{
  "text": "有人把Whisper做到了89ms中位延迟,还顺手把GPU峰值内存砍掉48%。Whisper离线转写很强,但一做实时字幕,就会遇到假设飘移,重算变重,静音误触发。Whisper Pipe最聪明的地方不是换模型,而是把实时转写拆成一条可控流水线。\n它用双缓冲把已确认文本和未确认音频分开,活动音频窗口最多30秒,历史不会越背越重。\n更关键的是词级时间戳,哪段文字提交了,就按最后一个词的结束时间,把旧音频精准裁掉。\n提交也不是拍脑袋,简单句走快速通道,模糊句走三次确认,字幕少闪,延迟也压住。\n结果很硬,P90延迟142ms,平均延迟下降81.1%,WER只比离线Whisper高约2%。\n资源更夸张,GPU峰值内存降低48%,平均利用率降低80.9%,150分钟内存零增长。\n现在用它,别走失效的源码安装,直接Pipe install WhisperPipe,然后用Pipe Stream启动。\n所以Whisper Pipe最值得看的不是单个数字,而是它让实时转写像一个能长期工作的系统。\n关注置身工坊,继续拆语音AI和开源项目。",
  "language": "zh",
  "duration": 75.328,
  "segments": [
    {
      "id": 0,
      "seek": 0,
      "start": 0.0,
      "end": 19.32,
      "text": "有人把Whisper做到了89ms中位延迟,还顺手把GPU峰值内存砍掉48%。Whisper离线转写很强,但一做实时字幕,就会遇到假设飘移,重算变重,静音误触发。Whisper Pipe最聪明的地方不是换模型,而是把实时转写拆成一条可控流水线。",
      "tokens": [
        50365,
        45820,
        16075,
        2471,
        271,
        610,
        10907,
        21381,
        21115,
        2592,
        5975,
        11160,
        17527,
        114,
        3316,
        253,
        11,
        14852,
        10178,
        118,
        11389,
        16075,
        38,
        8115,
        26555,
        108,
        40242,
        34742,
        39781,
        23230,
        235,
        29327,
        13318,
        4,
        1543,
        2471,
        271,
        610,
        18614,
        119,
        16853,
        123,
        17819,
        105,
        5676,
        247,
        4563,
        5702,
        118,
        11,
        8395,
        2257,
        10907,
        24726,
        15729,
        22381,
        40211,
        11,
        3111,
        12949,
        3330,
        229,
        4511,
        31706,
        7422,
        122,
        11808,
        246,
        8204,
        119,
        11,
        12624,
        19497,
        2129,
        246,
        12624,
        11,
        5363,
        247,
        18034,
        5233,
        107,
        6758,
        99,
        28926,
        1543,
        2471,
        271,
        610,
        11240,
        8661,
        8171,
        103,
        11100,
        1546,
        30146,
        7296,
        26075,
        95,
        41908,
        39823,
        11,
        11070,
        1541,
        16075,
        24726,
        15729,
        17819,
        105,
        5676,
        247,
        6852,
        228,
        11336,
        2257,
        48837,
        4429,
        48707,
        27854,
        15590,
        16853,
        123,
        1543,
        51331
      ],
      "temperature": 0,
      "avg_logprob": -0.0793113491751931,
      "compression_ratio": 1.181547619047619,
      "no_speech_prob": 8.655473039687056e-12,
      "source": "whisper",
      "text_original": "有人把Whisper做到了89ms中位延迟,还顺手把GPU峰值内存砍掉48%。Whisper离线转写很强,但一做实时字幕,就会遇到假设飘移,重算变重,静音误触发。Whisper pipe最聪明的地方不是换模型,而是把实时转写拆成一条可控流水线。",
      "correction_applied": true,
      "correction_source": "llm_correct_a_presegmentation"
    },
    {
      "id": 1,
      "seek": 0,
      "start": 20.32,
      "end": 26.46,
      "text": "它用双缓冲把已确认文本和未确认音频分开,活动音频窗口最多30秒,历史不会越背越重。",
      "tokens": [
        51381,
        11284,
        9254,
        2129,
        234,
        38109,
        241,
        5676,
        110,
        16075,
        15701,
        38114,
        106,
        7422,
        97,
        17174,
        8802,
        12565,
        29954,
        38114,
        106,
        7422,
        97,
        18034,
        39752,
        6627,
        18937,
        11,
        25956,
        34961,
        18034,
        39752,
        30771,
        245,
        18144,
        8661,
        6392,
        3446,
        20956,
        11,
        5014,
        228,
        45399,
        47928,
        25761,
        46329,
        25761,
        12624,
        1543,
        51688
      ],
      "temperature": 0,
      "avg_logprob": -0.0793113491751931,
      "compression_ratio": 1.181547619047619,
      "no_speech_prob": 8.655473039687056e-12,
      "source": "whisper"
    },
    {
      "id": 2,
      "seek": 2646,
      "start": 26.46,
      "end": 34.120000000000005,
      "text": "更关键的是词级时间戳,哪段文字提交了,就按最后一个词的结束时间,把旧音频精准裁掉。",
      "tokens": [
        50365,
        19002,
        28053,
        23049,
        106,
        24620,
        5233,
        235,
        26020,
        44848,
        1486,
        111,
        11,
        17028,
        28427,
        17174,
        22381,
        20949,
        28455,
        2289,
        11,
        3111,
        26613,
        8661,
        13547,
        20182,
        5233,
        235,
        1546,
        45641,
        44320,
        44848,
        11,
        16075,
        4479,
        100,
        18034,
        39752,
        34910,
        6336,
        228,
        8083,
        223,
        29327,
        1543,
        50748
      ],
      "temperature": 0,
      "avg_logprob": -0.102517713349441,
      "compression_ratio": 1.111913357400722,
      "no_speech_prob": 1.5042148082677897e-11,
      "source": "whisper",
      "text_original": "更关键的是词集时间戳,哪段文字提交了,就按最后一个词的结束时间,把旧音频精准裁掉。",
      "correction_applied": true,
      "correction_source": "llm_correct_a_presegmentation"
    },
    {
      "id": 3,
      "seek": 2646,
      "start": 35.38,
      "end": 41.2,
      "text": "提交也不是拍脑袋,简单句走快速通道,模糊句走三次确认,字幕少闪,延迟也压住。",
      "tokens": [
        50811,
        20949,
        28455,
        6404,
        7296,
        31113,
        27067,
        239,
        19688,
        233,
        11,
        11249,
        222,
        47446,
        34592,
        9575,
        10251,
        31217,
        19550,
        6025,
        11,
        41908,
        15239,
        232,
        34592,
        9575,
        10960,
        9487,
        38114,
        106,
        7422,
        97,
        11,
        22381,
        40211,
        15686,
        8259,
        103,
        11,
        17527,
        114,
        3316,
        253,
        6404,
        5014,
        233,
        21632,
        1543,
        51102
      ],
      "temperature": 0,
      "avg_logprob": -0.102517713349441,
      "compression_ratio": 1.111913357400722,
      "no_speech_prob": 1.5042148082677897e-11,
      "source": "whisper"
    },
    {
      "id": 4,
      "seek": 2646,
      "start": 42.3,
      "end": 49.38,
      "text": "结果很硬,P90延迟142ms,平均延迟下降81.1%,WER只比离线Whisper高约2%。",
      "tokens": [
        51157,
        45641,
        9319,
        4563,
        38114,
        105,
        11,
        47,
        7771,
        17527,
        114,
        3316,
        253,
        7271,
        17,
        2592,
        11,
        16716,
        14872,
        229,
        17527,
        114,
        3316,
        253,
        4438,
        47421,
        32875,
        13,
        16,
        8923,
        37937,
        17,
        14003,
        11706,
        18614,
        119,
        16853,
        123,
        2471,
        271,
        610,
        12979,
        16853,
        99,
        17,
        4,
        1543,
        51511
      ],
      "temperature": 0,
      "avg_logprob": -0.102517713349441,
      "compression_ratio": 1.111913357400722,
      "no_speech_prob": 1.5042148082677897e-11,
      "source": "whisper",
      "text_original": "结果很硬,P90延迟142ms,平均延迟下降81.1%,WE2只比离线Whisper高约2%。",
      "correction_applied": true,
      "correction_source": "llm_correct_a_presegmentation"
    },
    {
      "id": 5,
      "seek": 4938,
      "start": 49.38,
      "end": 56.78,
      "text": "资源更夸张,GPU峰值内存降低48%,平均利用率降低80.9%,150分钟内存零增长。",
      "tokens": [
        50365,
        5266,
        226,
        47402,
        19002,
        1787,
        116,
        44059,
        11,
        38,
        8115,
        26555,
        108,
        40242,
        34742,
        39781,
        47421,
        41377,
        13318,
        8923,
        16716,
        14872,
        229,
        23700,
        9254,
        44866,
        47421,
        41377,
        4702,
        13,
        24,
        8923,
        20120,
        6627,
        50064,
        34742,
        39781,
        6306,
        114,
        24228,
        252,
        32271,
        1543,
        50735
      ],
      "temperature": 0,
      "avg_logprob": -0.10327505111694336,
      "compression_ratio": 1.134375,
      "no_speech_prob": 1.3908199245071806e-11,
      "source": "whisper"
    },
    {
      "id": 6,
      "seek": 4938,
      "start": 57.86,
      "end": 63.160000000000004,
      "text": "现在用它,别走失效的源码安装,直接Pipe install WhisperPipe,然后用Pipe Stream启动。",
      "tokens": [
        50789,
        25040,
        9254,
        11284,
        11,
        18453,
        9575,
        28564,
        43076,
        1546,
        2523,
        228,
        23230,
        223,
        16206,
        8083,
        227,
        11,
        43297,
        47,
        6527,
        3625,
        41132,
        610,
        430,
        6527,
        11,
        26636,
        9254,
        47,
        6527,
        24904,
        2392,
        107,
        34961,
        1543,
        51054
      ],
      "temperature": 0,
      "avg_logprob": -0.10327505111694336,
      "compression_ratio": 1.134375,
      "no_speech_prob": 1.3908199245071806e-11,
      "source": "whisper",
      "text_original": "现在用它,别走失效的圆码安装,直接Pipe install Whisper Pipe,然后用Pipe Stream启动。",
      "correction_applied": true,
      "correction_source": "llm_correct_a_presegmentation"
    },
    {
      "id": 7,
      "seek": 4938,
      "start": 64.26,
      "end": 70.04,
      "text": "所以Whisper Pipe最值得看的不是单个数字,而是它让实时转写像一个能长期工作的系统。",
      "tokens": [
        51109,
        7239,
        2471,
        271,
        610,
        430,
        6527,
        8661,
        40242,
        5916,
        4200,
        1546,
        7296,
        47446,
        7549,
        33188,
        22381,
        11,
        11070,
        1541,
        11284,
        33650,
        24726,
        15729,
        17819,
        105,
        5676,
        247,
        12760,
        20182,
        8225,
        32271,
        16786,
        41315,
        1546,
        25368,
        10115,
        253,
        1543,
        51398
      ],
      "temperature": 0,
      "avg_logprob": -0.10327505111694336,
      "compression_ratio": 1.134375,
      "no_speech_prob": 1.3908199245071806e-11,
      "source": "whisper"
    },
    {
      "id": 8,
      "seek": 4938,
      "start": 71.14,
      "end": 74.16,
      "text": "关注置身工坊,继续拆语音AI和开源项目。",
      "tokens": [
        51453,
        28053,
        26432,
        34719,
        19847,
        23323,
        14872,
        232,
        11,
        10115,
        100,
        10115,
        255,
        6852,
        228,
        5233,
        255,
        18034,
        48698,
        12565,
        18937,
        47402,
        10178,
        117,
        11386,
        1543,
        51604
      ],
      "temperature": 0,
      "avg_logprob": -0.10327505111694336,
      "compression_ratio": 1.134375,
      "no_speech_prob": 1.3908199245071806e-11,
      "source": "whisper"
    }
  ],
  "transcript_selection": {
    "schema": "videonote.transcript_selection.v1",
    "status": "SELECTED",
    "requested_language": "zh",
    "duration": 75.328,
    "reference_path": null,
    "reference_kind": null,
    "reference_language": null,
    "selected_source": "whisper_segmented",
    "decision_reason": "source_unavailable_or_untrusted_whisper_passed",
    "whisper_rejected": false,
    "source_profile": {
      "segment_count": 0,
      "duration_coverage": 0.0,
      "usable_full": false,
      "usable_partial": false,
      "hallucination_region_count": 0,
      "hallucination_regions": [],
      "language": {
        "requested": "zh",
        "matched": false
      }
    },
    "whisper_profile": {
      "segment_count": 9,
      "multilingual": false,
      "first_start": 0.0,
      "final_end": 74.16,
      "duration_coverage": 0.984494,
      "maximum_gap_seconds": 1.26,
      "inverted_count": 0,
      "language": {
        "requested": "zh",
        "matched": true,
        "latin": 90,
        "han": 294,
        "kana": 0,
        "hangul": 0
      },
      "hallucination_region_count": 0,
      "hallucination_regions": [],
      "usable_full": true,
      "usable_partial": true
    },
    "comparison": {
      "window_seconds": 30.0,
      "compared_window_count": 0,
      "median_score": 0.0,
      "divergent_window_count": 0,
      "divergent_ratio": 0.0,
      "divergent_windows": []
    },
    "timestamp_alignment": {
      "applied": false,
      "estimated_offset_seconds": 0.0,
      "matched_segment_count": 0,
      "match_scores": []
    },
    "caption_timing_repair": null,
    "filled_source_profile": null,
    "whisper_duration_guard": {
      "media_duration": 75.328,
      "rejected_segment_count": 1,
      "rejected_segments": [
        {
          "index": 9,
          "start": 74.16,
          "end": 104.14,
          "reason": "material_end_timestamp_overrun"
        }
      ]
    }
  },
  "presegmentation_correction": {
    "applied": true,
    "correction_count": 4,
    "deterministic_fix_count": 0
  }
}