                      "word": "你",
                      "start": 6.58,
                      "end": 6.84,
                      "probability": 0.99951171875
                    }
                  ]
                }
              ],
              "language": "zh",
              "duration": 6.98
            },
            "status": "COMPLETE",
            "requested_language": "zh"
          },
          "suppression_proposed": null
        }
      ],
      "language": "zh",
      "packet_ids": [
        18
      ],
      "classification": "BEST_SUPPORTED_WITH_REVIEW",
      "independent_ASR": {
        "engine": "whisper-large-v3",
        "model_path": "/Users/sagawa/AI_WORK/cache/huggingface/hub/models--mlx-community--whisper-large-v3",
        "language": "zh",
        "audio_interval": [
          437.46,
          444.44
        ],
        "target_interval": [
          439.46,
          442.44
        ],
        "text": "索SymuSubtitle就能找到",
        "words": [
          {
            "word": "索",
            "start": 439.76,
            "end": 440.06,
            "probability": 1.0
          },
          {
            "word": "S",
            "start": 440.06,
            "end": 440.38,
            "probability": 0.42529296875
          },
          {
            "word": "ym",
            "start": 440.38,
            "end": 440.52,
            "probability": 0.23681640625
          },
          {
            "word": "u",
            "start": 440.52,
            "end": 440.62,
            "probability": 0.289306640625
          },
          {
            "word": " Sub",
            "start": 440.62,
            "end": 440.84,
            "probability": 0.58740234375
          },
          {
            "word": "tit",
            "start": 440.84,
            "end": 440.96,
            "probability": 0.99951171875
          },
          {
            "word": "le",
            "start": 440.96,
            "end": 441.21999999999997,
            "probability": 0.986328125
          },
          {
            "word": "就",
            "start": 441.21999999999997,
            "end": 441.79999999999995,
            "probability": 0.9912109375
          },
          {
            "word": "能",
            "start": 441.79999999999995,
            "end": 441.94,
            "probability": 1.0
          },
          {
            "word": "找",
            "start": 441.94,
            "end": 442.15999999999997,
            "probability": 1.0
          },
          {
            "word": "到",
            "start": 442.15999999999997,
            "end": 442.35999999999996,
            "probability": 1.0
          }
        ],
        "raw_result": {
          "text": "都在Github上搜索Symu Subtitle就能找到如果他也帮到了你",
          "segments": [
            {
              "id": 0,
              "seek": 0,
              "start": 0.0,
              "end": 1.48,
              "text": "都在Github上",
              "tokens": [
                50365,
                7182,
                3581,
                38,
                355,
                836,
                5708,
                50446
              ],
              "temperature": 0.0,
              "avg_logprob": -0.3237885634104411,
              "compression_ratio": 0.8641975308641975,
              "no_speech_prob": 0.8838439583778381,
              "words": [
                {
                  "word": "都",
                  "start": 0.0,
                  "end": 0.36,
                  "probability": 0.044677734375
                },
                {
                  "word": "在",
                  "start": 0.36,
                  "end": 0.86,
                  "probability": 0.986328125
                },
                {
                  "word": "G",
                  "start": 0.86,
                  "end": 1.0,
                  "probability": 0.71142578125
                },
                {
                  "word": "ith",
                  "start": 1.0,
                  "end": 1.12,
                  "probability": 0.86376953125
                },
                {
                  "word": "ub",
                  "start": 1.12,
                  "end": 1.3,
                  "probability": 1.0
                },
                {
                  "word": "上",
                  "start": 1.3,
                  "end": 1.48,
                  "probability": 0.998046875
                }
              ]
            },
            {
              "id": 1,
              "seek": 0,
              "start": 1.48,
              "end": 3.76,
              "text": "搜索Symu Subtitle",
              "tokens": [
                50446,
                162,
                14987,
                7732,
                95,
                50,
                4199,
                84,
                8511,
                27689,
                306,
                50563
              ],
              "temperature": 0.0,
              "avg_logprob": -0.3237885634104411,
              "compression_ratio": 0.8641975308641975,
              "no_speech_prob": 0.8838439583778381,
              "words": [
                {
                  "word": "搜",
                  "start": 1.48,
                  "end": 2.3,
                  "probability": 0.9443359375
                },
                {
                  "word": "索",
                  "start": 2.3,
                  "end": 2.6,
                  "probability": 1.0
                },
                {
                  "word": "S",
                  "start": 2.6,
                  "end": 2.92,
                  "probability": 0.42529296875
                },
                {
                  "word": "ym",
                  "start": 2.92,
                  "end": 3.06,
                  "probability": 0.23681640625
                },
                {
                  "word": "u",
                  "start": 3.06,
                  "end": 3.16,
                  "probability": 0.289306640625
                },
                {
                  "word": " Sub",
                  "start": 3.16,
                  "end": 3.38,
                  "probability": 0.58740234375
                },
                {
                  "word": "tit",
                  "start": 3.38,
                  "end": 3.5,
                  "probability": 0.99951171875
                },
                {
                  "word": "le",
                  "start": 3.5,
                  "end": 3.76,
                  "probability": 0.986328125
                }
              ]
            },
            {
              "id": 2,
              "seek": 0,
              "start": 3.76,
              "end": 4.9,
              "text": "就能找到",
              "tokens": [
                50563,
                3111,
                8225,
                25085,
                4511,
                50618
              ],
              "temperature": 0.0,
              "avg_logprob": -0.3237885634104411,
              "compression_ratio": 0.8641975308641975,
              "no_speech_prob": 0.8838439583778381,
              "words": [
                {
                  "word": "就",
                  "start": 3.76,
                  "end": 4.34,
                  "probability": 0.9912109375
                },
                {
                  "word": "能",
                  "start": 4.34,
                  "end": 4.48,
                  "probability": 1.0
                },
                {
                  "word": "找",
                  "start": 4.48,
                  "end": 4.7,
                  "probability": 1.0
                },
                {
                  "word": "到",
                  "start": 4.7,
                  "end": 4.9,
                  "probability": 1.0
                }
              ]
            },
            {
              "id": 3,
              "seek": 0,
              "start": 5.46,
              "end": 6.84,
              "text": "如果他也帮到了你",
              "tokens": [
                50618,
                13119,
                5000,
                6404,
                4845,
                106,
                21381,
                2166,
                50712
              ],
              "temperature": 0.0,
              "avg_logprob": -0.3237885634104411,
              "compression_ratio": 0.8641975308641975,
              "no_speech_prob": 0.8838439583778381,
              "words": [
                {
                  "word": "如果",
                  "start": 5.46,
                  "end": 5.88,
                  "probability": 0.97705078125
                },
                {
                  "word": "他",
                  "start": 5.88,
                  "end": 6.12,
                  "probability": 0.51123046875
                },
                {
                  "word": "也",
                  "start": 6.12,
                  "end": 6.28,
                  "probability": 0.998046875
                },
                {
                  "word": "帮",
                  "start": 6.28,
                  "end": 6.48,
                  "probability": 1.0
                },
                {
                  "word": "到了",
                  "start": 6.48,
                  "end": 6.58,
                  "probability": 0.9990234375
                },
                {
                  "word": "你",
                  "start": 6.58,
                  "end": 6.84,
                  "probability": 0.99951171875
                }
              ]
            }
          ],
          "language": "zh",
          "duration": 6.98
        },
        "status": "COMPLETE",
        "requested_language": "zh"
      },
      "source_text": "搜索CMU Subtitle就能找到",
      "review_status": "BEST_SUPPORTED_WITH_REVIEW",
      "canonical_timing_evidence": "CUE_ANCHORED",
      "index": 179
    },
    {
      "start": 443.1,
      "end": 444.36,
      "text": "如果它也帮到了你",
      "raw_asr_text": "如果它也帮到了你",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 18,
      "words": [
        {
          "word": "如果",
          "start": 443.1,
          "end": 443.48,
          "probability": 0.99658203125
        },
        {
          "word": "它",
          "start": 443.48,
          "end": 443.66,
          "probability": 0.72802734375
        },
        {
          "word": "也",
          "start": 443.66,
          "end": 443.8,
          "probability": 0.99951171875
        },
        {
          "word": "帮",
          "start": 443.8,
          "end": 443.94,
          "probability": 1.0
        },
        {
          "word": "到了",
          "start": 443.94,
          "end": 444.14,
          "probability": 0.99072265625
        },
        {
          "word": "你",
          "start": 444.14,
          "end": 444.36,
          "probability": 0.99951171875
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.1683034896850586,
        "compression_ratio": 0.9833333333333333,
        "no_speech_prob": 1.4493324075603997e-11
      },
      "lid_status": "UNCERTAIN",
      "alternatives": [
        {
          "language": "zh",
          "text": "项目代码和三个平台的安装包都在GitHub上搜索CMU Subtitle就能找到如果它也帮到了你别忘了点赞投币收藏支持一下呀我们下个版本见拜拜"
        }
      ],
      "id": 180,
      "review_annotation": "",
      "row_id": "180",
      "bounded_row_id": 180,
      "parent_row_ids": [
        180
      ],
      "parent_row_id": 180,
      "packet_id": 18,
      "LID_evidence": [
        {
          "region_id": 89,
          "probabilities": {
            "zh": 0.9033115889877081
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        89
      ],
      "language_confidence": 0.9033115889877081,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 434.5,
      "packet_end": 450.5173125,
      "raw_text": "如果它也帮到了你",
      "selected_text": "如果它也帮到了你",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            180
          ],
          "packet_id": 18,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        18
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 180
    },
    {
      "start": 444.36,
      "end": 447.54,
      "text": "别忘了点赞投币收藏支持一下呀",
      "raw_asr_text": "别忘了点赞投币收藏支持一下呀",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 18,
      "words": [
        {
          "word": "别",
          "start": 444.36,
          "end": 444.7,
          "probability": 0.9912109375
        },
        {
          "word": "忘",
          "start": 444.7,
          "end": 444.84,
          "probability": 1.0
        },
        {
          "word": "了",
          "start": 444.84,
          "end": 444.96,
          "probability": 1.0
        },
        {
          "word": "点",
          "start": 444.96,
          "end": 445.1,
          "probability": 0.99951171875
        },
        {
          "word": "赞",
          "start": 445.1,
          "end": 445.32,
          "probability": 0.9990234375
        },
        {
          "word": "投",
          "start": 445.32,
          "end": 445.58,
          "probability": 0.68408203125
        },
        {
          "word": "币",
          "start": 445.58,
          "end": 445.84,
          "probability": 0.9921875
        },
        {
          "word": "收",
          "start": 445.84,
          "end": 446.26,
          "probability": 0.9697265625
        },
        {
          "word": "藏",
          "start": 446.26,
          "end": 446.52,
          "probability": 1.0
        },
        {
          "word": "支持",
          "start": 446.52,
          "end": 446.96,
          "probability": 0.9912109375
        },
        {
          "word": "一下",
          "start": 446.96,
          "end": 447.34,
          "probability": 0.99951171875
        },
        {
          "word": "呀",
          "start": 447.34,
          "end": 447.54,
          "probability": 0.958984375
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.1683034896850586,
        "compression_ratio": 0.9833333333333333,
        "no_speech_prob": 1.4493324075603997e-11
      },
      "lid_status": "UNCERTAIN",
      "alternatives": [
        {
          "language": "zh",
          "text": "项目代码和三个平台的安装包都在GitHub上搜索CMU Subtitle就能找到如果它也帮到了你别忘了点赞投币收藏支持一下呀我们下个版本见拜拜"
        }
      ],
      "id": 181,
      "review_annotation": "",
      "row_id": "181",
      "bounded_row_id": 181,
      "parent_row_ids": [
        181
      ],
      "parent_row_id": 181,
      "packet_id": 18,
      "LID_evidence": [
        {
          "region_id": 89,
          "probabilities": {
            "zh": 0.9033115889877081
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        89
      ],
      "language_confidence": 0.9033115889877081,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 434.5,
      "packet_end": 450.5173125,
      "raw_text": "别忘了点赞投币收藏支持一下呀",
      "selected_text": "别忘了点赞投币收藏支持一下呀",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            181
          ],
          "packet_id": 18,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        18
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 181
    },
    {
      "start": 447.54,
      "end": 448.9,
      "text": "我们下个版本见",
      "raw_asr_text": "我们下个版本见",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 18,
      "words": [
        {
          "word": "我们",
          "start": 447.54,
          "end": 448.16,
          "probability": 0.99951171875
        },
        {
          "word": "下",
          "start": 448.16,
          "end": 448.34,
          "probability": 0.97607421875
        },
        {
          "word": "个",
          "start": 448.34,
          "end": 448.44,
          "probability": 0.99365234375
        },
        {
          "word": "版",
          "start": 448.44,
          "end": 448.58,
          "probability": 1.0
        },
        {
          "word": "本",
          "start": 448.58,
          "end": 448.72,
          "probability": 0.9990234375
        },
        {
          "word": "见",
          "start": 448.72,
          "end": 448.9,
          "probability": 0.9990234375
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.1683034896850586,
        "compression_ratio": 0.9833333333333333,
        "no_speech_prob": 1.4493324075603997e-11
      },
      "lid_status": "UNCERTAIN",
      "alternatives": [
        {
          "language": "zh",
          "text": "项目代码和三个平台的安装包都在GitHub上搜索CMU Subtitle就能找到如果它也帮到了你别忘了点赞投币收藏支持一下呀我们下个版本见拜拜"
        }
      ],
      "id": 182,
      "review_annotation": "",
      "row_id": "182",
      "bounded_row_id": 182,
      "parent_row_ids": [
        182
      ],
      "parent_row_id": 182,
      "packet_id": 18,
      "LID_evidence": [
        {
          "region_id": 89,
          "probabilities": {
            "zh": 0.9033115889877081
          },
          "status": "STABLE"
        },
        {
          "region_id": 90,
          "probabilities": {
            "zh": 0.9974267929792404
          },
          "status": "STABLE"
        }
      ],
      "VAD_region_ids": [
        89,
        90
      ],
      "language_confidence": 0.9974267929792404,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 434.5,
      "packet_end": 450.5173125,
      "raw_text": "我们下个版本见",
      "selected_text": "我们下个版本见",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            182
          ],
          "packet_id": 18,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        18
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 182
    },
    {
      "start": 448.9,
      "end": 449.58,
      "text": "拜拜",
      "raw_asr_text": "拜拜",
      "source_language": "zh",
      "source_lang": "zh",
      "asr_packet_id": 18,
      "words": [
        {
          "word": "拜拜",
          "start": 448.9,
          "end": 449.58,
          "probability": 0.97900390625
        }
      ],
      "asr_metrics": {
        "avg_logprob": -0.1683034896850586,
        "compression_ratio": 0.9833333333333333,
        "no_speech_prob": 1.4493324075603997e-11
      },
      "lid_status": "UNCERTAIN",
      "alternatives": [
        {
          "language": "zh",
          "text": "项目代码和三个平台的安装包都在GitHub上搜索CMU Subtitle就能找到如果它也帮到了你别忘了点赞投币收藏支持一下呀我们下个版本见拜拜"
        }
      ],
      "id": 183,
      "review_annotation": "",
      "row_id": "183",
      "bounded_row_id": 183,
      "parent_row_ids": [
        183
      ],
      "parent_row_id": 183,
      "packet_id": 18,
      "LID_evidence": [
        {
          "region_id": 90,
          "probabilities": {
            "zh": 0.9974267929792404
          },
          "status": "STABLE"
        },
        {
          "region_id": 91,
          "probabilities": {
            "zh": 0.9972147345542908
          },
          "status": "UNCERTAIN"
        }
      ],
      "VAD_region_ids": [
        90,
        91
      ],
      "language_confidence": 0.9974267929792404,
      "ASR_engine": "mlx-whisper",
      "ASR_route": "bounded_vad_lid",
      "packet_start": 434.5,
      "packet_end": 450.5173125,
      "raw_text": "拜拜",
      "selected_text": "拜拜",
      "provenance_chain": [
        {
          "stage": "bounded_asr",
          "parent_row_ids": [
            183
          ],
          "packet_id": 18,
          "engine": "mlx-whisper"
        }
      ],
      "language": "zh",
      "packet_ids": [
        18
      ],
      "classification": "VALID_SPOKEN_LANGUAGE",
      "index": 183
    }
  ],
  "transcript_selection": {
    "schema": "videonote.evidence-preserving-selector.v1",
    "status": "SELECTED",
    "selected_source": "bounded_multilingual_asr_local_review",
    "whisper_rejected": false,
    "source_profile": {
      "duration_coverage": null
    },
    "decision_reason": "bounded_backbone_preserved_local_QC",
    "handoff_report": "/Users/sagawa/AI_WORK/video_notes/url/20261015_132044__开源_同传字幕_1.0_正式版_Android_Mac_Windows_三端实时同传字幕_看生肉直播再也不用等/production_selector_handoff_report.json"
  },
  "canonical_stage": "full_document_canonical_gate",
  "canonical_ready_for_translation": false
}