{
  "text": "大家好,今天我們給大家講解這個Voice Agent的一個總體架構啊,以及它的這個完整流程。\n我們首先了解一下什麼是Voice Agent。你看這個Voice,Voice就是語音的意思,就是語音智能體。也就是說不僅有語音,還有智能體。\n那最簡單的比如說你用豆包的語音功能，或者用LGBT的語音功能，你跟它對話。那這個就是最基本的一個語音智能體。\n它裡面有哪些關鍵技術呢?有WebRTC，有STT，有TTS，有LLM，這幾個關鍵的技術。這幾個名詞呢，希望大家都去查一下。\n那首先，WebRTC解決什麼?WebRTC解決就是說，音視頻的問題。你比如說你的聲音怎麼採集的，你聲音怎麼播放的，聲音怎麼傳輸的，這個是由WebRTC來解決的。\n那麼STT是什麼?STT就是\n語音轉文字。\n也就是說\n你語音採集了以後,需要把它轉成文字,這個文字再交給大語音模型。\n也就是LM\n決定說什麼?\n也就是說,LM是把文字拿到以後,它會加工,會思考,會返回。\n返回完以後,\n這不有文字了嗎?\n文字在\n第一個,它首先會發到\n這個端上,也就是比如,一般我們叫前端,比如網頁裡面,或者手機裡面。\n然後不經意發,有可能要播語音,這個播語音就要把這個文字再轉成語音,這個過程叫TTS。\n所以說,這裡面有幾個關鍵概念,叫Web RC\nSTD\nTTS\nLM\n對吧\n好,我們來看一下它總體架構\n首先是\n前端,前端呢,它有一個\n比如說可能是瀏覽器\n有可能是App\n甚至是一個硬件啊,對吧\n好,那麼麥克風採集\n那音頻\n同時音頻可能會播放,然後屏幕上可能有文字\n大家用這個\n用豆包的時候,你在跟它對話的時候,它不經常有這個\n會有文字顯示嗎?另外一個它會播放語音\n那麼音頻會上行\n到後端\n後端呢\n它這裡面呢,它會比如說,它一般是Web RC\n處理啊,處理完\n以後呢,它會\n把這個聲音提取出來\n然後把這個聲音呢,丟給\nSTT\nSTT呢\n呃\n有幾種,一種是你本地的這種服務,對吧\n你可能從歐拉馬來接入\n第二種呢,有可能比如說你是阿里的\n對吧\n也有可能是騰訊的\n也有可能是火山引擎的\n也可能是MiniMax\n是吧\nSTT,MiniMax\n目前\n不一定有\n然後你可能Deepgram\n有可能是ElevenLab\n這裡面就有中英文\n這裡面我們主要以中英文給大家解釋就行了\n那麼由它轉動文字又返回給\n那麼由它轉動文字又返回給這個後端，後端這塊它又會調這個LLM，也就是調的大語言模型\n這裡面選擇就多了,你比如說你可以用Deepseek\n可以用這個\nMiniMax,可以用這個Kimi\n可以用這個OpenAI,對吧\n你可以用\n這個很多很多模型啊,都可以用\n那\n丟給大元模型,大元模型會\n對你的話進行加工\n加工完以後\n他比如說你今天問了一個\n舉個例子,什麼是人工智能對吧\n他會肯定回答\n一堆這個內容嘛,然後\n他返回完\n返回他以後,然後後端呢\n又\n緊接著\n調這個TTS模型,這個TTS模型也有\n比如火山引擎呢,阿里騰訊呢\n對吧,Deepgram,ElevenLav,他都會把\n文字再轉成語音\n轉成語音呢,他會有一個\n實時的流，這個實時流呢，再通過WebRTC技術再傳給前端\n前端把聲音播放出來\n同時文字部分呢,他也會到前端\n那這時候呢\n這時候呢\n就完成了整個的\n完成了整個的這個\nvoice agent的一個流程\n那麼為什麼我們又稱為agent呢\n就是\n不僅是voice,為什麼agent呢\n原因就是因為這裡面有R,就是大語言模型的介入\n它是一個智能的\n所以說我們就要\nvoice agent\n你既然是agent,那麼\n他可能也可以做RE,也就是檢索增加生成\n比如說\n我經常問的問題\n那麼我們可以把這個問題\n問的問題以及回答的過程把它傳到相量信息庫裡面\n那你問題問多了以後,那麼以後他在\n你再問他問題的,他就更懂你了,對吧\n更懂你了,知道你以前問過,他會知道怎麼再去回答你\n這個就是\n接入這個大語言模型的好處\n這個你還在這部分進行加強\n就這部分加強\n好\n那下面這邊有解釋\n好,我們看到總體的架構,我們再看一下這個\n還有完整流程,那這個裡面當然就更詳細了\n比如說瀏覽器這邊設計到麥克風採集\n對吧\n重採樣\n分幷\n這個技術名稱我們不需要長那麼細\n暫時\n然後二進制上行\n對吧\n然後\n這邊還會分發\n那麼這邊有VAD檢測,這個VAD是幹什麼\n你跟這個\nAI\n對話的時候你不經常可能會\n打斷他嗎?比如說\n他正在回答你某一個\n某一段話的時候你突然間打斷他了\n那如果是有打斷的話\n對吧\n有打斷的話\n那麼\n這時候會丟給LM\n對吧\n會切這個句子\n然後又貼貼的再返回\n對吧\n如果不打斷的話\n不打的話他就\n他就整個就是\n朝下去走\n這個意思\n整個朝下去走\n直接從這\nTTS\n這走\n然後就返回\n再返回到前端\n然後進行播放\n播放這個聲音\n中間發生的過程就是\nSTT code端\n你要走\nSTT的服務\n對吧\n你比如說\n阿里的呀\nDeepgram呀\n之類的\n那麼\n這個如果是LLM的話\n你可能接入OpenAI\n也可能接入Deepseek\n有可能接入Milimax\nKimi\n都可以\n好\n那麼\n那麼TTS這塊\n通過SSE的流\n來進行播放\n那麼SSE什麼呢\n這個是AI裡面的一個技術\n就是一個流式的\n返回的\n一個\n處理技術\n就是文字的一個流式\n同時還有音頻的\n最後到\n網頁這段它有音頻的一個流式\n所以這個是更詳細的一個完整流程\n這裡面涉及到好些技術\n好技術\n大家這些技術\n我們會一步步拆解\n給大家講的\n那麼\n我們對於一個\n總體的架構以及完整流程\n有所了解以後\n你才能搞懂\n我們後續要給大家講的\n比如說\n從Web端到Server端的整個的一個\nVoice Agent的一個整個流程\n那麼我們這個例子\n寫的不是很大\n但是通過這個例子\n你\n能對Voice Agent有一個\n完整的流程的一個理解\n並且\n它是正式辦理能跑起來的一個項目\n我們首先是把理論先搞清楚\n這裡面\n我們可以看到\n這裡面我們涉及的各種服務\n比如\nSTT\nIM\n這些\n大家都可以去採用不同的\nProvider\n有著供應商的服務\n不一定說是非要某一間的\n對吧\n!\n比如說如果說我們英文的我建議大家用Deepgram\nEleven Lab\n如果說是中文的建議大家用\n阿里的、騰訊的、火山引擎的，像這些\n包括MiniMax，像這些\n這樣的話是比較好的\n因為\n肯定是\n各自的模型\n它對中英文支持的出來的效果是不一樣的\n當然同時它還支持方言的\n比如說粵語\n對吧\n等等\n像這些方言的四川話\n普通話\n像方言普通話都支持的\n所以說這個\n你要不同的去\n另外還有一個\nVAD打斷\nVED打斷的\n不同的\n比如STT\n就是\n它叫\n實際上叫\nASR技術\n不同的\n廠商提供出來的\n質量也是不一樣的\n這個都去接\n然後我們這裡面當然就接了\n有阿里的\n有Deepgram\n這些\n已經有兩個比較好用的給大家接進來了\n告訴大家怎麼去接的\n好\n那麼關於它的這個總體的架構儀完整流程的我們就先講到這裡謝謝大家觀看",
  "language": "zh",
  "duration": 497.580438,
  "segments": [
    {
      "id": 0,
      "seek": 0,
      "start": 0.76,
      "end": 9.72,
      "text": "大家好,今天我們給大家講解這個Voice Agent的一個總體架構啊,以及它的這個完整流程。",
      "tokens": [
        50403,
        15248,
        11,
        12074,
        5884,
        17798,
        6868,
        11932,
        17278,
        6287,
        13339,
        27174,
        1546,
        8990,
        26575,
        23987,
        7360,
        114,
        43362,
        4905,
        11,
        40282,
        45224,
        6287,
        14128,
        27662,
        27854,
        29649,
        1543,
        50851
      ],
      "temperature": 0,
      "avg_logprob": -0.27545283635457357,
      "compression_ratio": 1.4195804195804196,
      "no_speech_prob": 9.389510870205786e-12,
      "source": "whisper"
    },
    {
      "id": 1,
      "seek": 0,
      "start": 10.5,
      "end": 20.48,
      "text": "我們首先了解一下什麼是Voice Agent。你看這個Voice,Voice就是語音的意思,就是語音智能體。也就是說不僅有語音,還有智能體。",
      "tokens": [
        50890,
        5884,
        36490,
        2289,
        17278,
        8861,
        7598,
        1541,
        13339,
        27174,
        1543,
        16529,
        6287,
        13339,
        11,
        13339,
        5620,
        31348,
        18034,
        1546,
        16697,
        11,
        5620,
        31348,
        18034,
        5094,
        118,
        8225,
        23987,
        1543,
        6404,
        26222,
        1960,
        9502,
        227,
        2412,
        31348,
        18034,
        11,
        15569,
        5094,
        118,
        8225,
        23987,
        1543,
        51389
      ],
      "temperature": 0,
      "avg_logprob": -0.27545283635457357,
      "compression_ratio": 1.4195804195804196,
      "no_speech_prob": 9.389510870205786e-12,
      "source": "whisper"
    },
    {
      "id": 2,
      "seek": 0,
      "start": 22.02,
      "end": 29.96,
      "text": "那最簡單的比如說你用豆包的語音功能，或者用LGBT的語音功能，你跟它對話。那這個就是最基本的一個語音智能體。",
      "tokens": [
        51466,
        4184,
        8661,
        31995,
        1546,
        36757,
        4622,
        2166,
        9254,
        45169,
        23305,
        1546,
        29801,
        11,
        31148,
        9254,
        43,
        8769,
        51,
        1546,
        29801,
        11,
        2166,
        9678,
        11284,
        2855,
        11103,
        1543,
        4184,
        6287,
        5620,
        8661,
        37946,
        1546,
        8990,
        31348,
        18034,
        5094,
        118,
        8225,
        23987,
        1543,
        51863
      ],
      "temperature": 0,
      "avg_logprob": -0.27545283635457357,
      "compression_ratio": 1.4195804195804196,
      "no_speech_prob": 9.389510870205786e-12,
      "source": "whisper",
      "text_original": "那最簡單的比如說你用豆包的書,或者用LGBT的書,你跟它對話。那這個就是最基本的一個語音智能體。",
      "correction_applied": true,
      "correction_source": "llm_correct_a_presegmentation"
    },
    {
      "id": 3,
      "seek": 3000,
      "start": 30.76,
      "end": 42.28,
      "text": "它裡面有哪些關鍵技術呢?有WebRTC，有STT，有TTS，有LLM，這幾個關鍵的技術。這幾個名詞呢，希望大家都去查一下。",
      "tokens": [
        50403,
        11284,
        32399,
        2412,
        17028,
        13824,
        14899,
        41668,
        113,
        32502,
        9890,
        241,
        6240,
        30,
        2412,
        4360,
        65,
        28987,
        11,
        2412,
        6840,
        35,
        11,
        2412,
        51,
        7327,
        11,
        2412,
        40,
        6324,
        11,
        2664,
        23575,
        3338,
        14899,
        41668,
        113,
        1546,
        32502,
        9890,
        241,
        1543,
        2664,
        23575,
        3338,
        15940,
        6716,
        252,
        6240,
        11,
        29955,
        46525,
        6734,
        42623,
        8861,
        1543,
        50979
      ],
      "temperature": 0,
      "avg_logprob": -0.19705173833583428,
      "compression_ratio": 1.311023622047244,
      "no_speech_prob": 1.0058152227765405e-11,
      "source": "whisper",
      "text_original": "它裡面有哪些關鍵技術呢?有Web RC,有STD,有TTS,有IIM,這幾個關鍵的技術。這幾個名詞呢,希望大家都去查一下。",
      "correction_applied": true,
      "correction_source": "llm_correct_a_presegmentation"
    },
    {
      "id": 4,
      "seek": 3000,
      "start": 42.8,
      "end": 55.6,
      "text": "那首先，WebRTC解決什麼?WebRTC解決就是說，音視頻的問題。你比如說你的聲音怎麼採集的，你聲音怎麼播放的，聲音怎麼傳輸的，這個是由WebRTC來解決的。",
      "tokens": [
        51005,
        4184,
        36490,
        11,
        4360,
        65,
        28987,
        17278,
        33540,
        7598,
        30,
        4360,
        65,
        28987,
        17278,
        33540,
        26222,
        11,
        18034,
        27333,
        29940,
        1546,
        17197,
        1543,
        2166,
        36757,
        4622,
        18961,
        29454,
        18034,
        11614,
        6900,
        94,
        26020,
        1546,
        11,
        2166,
        29454,
        18034,
        11614,
        49993,
        12744,
        1546,
        11,
        29454,
        18034,
        11614,
        40852,
        12290,
        116,
        1546,
        11,
        6287,
        1541,
        23786,
        4360,
        65,
        28987,
        3763,
        17278,
        33540,
        1546,
        1543,
        51645
      ],
      "temperature": 0,
      "avg_logprob": -0.19705173833583428,
      "compression_ratio": 1.311023622047244,
      "no_speech_prob": 1.0058152227765405e-11,
      "source": "whisper",
      "text_original": "那首先,Web RC解決什麼?Web RC解決就是說,音視頻的問題。你比如說你的聲音怎麼採集的,你聲音怎麼播放的,聲音怎麼傳輸的,這個是由Web RC來解決的。",
      "correction_applied": true,
      "correction_source": "llm_correct_a_presegmentation"
    },
    {
      "id": 5,
      "seek": 5560,
      "start": 56.120000000000005,
      "end": 58.42,
      "text": "那麼STT是什麼?STT就是",
      "tokens": [
        50391,
        20504,
        6840,
        35,
        35851,
        30,
        6840,
        35,
        5620,
        50506
      ],
      "temperature": 0,
      "avg_logprob": -0.1852966628911841,
      "compression_ratio": 1.44,
      "no_speech_prob": 1.9161490102947987e-11,
      "source": "whisper",
      "text_original": "那麼STD是什麼?STD就是",
      "correction_applied": true,
      "correction_source": "llm_correct_a_presegmentation"
    },
    {
      "id": 6,
      "seek": 5560,
      "start": 58.92,
      "end": 60.2,
      "text": "語音轉文字。",
      "tokens": [
        50531,
        31348,
        18034,
        32739,
        17174,
        22381,
        1543,
        50595
      ],
      "temperature": 0,
      "avg_logprob": -0.1852966628911841,
      "compression_ratio": 1.44,
      "no_speech_prob": 1.9161490102947987e-11,
      "source": "whisper"
    },
    {
      "id": 7,
      "seek": 5560,
      "start": 60.46,
      "end": 61.24,
      "text": "也就是說",
      "tokens": [
        50608,
        6404,
        26222,
        50647
      ],
      "temperature": 0,
      "avg_logprob": -0.1852966628911841,
      "compression_ratio": 1.44,
      "no_speech_prob": 1.9161490102947987e-11,
      "source": "whisper"
    },
    {
      "id": 8,
      "seek": 5560,
      "start": 61.74,
      "end": 67.12,
      "text": "你語音採集了以後,需要把它轉成文字,這個文字再交給大語音模型。",
      "tokens": [
        50672,
        2166,
        31348,
        18034,
        6900,
        94,
        26020,
        2289,
        3588,
        5661,
        11,
        35748,
        42061,
        32739,
        11336,
        17174,
        22381,
        11,
        6287,
        17174,
        22381,
        8623,
        28455,
        17798,
        3582,
        31348,
        18034,
        41908,
        39823,
        1543,
        50941
      ],
      "temperature": 0,
      "avg_logprob": -0.1852966628911841,
      "compression_ratio": 1.44,
      "no_speech_prob": 1.9161490102947987e-11,
      "source": "whisper"
    },
    {
      "id": 9,
      "seek": 5560,
      "start": 67.64,
      "end": 68.66,
      "text": "也就是LM",
      "tokens": [
        50967,
        6404,
        5620,
        43,
        44,
        51018
      ],
      "temperature": 0,
      "avg_logprob": -0.1852966628911841,
      "compression_ratio": 1.44,
      "no_speech_prob": 1.9161490102947987e-11,
      "source": "whisper"
    },
    {
      "id": 10,
      "seek": 5560,
      "start": 69.16,
      "end": 70.2,
      "text": "決定說什麼?",
      "tokens": [
        51043,
        33540,
        12088,
        4622,
        7598,
        30,
        51095
      ],
      "temperature": 0,
      "avg_logprob": -0.1852966628911841,
      "compression_ratio": 1.44,
      "no_speech_prob": 1.9161490102947987e-11,
      "source": "whisper"
    },
    {
      "id": 11,
      "seek": 5560,
      "start": 70.44,
      "end": 75.56,
      "text": "也就是說,LM是把文字拿到以後,它會加工,會思考,會返回。",
      "tokens": [
        51107,
        6404,
        26222,
        11,
        43,
        44,
        1541,
        16075,
        17174,
        22381,
        24351,
        4511,
        3588,
        5661,
        11,
        11284,
        6236,
        9990,
        23323,
        11,
        6236,
        8870,
        26504,
        11,
        6236,
        3316,
        242,
        8350,
        1543,
        51363
      ],
      "temperature": 0,
      "avg_logprob": -0.1852966628911841,
      "compression_ratio": 1.44,
      "no_speech_prob": 1.9161490102947987e-11,
      "source": "whisper"
    },
    {
      "id": 12,
      "seek": 5560,
      "start": 75.82,
      "end": 76.84,
      "text": "返回完以後,",
      "tokens": [
        51376,
        3316,
        242,
        8350,
        14128,
        3588,
        5661,
        11,
        51427
      ],
      "temperature": 0,
      "avg_logprob": -0.1852966628911841,
      "compression_ratio": 1.44,
      "no_speech_prob": 1.9161490102947987e-11,
      "source": "whisper"
    },
    {
      "id": 13,
      "seek": 5560,
      "start": 77.1,
      "end": 78.12,
      "text": "這不有文字了嗎?",
      "tokens": [
        51440,
        2664,
        1960,
        2412,
        17174,
        22381,
        2289,
        7434,
        30,
        51491
      ],
      "temperature": 0,
      "avg_logprob": -0.1852966628911841,
      "compression_ratio": 1.44,
      "no_speech_prob": 1.9161490102947987e-11,
      "source": "whisper"
    },
    {
      "id": 14,
      "seek": 5560,
      "start": 78.38,
      "end": 79.4,
      "text": "文字在",
      "tokens": [
        51504,
        17174,
        22381,
        3581,
        51555
      ],
      "temperature": 0,
      "avg_logprob": -0.1852966628911841,
      "compression_ratio": 1.44,
      "no_speech_prob": 1.9161490102947987e-11,
      "source": "whisper"
    },
    {
      "id": 15,
      "seek": 5560,
      "start": 79.66,
      "end": 81.46000000000001,
      "text": "第一個,它首先會發到",
      "tokens": [
        51568,
        40760,
        11,
        11284,
        36490,
        6236,
        14637,
        4511,
        51658
      ],
      "temperature": 0,
      "avg_logprob": -0.1852966628911841,
      "compression_ratio": 1.44,
      "no_speech_prob": 1.9161490102947987e-11,
      "source": "whisper"
    },
    {
      "id": 16,
      "seek": 8146,
      "start": 82.22,
      "end": 87.1,
      "text": "這個端上,也就是比如,一般我們叫前端,比如網頁裡面,或者手機裡面。",
      "tokens": [
        50403,
        6287,
        11957,
        107,
        5708,
        11,
        6404,
        5620,
        36757,
        11,
        2257,
        49640,
        5884,
        19855,
        8945,
        11957,
        107,
        11,
        36757,
        26352,
        8313,
        223,
        32399,
        11,
        31148,
        11389,
        17543,
        32399,
        1543,
        50647
      ],
      "temperature": 0,
      "avg_logprob": -0.2805343060880094,
      "compression_ratio": 1.3017543859649123,
      "no_speech_prob": 2.001397138295946e-11,
      "source": "whisper"
    },
    {
      "id": 17,
      "seek": 8146,
      "start": 88.11999999999999,
      "end": 94.0,
      "text": "然後不經意發,有可能要播語音,這個播語音就要把這個文字再轉成語音,這個過程叫TTS。",
      "tokens": [
        50698,
        10213,
        1960,
        15145,
        9042,
        14637,
        11,
        2412,
        16657,
        4275,
        49993,
        31348,
        18034,
        11,
        6287,
        49993,
        31348,
        18034,
        3111,
        4275,
        16075,
        6287,
        17174,
        22381,
        8623,
        32739,
        11336,
        31348,
        18034,
        11,
        6287,
        8816,
        29649,
        19855,
        51,
        7327,
        1543,
        50992
      ],
      "temperature": 0,
      "avg_logprob": -0.2805343060880094,
      "compression_ratio": 1.3017543859649123,
      "no_speech_prob": 2.001397138295946e-11,
      "source": "whisper"
    },
    {
      "id": 18,
      "seek": 8146,
      "start": 95.02,
      "end": 97.58,
      "text": "所以說,這裡面有幾個關鍵概念,叫Web RC",
      "tokens": [
        51043,
        7239,
        4622,
        11,
        18771,
        8833,
        2412,
        23575,
        3338,
        14899,
        41668,
        113,
        26181,
        33679,
        11,
        19855,
        4360,
        65,
        28987,
        51171
      ],
      "temperature": 0,
      "avg_logprob": -0.2805343060880094,
      "compression_ratio": 1.3017543859649123,
      "no_speech_prob": 2.001397138295946e-11,
      "source": "whisper"
    },
    {
      "id": 19,
      "seek": 8146,
      "start": 97.83999999999999,
      "end": 98.61999999999999,
      "text": "STD",
      "tokens": [
        51184,
        6840,
        35,
        51223
      ],
      "temperature": 0,
      "avg_logprob": -0.2805343060880094,
      "compression_ratio": 1.3017543859649123,
      "no_speech_prob": 2.001397138295946e-11,
      "source": "whisper"
    },
    {
      "id": 20,
      "seek": 8146,
      "start": 98.86,
      "end": 99.63999999999999,
      "text": "TTS",
      "tokens": [
        51235,
        51,
        7327,
        51274
      ],
      "temperature": 0,
      "avg_logprob": -0.2805343060880094,
      "compression_ratio": 1.3017543859649123,
      "no_speech_prob": 2.001397138295946e-11,
      "source": "whisper"
    },
    {
      "id": 21,
      "seek": 8146,
      "start": 99.89999999999999,
      "end": 100.66,
      "text": "LM",
      "tokens": [
        51287,
        43,
        44,
        51325
      ],
      "temperature": 0,
      "avg_logprob": -0.2805343060880094,
      "compression_ratio": 1.3017543859649123,
      "no_speech_prob": 2.001397138295946e-11,
      "source": "whisper"
    },
    {
      "id": 22,
      "seek": 8146,
      "start": 101.41999999999999,
      "end": 101.67999999999999,
      "text": "對吧",
      "tokens": [
        51363,
        47905,
        51376
      ],
      "temperature": 0,
      "avg_logprob": -0.2805343060880094,
      "compression_ratio": 1.3017543859649123,
      "no_speech_prob": 2.001397138295946e-11,
      "source": "whisper"
    },
    {
      "id": 23,
      "seek": 8146,
      "start": 102.19999999999999,
      "end": 103.97999999999999,
      "text": "好,我們來看一下它總體架構",
      "tokens": [
        51402,
        2131,
        11,
        5884,
        3763,
        28324,
        11284,
        26575,
        23987,
        7360,
        114,
        43362,
        51491
      ],
      "temperature": 0,
      "avg_logprob": -0.2805343060880094,
      "compression_ratio": 1.3017543859649123,
      "no_speech_prob": 2.001397138295946e-11,
      "source": "whisper"
    },
    {
      "id": 24,
      "seek": 8146,
      "start": 104.75999999999999,
      "end": 105.25999999999999,
      "text": "首先是",
      "tokens": [
        51530,
        36490,
        1541,
        51555
      ],
      "temperature": 0,
      "avg_logprob": -0.2805343060880094,
      "compression_ratio": 1.3017543859649123,
      "no_speech_prob": 2.001397138295946e-11,
      "source": "whisper"
    },
    {
      "id": 25,
      "seek": 8146,
      "start": 106.03999999999999,
      "end": 108.33999999999999,
      "text": "前端,前端呢,它有一個",
      "tokens": [
        51594,
        8945,
        11957,
        107,
        11,
        8945,
        11957,
        107,
        6240,
        11,
        11284,
        2412,
        8990,
        51709
      ],
      "temperature": 0,
      "avg_logprob": -0.2805343060880094,
      "compression_ratio": 1.3017543859649123,
      "no_speech_prob": 2.001397138295946e-11,
      "source": "whisper"
    },
    {
      "id": 26,
      "seek": 8146,
      "start": 108.86,
      "end": 109.88,
      "text": "比如說可能是瀏覽器",
      "tokens": [
        51735,
        36757,
        4622,
        16657,
        1541,
        163,
        222,
        237,
        2888,
        121,
        34386,
        51786
      ],
      "temperature": 0,
      "avg_logprob": -0.2805343060880094,
      "compression_ratio": 1.3017543859649123,
      "no_speech_prob": 2.001397138295946e-11,
      "source": "whisper"
    },
    {
      "id": 27,
      "seek": 10988,
      "start": 109.88,
      "end": 111.67999999999999,
      "text": "有可能是App",
      "tokens": [
        50365,
        2412,
        16657,
        1541,
        9132,
        50455
      ],
      "temperature": 0,
      "avg_logprob": -0.19052840567923882,
      "compression_ratio": 1.3462897526501767,
      "no_speech_prob": 1.547524608458417e-11,
      "source": "whisper"
    },
    {
      "id": 28,
      "seek": 10988,
      "start": 111.92,
      "end": 113.19999999999999,
      "text": "甚至是一個硬件啊,對吧",
      "tokens": [
        50467,
        17556,
        20844,
        1541,
        8990,
        38114,
        105,
        20485,
        4905,
        11,
        47905,
        50531
      ],
      "temperature": 0,
      "avg_logprob": -0.19052840567923882,
      "compression_ratio": 1.3462897526501767,
      "no_speech_prob": 1.547524608458417e-11,
      "source": "whisper"
    },
    {
      "id": 29,
      "seek": 10988,
      "start": 113.46,
      "end": 115.0,
      "text": "好,那麼麥克風採集",
      "tokens": [
        50544,
        2131,
        11,
        20504,
        3994,
        98,
        24881,
        22713,
        6900,
        94,
        26020,
        50621
      ],
      "temperature": 0,
      "avg_logprob": -0.19052840567923882,
      "compression_ratio": 1.3462897526501767,
      "no_speech_prob": 1.547524608458417e-11,
      "source": "whisper"
    },
    {
      "id": 30,
      "seek": 10988,
      "start": 115.52,
      "end": 116.28,
      "text": "那音頻",
      "tokens": [
        50647,
        4184,
        18034,
        29940,
        50685
      ],
      "temperature": 0,
      "avg_logprob": -0.19052840567923882,
      "compression_ratio": 1.3462897526501767,
      "no_speech_prob": 1.547524608458417e-11,
      "source": "whisper"
    },
    {
      "id": 31,
      "seek": 10988,
      "start": 116.8,
      "end": 119.86,
      "text": "同時音頻可能會播放,然後屏幕上可能有文字",
      "tokens": [
        50711,
        13089,
        6611,
        18034,
        29940,
        16657,
        6236,
        49993,
        12744,
        11,
        10213,
        9636,
        237,
        40211,
        5708,
        16657,
        2412,
        17174,
        22381,
        50864
      ],
      "temperature": 0,
      "avg_logprob": -0.19052840567923882,
      "compression_ratio": 1.3462897526501767,
      "no_speech_prob": 1.547524608458417e-11,
      "source": "whisper"
    },
    {
      "id": 32,
      "seek": 10988,
      "start": 120.11999999999999,
      "end": 121.14,
      "text": "大家用這個",
      "tokens": [
        50877,
        6868,
        9254,
        6287,
        50928
      ],
      "temperature": 0,
      "avg_logprob": -0.19052840567923882,
      "compression_ratio": 1.3462897526501767,
      "no_speech_prob": 1.547524608458417e-11,
      "source": "whisper"
    },
    {
      "id": 33,
      "seek": 10988,
      "start": 121.39999999999999,
      "end": 123.96,
      "text": "用豆包的時候,你在跟它對話的時候,它不經常有這個",
      "tokens": [
        50941,
        9254,
        45169,
        23305,
        20643,
        11,
        41410,
        9678,
        11284,
        2855,
        11103,
        20643,
        11,
        11284,
        1960,
        15145,
        11279,
        2412,
        6287,
        51069
      ],
      "temperature": 0,
      "avg_logprob": -0.19052840567923882,
      "compression_ratio": 1.3462897526501767,
      "no_speech_prob": 1.547524608458417e-11,
      "source": "whisper"
    },
    {
      "id": 34,
      "seek": 10988,
      "start": 125.24,
      "end": 127.8,
      "text": "會有文字顯示嗎?另外一個它會播放語音",
      "tokens": [
        51133,
        6236,
        2412,
        17174,
        22381,
        46605,
        25696,
        7434,
        30,
        26202,
        8990,
        11284,
        6236,
        49993,
        12744,
        31348,
        18034,
        51261
      ],
      "temperature": 0,
      "avg_logprob": -0.19052840567923882,
      "compression_ratio": 1.3462897526501767,
      "no_speech_prob": 1.547524608458417e-11,
      "source": "whisper"
    },
    {
      "id": 35,
      "seek": 10988,
      "start": 128.82,
      "end": 130.1,
      "text": "那麼音頻會上行",
      "tokens": [
        51312,
        20504,
        18034,
        29940,
        6236,
        5708,
        8082,
        51376
      ],
      "temperature": 0,
      "avg_logprob": -0.19052840567923882,
      "compression_ratio": 1.3462897526501767,
      "no_speech_prob": 1.547524608458417e-11,
      "source": "whisper"
    },
    {
      "id": 36,
      "seek": 10988,
      "start": 130.35999999999999,
      "end": 131.12,
      "text": "到後端",
      "tokens": [
        51389,
        4511,
        5661,
        11957,
        107,
        51427
      ],
      "temperature": 0,
      "avg_logprob": -0.19052840567923882,
      "compression_ratio": 1.3462897526501767,
      "no_speech_prob": 1.547524608458417e-11,
      "source": "whisper"
    },
    {
      "id": 37,
      "seek": 10988,
      "start": 131.64,
      "end": 132.16,
      "text": "後端呢",
      "tokens": [
        51453,
        5661,
        11957,
        107,
        6240,
        51479
      ],
      "temperature": 0,
      "avg_logprob": -0.19052840567923882,
      "compression_ratio": 1.3462897526501767,
      "no_speech_prob": 1.547524608458417e-11,
      "source": "whisper"
    },
    {
      "id": 38,
      "seek": 10988,
      "start": 132.66,
      "end": 135.48,
      "text": "它這裡面呢,它會比如說,它一般是Web RC",
      "tokens": [
        51504,
        11284,
        18771,
        8833,
        6240,
        11,
        11284,
        6236,
        36757,
        4622,
        11,
        11284,
        2257,
        49640,
        1541,
        4360,
        65,
        28987,
        51645
      ],
      "temperature": 0,
      "avg_logprob": -0.19052840567923882,
      "compression_ratio": 1.3462897526501767,
      "no_speech_prob": 1.547524608458417e-11,
      "source": "whisper"
    },
    {
      "id": 39,
      "seek": 10988,
      "start": 136.5,
      "end": 137.51999999999998,
      "text": "處理啊,處理完",
      "tokens": [
        51696,
        39289,
        13876,
        4905,
        11,
        39289,
        13876,
        14128,
        51747
      ],
      "temperature": 0,
      "avg_logprob": -0.19052840567923882,
      "compression_ratio": 1.3462897526501767,
      "no_speech_prob": 1.547524608458417e-11,
      "source": "whisper"
    },
    {
      "id": 40,
      "seek": 13752,
      "start": 138.04000000000002,
      "end": 138.8,
      "text": "以後呢,它會",
      "tokens": [
        50391,
        3588,
        5661,
        6240,
        11,
        11284,
        6236,
        50429
      ],
      "temperature": 0,
      "avg_logprob": -0.2580117271060035,
      "compression_ratio": 1.3633217993079585,
      "no_speech_prob": 2.407556966288471e-11,
      "source": "whisper"
    },
    {
      "id": 41,
      "seek": 13752,
      "start": 139.06,
      "end": 140.84,
      "text": "把這個聲音提取出來",
      "tokens": [
        50442,
        16075,
        6287,
        29454,
        18034,
        20949,
        29436,
        29741,
        50531
      ],
      "temperature": 0,
      "avg_logprob": -0.2580117271060035,
      "compression_ratio": 1.3633217993079585,
      "no_speech_prob": 2.407556966288471e-11,
      "source": "whisper"
    },
    {
      "id": 42,
      "seek": 13752,
      "start": 141.10000000000002,
      "end": 142.64000000000001,
      "text": "然後把這個聲音呢,丟給",
      "tokens": [
        50544,
        10213,
        16075,
        6287,
        29454,
        18034,
        6240,
        11,
        940,
        253,
        17798,
        50621
      ],
      "temperature": 0,
      "avg_logprob": -0.2580117271060035,
      "compression_ratio": 1.3633217993079585,
      "no_speech_prob": 2.407556966288471e-11,
      "source": "whisper"
    },
    {
      "id": 43,
      "seek": 13752,
      "start": 142.9,
      "end": 143.66000000000003,
      "text": "STT",
      "tokens": [
        50634,
        6840,
        51,
        50672
      ],
      "temperature": 0,
      "avg_logprob": -0.2580117271060035,
      "compression_ratio": 1.3633217993079585,
      "no_speech_prob": 2.407556966288471e-11,
      "source": "whisper"
    },
    {
      "id": 44,
      "seek": 13752,
      "start": 143.92000000000002,
      "end": 144.68,
      "text": "STT呢",
      "tokens": [
        50685,
        6840,
        51,
        6240,
        50723
      ],
      "temperature": 0,
      "avg_logprob": -0.2580117271060035,
      "compression_ratio": 1.3633217993079585,
      "no_speech_prob": 2.407556966288471e-11,
      "source": "whisper"
    },
    {
      "id": 45,
      "seek": 13752,
      "start": 144.94,
      "end": 145.46,
      "text": "呃",
      "tokens": [
        50736,
        3606,
        225,
        50762
      ],
      "temperature": 0,
      "avg_logprob": -0.2580117271060035,
      "compression_ratio": 1.3633217993079585,
      "no_speech_prob": 2.407556966288471e-11,
      "source": "whisper"
    },
    {
      "id": 46,
      "seek": 13752,
      "start": 145.96,
      "end": 148.52,
      "text": "有幾種,一種是你本地的這種服務,對吧",
      "tokens": [
        50787,
        2412,
        23575,
        16516,
        11,
        2257,
        16516,
        1541,
        2166,
        8802,
        10928,
        1546,
        33005,
        27408,
        39517,
        11,
        47905,
        50915
      ],
      "temperature": 0,
      "avg_logprob": -0.2580117271060035,
      "compression_ratio": 1.3633217993079585,
      "no_speech_prob": 2.407556966288471e-11,
      "source": "whisper"
    },
    {
      "id": 47,
      "seek": 13752,
      "start": 148.78,
      "end": 150.06,
      "text": "你可能從歐拉馬來接入",
      "tokens": [
        50928,
        2166,
        16657,
        21068,
        4287,
        238,
        29717,
        29098,
        3763,
        14468,
        14028,
        50992
      ],
      "temperature": 0,
      "avg_logprob": -0.2580117271060035,
      "compression_ratio": 1.3633217993079585,
      "no_speech_prob": 2.407556966288471e-11,
      "source": "whisper"
    },
    {
      "id": 48,
      "seek": 13752,
      "start": 150.32000000000002,
      "end": 152.36,
      "text": "第二種呢,有可能比如說你是阿里的",
      "tokens": [
        51005,
        19693,
        16516,
        6240,
        11,
        2412,
        16657,
        36757,
        4622,
        32526,
        13594,
        15759,
        1546,
        51107
      ],
      "temperature": 0,
      "avg_logprob": -0.2580117271060035,
      "compression_ratio": 1.3633217993079585,
      "no_speech_prob": 2.407556966288471e-11,
      "source": "whisper"
    },
    {
      "id": 49,
      "seek": 13752,
      "start": 152.62,
      "end": 152.88,
      "text": "對吧",
      "tokens": [
        51120,
        47905,
        51133
      ],
      "temperature": 0,
      "avg_logprob": -0.2580117271060035,
      "compression_ratio": 1.3633217993079585,
      "no_speech_prob": 2.407556966288471e-11,
      "source": "whisper"
    },
    {
      "id": 50,
      "seek": 13752,
      "start": 153.14000000000001,
      "end": 154.42000000000002,
      "text": "也有可能是騰訊的",
      "tokens": [
        51146,
        6404,
        2412,
        16657,
        1541,
        38168,
        108,
        5396,
        232,
        1546,
        51210
      ],
      "temperature": 0,
      "avg_logprob": -0.2580117271060035,
      "compression_ratio": 1.3633217993079585,
      "no_speech_prob": 2.407556966288471e-11,
      "source": "whisper"
    },
    {
      "id": 51,
      "seek": 13752,
      "start": 154.68,
      "end": 156.72,
      "text": "也有可能是火山引擎的",
      "tokens": [
        51223,
        6404,
        2412,
        16657,
        1541,
        26171,
        24113,
        41301,
        13167,
        236,
        1546,
        51325
      ],
      "temperature": 0,
      "avg_logprob": -0.2580117271060035,
      "compression_ratio": 1.3633217993079585,
      "no_speech_prob": 2.407556966288471e-11,
      "source": "whisper"
    },
    {
      "id": 52,
      "seek": 13752,
      "start": 156.98000000000002,
      "end": 158.0,
      "text": "也可能是MiniMax",
      "tokens": [
        51338,
        6404,
        16657,
        1541,
        44,
        3812,
        36025,
        51389
      ],
      "temperature": 0,
      "avg_logprob": -0.2580117271060035,
      "compression_ratio": 1.3633217993079585,
      "no_speech_prob": 2.407556966288471e-11,
      "source": "whisper"
    },
    {
      "id": 53,
      "seek": 13752,
      "start": 158.26000000000002,
      "end": 158.52,
      "text": "是吧",
      "tokens": [
        51402,
        1541,
        6062,
        51415
      ],
      "temperature": 0,
      "avg_logprob": -0.2580117271060035,
      "compression_ratio": 1.3633217993079585,
      "no_speech_prob": 2.407556966288471e-11,
      "source": "whisper"
    },
    {
      "id": 54,
      "seek": 13752,
      "start": 159.54000000000002,
      "end": 160.56,
      "text": "STT,MiniMax",
      "tokens": [
        51466,
        6840,
        51,
        11,
        44,
        3812,
        36025,
        51517
      ],
      "temperature": 0,
      "avg_logprob": -0.2580117271060035,
      "compression_ratio": 1.3633217993079585,
      "no_speech_prob": 2.407556966288471e-11,
      "source": "whisper"
    },
    {
      "id": 55,
      "seek": 13752,
      "start": 160.82000000000002,
      "end": 161.32000000000002,
      "text": "目前",
      "tokens": [
        51530,
        39004,
        51555
      ],
      "temperature": 0,
      "avg_logprob": -0.2580117271060035,
      "compression_ratio": 1.3633217993079585,
      "no_speech_prob": 2.407556966288471e-11,
      "source": "whisper"
    },
    {
      "id": 56,
      "seek": 13752,
      "start": 161.84,
      "end": 162.36,
      "text": "不一定有",
      "tokens": [
        51581,
        1960,
        24272,
        2412,
        51607
      ],
      "temperature": 0,
      "avg_logprob": -0.2580117271060035,
      "compression_ratio": 1.3633217993079585,
      "no_speech_prob": 2.407556966288471e-11,
      "source": "whisper"
    },
    {
      "id": 57,
      "seek": 13752,
      "start": 162.60000000000002,
      "end": 163.88,
      "text": "然後你可能Deepgram",
      "tokens": [
        51619,
        10213,
        2166,
        16657,
        11089,
        595,
        1342,
        51683
      ],
      "temperature": 0,
      "avg_logprob": -0.2580117271060035,
      "compression_ratio": 1.3633217993079585,
      "no_speech_prob": 2.407556966288471e-11,
      "source": "whisper"
    },
    {
      "id": 58,
      "seek": 13752,
      "start": 164.14000000000001,
      "end": 165.42000000000002,
      "text": "有可能是ElevenLab",
      "tokens": [
        51696,
        2412,
        16657,
        1541,
        36,
        306,
        553,
        37880,
        51760
      ],
      "temperature": 0,
      "avg_logprob": -0.2580117271060035,
      "compression_ratio": 1.3633217993079585,
      "no_speech_prob": 2.407556966288471e-11,
      "source": "whisper"
    },
    {
      "id": 59,
      "seek": 13752,
      "start": 165.68,
      "end": 167.22,
      "text": "這裡面就有中英文",
      "tokens": [
        51773,
        18771,
        8833,
        3111,
        2412,
        5975,
        27869,
        17174,
        51850
      ],
      "temperature": 0,
      "avg_logprob": -0.2580117271060035,
      "compression_ratio": 1.3633217993079585,
      "no_speech_prob": 2.407556966288471e-11,
      "source": "whisper"
    },
    {
      "id": 60,
      "seek": 16722,
      "start": 167.48,
      "end": 169.78,
      "text": "這裡面我們主要以中英文給大家解釋就行了",
      "tokens": [
        50378,
        18771,
        8833,
        5884,
        13557,
        4275,
        3588,
        5975,
        27869,
        17174,
        17798,
        6868,
        17278,
        5873,
        233,
        3111,
        46606,
        50493
      ],
      "temperature": 0,
      "avg_logprob": -0.20930714674398931,
      "compression_ratio": 1.4087591240875912,
      "no_speech_prob": 2.1481701834025202e-11,
      "source": "whisper"
    },
    {
      "id": 61,
      "seek": 16722,
      "start": 170.3,
      "end": 172.6,
      "text": "那麼由它轉動文字又返回給",
      "tokens": [
        50519,
        20504,
        23786,
        11284,
        32739,
        12572,
        17174,
        22381,
        17047,
        3316,
        242,
        8350,
        17798,
        50634
      ],
      "temperature": 0,
      "avg_logprob": -0.20930714674398931,
      "compression_ratio": 1.4087591240875912,
      "no_speech_prob": 2.1481701834025202e-11,
      "source": "whisper"
    },
    {
      "id": 62,
      "seek": 16722,
      "start": 172.85999999999999,
      "end": 177.2,
      "text": "那麼由它轉動文字又返回給這個後端，後端這塊它又會調這個LLM，也就是調的大語言模型",
      "tokens": [
        50647,
        6287,
        5661,
        11957,
        107,
        11,
        5661,
        11957,
        107,
        2664,
        26268,
        11284,
        17047,
        6236,
        29327,
        1546,
        6287,
        6324,
        46,
        11,
        6404,
        5620,
        29327,
        1546,
        3582,
        14812,
        41908,
        39823,
        50864
      ],
      "temperature": 0,
      "avg_logprob": -0.20930714674398931,
      "compression_ratio": 1.4087591240875912,
      "no_speech_prob": 2.1481701834025202e-11,
      "source": "whisper",
      "text_original": "這個後端,後端這塊它又會掉的這個IMO,也就是掉的大元模型",
      "correction_applied": true,
      "correction_source": "llm_correct_a_presegmentation"
    },
    {
      "id": 63,
      "seek": 16722,
      "start": 177.46,
      "end": 179.76,
      "text": "這裡面選擇就多了,你比如說你可以用Deepseek",
      "tokens": [
        50877,
        18771,
        8833,
        21625,
        13167,
        229,
        3111,
        6392,
        2289,
        11,
        2166,
        36757,
        4622,
        42766,
        9254,
        11089,
        595,
        405,
        916,
        50992
      ],
      "temperature": 0,
      "avg_logprob": -0.20930714674398931,
      "compression_ratio": 1.4087591240875912,
      "no_speech_prob": 2.1481701834025202e-11,
      "source": "whisper"
    },
    {
      "id": 64,
      "seek": 16722,
      "start": 180.02,
      "end": 181.3,
      "text": "可以用這個",
      "tokens": [
        51005,
        6723,
        9254,
        6287,
        51069
      ],
      "temperature": 0,
      "avg_logprob": -0.20930714674398931,
      "compression_ratio": 1.4087591240875912,
      "no_speech_prob": 2.1481701834025202e-11,
      "source": "whisper"
    },
    {
      "id": 65,
      "seek": 16722,
      "start": 182.06,
      "end": 184.12,
      "text": "MiniMax,可以用這個Kimi",
      "tokens": [
        51107,
        44,
        3812,
        36025,
        11,
        6723,
        9254,
        6287,
        42,
        10121,
        51210
      ],
      "temperature": 0,
      "avg_logprob": -0.20930714674398931,
      "compression_ratio": 1.4087591240875912,
      "no_speech_prob": 2.1481701834025202e-11,
      "source": "whisper"
    },
    {
      "id": 66,
      "seek": 16722,
      "start": 184.38,
      "end": 186.94,
      "text": "可以用這個OpenAI,對吧",
      "tokens": [
        51223,
        6723,
        9254,
        6287,
        45569,
        48698,
        11,
        47905,
        51351
      ],
      "temperature": 0,
      "avg_logprob": -0.20930714674398931,
      "compression_ratio": 1.4087591240875912,
      "no_speech_prob": 2.1481701834025202e-11,
      "source": "whisper"
    },
    {
      "id": 67,
      "seek": 16722,
      "start": 187.44,
      "end": 188.46,
      "text": "你可以用",
      "tokens": [
        51376,
        42766,
        9254,
        51427
      ],
      "temperature": 0,
      "avg_logprob": -0.20930714674398931,
      "compression_ratio": 1.4087591240875912,
      "no_speech_prob": 2.1481701834025202e-11,
      "source": "whisper"
    },
    {
      "id": 68,
      "seek": 16722,
      "start": 188.72,
      "end": 190.52,
      "text": "這個很多很多模型啊,都可以用",
      "tokens": [
        51440,
        6287,
        20778,
        20778,
        41908,
        39823,
        4905,
        11,
        7182,
        6723,
        9254,
        51530
      ],
      "temperature": 0,
      "avg_logprob": -0.20930714674398931,
      "compression_ratio": 1.4087591240875912,
      "no_speech_prob": 2.1481701834025202e-11,
      "source": "whisper"
    },
    {
      "id": 69,
      "seek": 16722,
      "start": 191.02,
      "end": 191.54,
      "text": "那",
      "tokens": [
        51555,
        4184,
        51581
      ],
      "temperature": 0,
      "avg_logprob": -0.20930714674398931,
      "compression_ratio": 1.4087591240875912,
      "no_speech_prob": 2.1481701834025202e-11,
      "source": "whisper"
    },
    {
      "id": 70,
      "seek": 16722,
      "start": 191.8,
      "end": 193.57999999999998,
      "text": "丟給大元模型,大元模型會",
      "tokens": [
        51594,
        940,
        253,
        17798,
        3582,
        14812,
        41908,
        39823,
        11,
        3582,
        14812,
        41908,
        39823,
        6236,
        51683
      ],
      "temperature": 0,
      "avg_logprob": -0.20930714674398931,
      "compression_ratio": 1.4087591240875912,
      "no_speech_prob": 2.1481701834025202e-11,
      "source": "whisper"
    },
    {
      "id": 71,
      "seek": 19358,
      "start": 193.84,
      "end": 195.62,
      "text": "對你的話進行加工",
      "tokens": [
        50378,
        2855,
        18961,
        43822,
        18214,
        8082,
        9990,
        23323,
        50467
      ],
      "temperature": 0,
      "avg_logprob": -0.21067471313476563,
      "compression_ratio": 1.2831858407079646,
      "no_speech_prob": 2.1814834660904836e-11,
      "source": "whisper",
      "text_original": "對你的畫進行加工",
      "correction_applied": true,
      "correction_source": "llm_correct_a_presegmentation"
    },
    {
      "id": 72,
      "seek": 19358,
      "start": 195.88000000000002,
      "end": 197.16000000000003,
      "text": "加工完以後",
      "tokens": [
        50480,
        9990,
        23323,
        14128,
        3588,
        5661,
        50544
      ],
      "temperature": 0,
      "avg_logprob": -0.21067471313476563,
      "compression_ratio": 1.2831858407079646,
      "no_speech_prob": 2.1814834660904836e-11,
      "source": "whisper"
    },
    {
      "id": 73,
      "seek": 19358,
      "start": 197.42000000000002,
      "end": 198.96,
      "text": "他比如說你今天問了一個",
      "tokens": [
        50557,
        5000,
        36757,
        4622,
        2166,
        12074,
        11361,
        2289,
        8990,
        50634
      ],
      "temperature": 0,
      "avg_logprob": -0.21067471313476563,
      "compression_ratio": 1.2831858407079646,
      "no_speech_prob": 2.1814834660904836e-11,
      "source": "whisper"
    },
    {
      "id": 74,
      "seek": 19358,
      "start": 199.22,
      "end": 201.0,
      "text": "舉個例子,什麼是人工智能對吧",
      "tokens": [
        50647,
        42828,
        3338,
        17797,
        7626,
        11,
        7598,
        1541,
        4035,
        23323,
        5094,
        118,
        8225,
        47905,
        50736
      ],
      "temperature": 0,
      "avg_logprob": -0.21067471313476563,
      "compression_ratio": 1.2831858407079646,
      "no_speech_prob": 2.1814834660904836e-11,
      "source": "whisper"
    },
    {
      "id": 75,
      "seek": 19358,
      "start": 201.26000000000002,
      "end": 202.28,
      "text": "他會肯定回答",
      "tokens": [
        50749,
        5000,
        6236,
        14356,
        107,
        12088,
        8350,
        28223,
        50800
      ],
      "temperature": 0,
      "avg_logprob": -0.21067471313476563,
      "compression_ratio": 1.2831858407079646,
      "no_speech_prob": 2.1814834660904836e-11,
      "source": "whisper"
    },
    {
      "id": 76,
      "seek": 19358,
      "start": 203.06,
      "end": 204.58,
      "text": "一堆這個內容嘛,然後",
      "tokens": [
        50839,
        2257,
        10726,
        228,
        6287,
        28472,
        25750,
        20722,
        11,
        10213,
        50915
      ],
      "temperature": 0,
      "avg_logprob": -0.21067471313476563,
      "compression_ratio": 1.2831858407079646,
      "no_speech_prob": 2.1814834660904836e-11,
      "source": "whisper"
    },
    {
      "id": 77,
      "seek": 19358,
      "start": 205.10000000000002,
      "end": 205.86,
      "text": "他返回完",
      "tokens": [
        50941,
        5000,
        3316,
        242,
        8350,
        14128,
        50979
      ],
      "temperature": 0,
      "avg_logprob": -0.21067471313476563,
      "compression_ratio": 1.2831858407079646,
      "no_speech_prob": 2.1814834660904836e-11,
      "source": "whisper"
    },
    {
      "id": 78,
      "seek": 19358,
      "start": 206.12,
      "end": 207.92000000000002,
      "text": "返回他以後,然後後端呢",
      "tokens": [
        50992,
        3316,
        242,
        8350,
        5000,
        3588,
        5661,
        11,
        10213,
        5661,
        11957,
        107,
        6240,
        51082
      ],
      "temperature": 0,
      "avg_logprob": -0.21067471313476563,
      "compression_ratio": 1.2831858407079646,
      "no_speech_prob": 2.1814834660904836e-11,
      "source": "whisper"
    },
    {
      "id": 79,
      "seek": 19358,
      "start": 208.68,
      "end": 208.94,
      "text": "又",
      "tokens": [
        51120,
        17047,
        51133
      ],
      "temperature": 0,
      "avg_logprob": -0.21067471313476563,
      "compression_ratio": 1.2831858407079646,
      "no_speech_prob": 2.1814834660904836e-11,
      "source": "whisper"
    },
    {
      "id": 80,
      "seek": 19358,
      "start": 209.20000000000002,
      "end": 209.96,
      "text": "緊接著",
      "tokens": [
        51146,
        30172,
        14468,
        19382,
        51184
      ],
      "temperature": 0,
      "avg_logprob": -0.21067471313476563,
      "compression_ratio": 1.2831858407079646,
      "no_speech_prob": 2.1814834660904836e-11,
      "source": "whisper"
    },
    {
      "id": 81,
      "seek": 19358,
      "start": 210.22000000000003,
      "end": 212.26000000000002,
      "text": "調這個TTS模型,這個TTS模型也有",
      "tokens": [
        51197,
        32943,
        6287,
        51,
        7327,
        41908,
        39823,
        11,
        6287,
        51,
        7327,
        41908,
        39823,
        6404,
        2412,
        51299
      ],
      "temperature": 0,
      "avg_logprob": -0.21067471313476563,
      "compression_ratio": 1.2831858407079646,
      "no_speech_prob": 2.1814834660904836e-11,
      "source": "whisper"
    },
    {
      "id": 82,
      "seek": 19358,
      "start": 212.52,
      "end": 214.58,
      "text": "比如火山引擎呢,阿里騰訊呢",
      "tokens": [
        51312,
        36757,
        26171,
        24113,
        41301,
        13167,
        236,
        6240,
        11,
        13594,
        15759,
        38168,
        108,
        5396,
        232,
        6240,
        51415
      ],
      "temperature": 0,
      "avg_logprob": -0.21067471313476563,
      "compression_ratio": 1.2831858407079646,
      "no_speech_prob": 2.1814834660904836e-11,
      "source": "whisper"
    },
    {
      "id": 83,
      "seek": 21458,
      "start": 214.84,
      "end": 217.66000000000003,
      "text": "對吧,Deepgram,ElevenLav,他都會把",
      "tokens": [
        50378,
        47905,
        11,
        11089,
        595,
        1342,
        11,
        36,
        306,
        553,
        43,
        706,
        11,
        5000,
        7182,
        6236,
        16075,
        50519
      ],
      "temperature": 0,
      "avg_logprob": -0.3078677261931987,
      "compression_ratio": 1.3806228373702423,
      "no_speech_prob": 2.0916712806240412e-11,
      "source": "whisper"
    },
    {
      "id": 84,
      "seek": 21458,
      "start": 218.16000000000003,
      "end": 219.18,
      "text": "文字再轉成語音",
      "tokens": [
        50544,
        17174,
        22381,
        8623,
        32739,
        11336,
        31348,
        18034,
        50595
      ],
      "temperature": 0,
      "avg_logprob": -0.3078677261931987,
      "compression_ratio": 1.3806228373702423,
      "no_speech_prob": 2.0916712806240412e-11,
      "source": "whisper"
    },
    {
      "id": 85,
      "seek": 21458,
      "start": 219.44000000000003,
      "end": 220.72000000000003,
      "text": "轉成語音呢,他會有一個",
      "tokens": [
        50608,
        32739,
        11336,
        31348,
        18034,
        6240,
        11,
        5000,
        6236,
        2412,
        8990,
        50672
      ],
      "temperature": 0,
      "avg_logprob": -0.3078677261931987,
      "compression_ratio": 1.3806228373702423,
      "no_speech_prob": 2.0916712806240412e-11,
      "source": "whisper"
    },
    {
      "id": 86,
      "seek": 21458,
      "start": 220.98000000000002,
      "end": 224.56,
      "text": "實時的流，這個實時流呢，再通過WebRTC技術再傳給前端",
      "tokens": [
        50685,
        10376,
        10376,
        1546,
        27854,
        11,
        6287,
        10376,
        10376,
        27854,
        6240,
        11,
        8623,
        19550,
        8816,
        4360,
        65,
        5879,
        36,
        32502,
        9890,
        241,
        8623,
        40852,
        17798,
        8945,
        11957,
        107,
        50864
      ],
      "temperature": 0,
      "avg_logprob": -0.3078677261931987,
      "compression_ratio": 1.3806228373702423,
      "no_speech_prob": 2.0916712806240412e-11,
      "source": "whisper",
      "text_original": "實實的流,這個實實流呢,再通過WebSEE技術再傳給前端",
      "correction_applied": true,
      "correction_source": "llm_correct_a_presegmentation"
    },
    {
      "id": 87,
      "seek": 21458,
      "start": 224.82000000000002,
      "end": 226.10000000000002,
      "text": "前端把聲音播放出來",
      "tokens": [
        50877,
        8945,
        11957,
        107,
        16075,
        29454,
        18034,
        49993,
        12744,
        29741,
        50941
      ],
      "temperature": 0,
      "avg_logprob": -0.3078677261931987,
      "compression_ratio": 1.3806228373702423,
      "no_speech_prob": 2.0916712806240412e-11,
      "source": "whisper"
    },
    {
      "id": 88,
      "seek": 21458,
      "start": 226.62,
      "end": 228.66000000000003,
      "text": "同時文字部分呢,他也會到前端",
      "tokens": [
        50967,
        13089,
        6611,
        17174,
        22381,
        32174,
        6240,
        11,
        5000,
        6404,
        6236,
        4511,
        8945,
        11957,
        107,
        51069
      ],
      "temperature": 0,
      "avg_logprob": -0.3078677261931987,
      "compression_ratio": 1.3806228373702423,
      "no_speech_prob": 2.0916712806240412e-11,
      "source": "whisper"
    },
    {
      "id": 89,
      "seek": 21458,
      "start": 229.18,
      "end": 229.94,
      "text": "那這時候呢",
      "tokens": [
        51095,
        4184,
        2664,
        14010,
        6240,
        51133
      ],
      "temperature": 0,
      "avg_logprob": -0.3078677261931987,
      "compression_ratio": 1.3806228373702423,
      "no_speech_prob": 2.0916712806240412e-11,
      "source": "whisper"
    },
    {
      "id": 90,
      "seek": 21458,
      "start": 230.96,
      "end": 231.48000000000002,
      "text": "這時候呢",
      "tokens": [
        51184,
        2664,
        14010,
        6240,
        51210
      ],
      "temperature": 0,
      "avg_logprob": -0.3078677261931987,
      "compression_ratio": 1.3806228373702423,
      "no_speech_prob": 2.0916712806240412e-11,
      "source": "whisper"
    },
    {
      "id": 91,
      "seek": 21458,
      "start": 231.74,
      "end": 233.26000000000002,
      "text": "就完成了整個的",
      "tokens": [
        51223,
        3111,
        41509,
        2289,
        27662,
        3338,
        1546,
        51299
      ],
      "temperature": 0,
      "avg_logprob": -0.3078677261931987,
      "compression_ratio": 1.3806228373702423,
      "no_speech_prob": 2.0916712806240412e-11,
      "source": "whisper"
    },
    {
      "id": 92,
      "seek": 21458,
      "start": 234.3,
      "end": 235.32000000000002,
      "text": "完成了整個的這個",
      "tokens": [
        51351,
        41509,
        2289,
        27662,
        3338,
        1546,
        6287,
        51402
      ],
      "temperature": 0,
      "avg_logprob": -0.3078677261931987,
      "compression_ratio": 1.3806228373702423,
      "no_speech_prob": 2.0916712806240412e-11,
      "source": "whisper"
    },
    {
      "id": 93,
      "seek": 21458,
      "start": 235.58,
      "end": 236.60000000000002,
      "text": "voice agent的一個流程",
      "tokens": [
        51415,
        42268,
        9461,
        1546,
        8990,
        27854,
        29649,
        51466
      ],
      "temperature": 0,
      "avg_logprob": -0.3078677261931987,
      "compression_ratio": 1.3806228373702423,
      "no_speech_prob": 2.0916712806240412e-11,
      "source": "whisper"
    },
    {
      "id": 94,
      "seek": 21458,
      "start": 237.36,
      "end": 239.16000000000003,
      "text": "那麼為什麼我們又稱為agent呢",
      "tokens": [
        51504,
        20504,
        18987,
        5884,
        17047,
        49606,
        6344,
        559,
        317,
        6240,
        51594
      ],
      "temperature": 0,
      "avg_logprob": -0.3078677261931987,
      "compression_ratio": 1.3806228373702423,
      "no_speech_prob": 2.0916712806240412e-11,
      "source": "whisper"
    },
    {
      "id": 95,
      "seek": 21458,
      "start": 239.92000000000002,
      "end": 240.18,
      "text": "就是",
      "tokens": [
        51632,
        5620,
        51645
      ],
      "temperature": 0,
      "avg_logprob": -0.3078677261931987,
      "compression_ratio": 1.3806228373702423,
      "no_speech_prob": 2.0916712806240412e-11,
      "source": "whisper"
    },
    {
      "id": 96,
      "seek": 21458,
      "start": 240.94,
      "end": 242.22000000000003,
      "text": "不僅是voice,為什麼agent呢",
      "tokens": [
        51683,
        1960,
        9502,
        227,
        1541,
        42268,
        11,
        18987,
        559,
        317,
        6240,
        51747
      ],
      "temperature": 0,
      "avg_logprob": -0.3078677261931987,
      "compression_ratio": 1.3806228373702423,
      "no_speech_prob": 2.0916712806240412e-11,
      "source": "whisper"
    },
    {
      "id": 97,
      "seek": 24222,
      "start": 242.22,
      "end": 245.3,
      "text": "原因就是因為這裡面有R,就是大語言模型的介入",
      "tokens": [
        50365,
        42577,
        5620,
        11471,
        18771,
        8833,
        2412,
        49,
        11,
        5620,
        3582,
        31348,
        12009,
        41908,
        39823,
        1546,
        30312,
        14028,
        50519
      ],
      "temperature": 0,
      "avg_logprob": -0.19754871208510713,
      "compression_ratio": 1.4414414414414414,
      "no_speech_prob": 2.0629683725181813e-11,
      "source": "whisper"
    },
    {
      "id": 98,
      "seek": 24222,
      "start": 245.54,
      "end": 246.32,
      "text": "它是一個智能的",
      "tokens": [
        50531,
        11284,
        1541,
        8990,
        5094,
        118,
        8225,
        1546,
        50570
      ],
      "temperature": 0,
      "avg_logprob": -0.19754871208510713,
      "compression_ratio": 1.4414414414414414,
      "no_speech_prob": 2.0629683725181813e-11,
      "source": "whisper"
    },
    {
      "id": 99,
      "seek": 24222,
      "start": 247.08,
      "end": 248.36,
      "text": "所以說我們就要",
      "tokens": [
        50608,
        7239,
        4622,
        5884,
        3111,
        4275,
        50672
      ],
      "temperature": 0,
      "avg_logprob": -0.19754871208510713,
      "compression_ratio": 1.4414414414414414,
      "no_speech_prob": 2.0629683725181813e-11,
      "source": "whisper"
    },
    {
      "id": 100,
      "seek": 24222,
      "start": 248.88,
      "end": 249.9,
      "text": "voice agent",
      "tokens": [
        50698,
        42268,
        9461,
        50749
      ],
      "temperature": 0,
      "avg_logprob": -0.19754871208510713,
      "compression_ratio": 1.4414414414414414,
      "no_speech_prob": 2.0629683725181813e-11,
      "source": "whisper"
    },
    {
      "id": 101,
      "seek": 24222,
      "start": 250.16,
      "end": 251.44,
      "text": "你既然是agent,那麼",
      "tokens": [
        50762,
        2166,
        4479,
        95,
        5823,
        1541,
        559,
        317,
        11,
        20504,
        50826
      ],
      "temperature": 0,
      "avg_logprob": -0.19754871208510713,
      "compression_ratio": 1.4414414414414414,
      "no_speech_prob": 2.0629683725181813e-11,
      "source": "whisper"
    },
    {
      "id": 102,
      "seek": 24222,
      "start": 251.7,
      "end": 255.02,
      "text": "他可能也可以做RE,也就是檢索增加生成",
      "tokens": [
        50839,
        5000,
        16657,
        47935,
        10907,
        3850,
        11,
        6404,
        5620,
        40741,
        95,
        7732,
        95,
        24228,
        252,
        9990,
        8244,
        11336,
        51005
      ],
      "temperature": 0,
      "avg_logprob": -0.19754871208510713,
      "compression_ratio": 1.4414414414414414,
      "no_speech_prob": 2.0629683725181813e-11,
      "source": "whisper"
    },
    {
      "id": 103,
      "seek": 24222,
      "start": 255.28,
      "end": 255.78,
      "text": "比如說",
      "tokens": [
        51018,
        36757,
        4622,
        51043
      ],
      "temperature": 0,
      "avg_logprob": -0.19754871208510713,
      "compression_ratio": 1.4414414414414414,
      "no_speech_prob": 2.0629683725181813e-11,
      "source": "whisper"
    },
    {
      "id": 104,
      "seek": 24222,
      "start": 256.04,
      "end": 257.06,
      "text": "我經常問的問題",
      "tokens": [
        51056,
        1654,
        15145,
        11279,
        11361,
        1546,
        17197,
        51107
      ],
      "temperature": 0,
      "avg_logprob": -0.19754871208510713,
      "compression_ratio": 1.4414414414414414,
      "no_speech_prob": 2.0629683725181813e-11,
      "source": "whisper"
    },
    {
      "id": 105,
      "seek": 24222,
      "start": 257.32,
      "end": 258.6,
      "text": "那麼我們可以把這個問題",
      "tokens": [
        51120,
        20504,
        5884,
        6723,
        16075,
        6287,
        17197,
        51184
      ],
      "temperature": 0,
      "avg_logprob": -0.19754871208510713,
      "compression_ratio": 1.4414414414414414,
      "no_speech_prob": 2.0629683725181813e-11,
      "source": "whisper"
    },
    {
      "id": 106,
      "seek": 24222,
      "start": 258.86,
      "end": 261.94,
      "text": "問的問題以及回答的過程把它傳到相量信息庫裡面",
      "tokens": [
        51197,
        11361,
        1546,
        17197,
        40282,
        8350,
        28223,
        1546,
        8816,
        29649,
        42061,
        40852,
        4511,
        15106,
        26748,
        17665,
        26460,
        6346,
        104,
        32399,
        51351
      ],
      "temperature": 0,
      "avg_logprob": -0.19754871208510713,
      "compression_ratio": 1.4414414414414414,
      "no_speech_prob": 2.0629683725181813e-11,
      "source": "whisper"
    },
    {
      "id": 107,
      "seek": 24222,
      "start": 262.18,
      "end": 265.26,
      "text": "那你問題問多了以後,那麼以後他在",
      "tokens": [
        51363,
        4184,
        2166,
        17197,
        11361,
        6392,
        2289,
        3588,
        5661,
        11,
        20504,
        3588,
        5661,
        5000,
        3581,
        51517
      ],
      "temperature": 0,
      "avg_logprob": -0.19754871208510713,
      "compression_ratio": 1.4414414414414414,
      "no_speech_prob": 2.0629683725181813e-11,
      "source": "whisper"
    },
    {
      "id": 108,
      "seek": 24222,
      "start": 265.78,
      "end": 268.08,
      "text": "你再問他問題的,他就更懂你了,對吧",
      "tokens": [
        51543,
        2166,
        8623,
        11361,
        5000,
        17197,
        1546,
        11,
        5000,
        3111,
        19002,
        29624,
        2166,
        2289,
        11,
        47905,
        51658
      ],
      "temperature": 0,
      "avg_logprob": -0.19754871208510713,
      "compression_ratio": 1.4414414414414414,
      "no_speech_prob": 2.0629683725181813e-11,
      "source": "whisper"
    },
    {
      "id": 109,
      "seek": 24222,
      "start": 268.34,
      "end": 271.66,
      "text": "更懂你了,知道你以前問過,他會知道怎麼再去回答你",
      "tokens": [
        51671,
        19002,
        29624,
        2166,
        2289,
        11,
        7758,
        2166,
        44996,
        11361,
        8816,
        11,
        5000,
        6236,
        7758,
        11614,
        8623,
        6734,
        8350,
        28223,
        2166,
        51837
      ],
      "temperature": 0,
      "avg_logprob": -0.19754871208510713,
      "compression_ratio": 1.4414414414414414,
      "no_speech_prob": 2.0629683725181813e-11,
      "source": "whisper"
    },
    {
      "id": 110,
      "seek": 27166,
      "start": 272.18,
      "end": 272.94,
      "text": "這個就是",
      "tokens": [
        50391,
        6287,
        5620,
        50429
      ],
      "temperature": 0,
      "avg_logprob": -0.26263707997847574,
      "compression_ratio": 1.3655913978494623,
      "no_speech_prob": 2.1979588288312257e-11,
      "source": "whisper"
    },
    {
      "id": 111,
      "seek": 27166,
      "start": 273.46000000000004,
      "end": 276.02000000000004,
      "text": "接入這個大語言模型的好處",
      "tokens": [
        50455,
        14468,
        14028,
        6287,
        3582,
        31348,
        12009,
        41908,
        39823,
        1546,
        2131,
        39289,
        50583
      ],
      "temperature": 0,
      "avg_logprob": -0.26263707997847574,
      "compression_ratio": 1.3655913978494623,
      "no_speech_prob": 2.1979588288312257e-11,
      "source": "whisper"
    },
    {
      "id": 112,
      "seek": 27166,
      "start": 276.26000000000005,
      "end": 278.58000000000004,
      "text": "這個你還在這部分進行加強",
      "tokens": [
        50595,
        6287,
        2166,
        7824,
        3581,
        2664,
        32174,
        18214,
        8082,
        9990,
        22752,
        50711
      ],
      "temperature": 0,
      "avg_logprob": -0.26263707997847574,
      "compression_ratio": 1.3655913978494623,
      "no_speech_prob": 2.1979588288312257e-11,
      "source": "whisper"
    },
    {
      "id": 113,
      "seek": 27166,
      "start": 278.82000000000005,
      "end": 279.6,
      "text": "就這部分加強",
      "tokens": [
        50723,
        3111,
        2664,
        32174,
        9990,
        22752,
        50762
      ],
      "temperature": 0,
      "avg_logprob": -0.26263707997847574,
      "compression_ratio": 1.3655913978494623,
      "no_speech_prob": 2.1979588288312257e-11,
      "source": "whisper"
    },
    {
      "id": 114,
      "seek": 27166,
      "start": 280.36,
      "end": 280.88000000000005,
      "text": "好",
      "tokens": [
        50800,
        2131,
        50826
      ],
      "temperature": 0,
      "avg_logprob": -0.26263707997847574,
      "compression_ratio": 1.3655913978494623,
      "no_speech_prob": 2.1979588288312257e-11,
      "source": "whisper"
    },
    {
      "id": 115,
      "seek": 27166,
      "start": 281.14000000000004,
      "end": 282.66,
      "text": "那下面這邊有解釋",
      "tokens": [
        50839,
        4184,
        47150,
        22821,
        2412,
        17278,
        5873,
        233,
        50915
      ],
      "temperature": 0,
      "avg_logprob": -0.26263707997847574,
      "compression_ratio": 1.3655913978494623,
      "no_speech_prob": 2.1979588288312257e-11,
      "source": "whisper"
    },
    {
      "id": 116,
      "seek": 27166,
      "start": 283.44,
      "end": 285.74,
      "text": "好,我們看到總體的架構,我們再看一下這個",
      "tokens": [
        50954,
        2131,
        11,
        5884,
        18032,
        26575,
        23987,
        1546,
        7360,
        114,
        43362,
        11,
        5884,
        8623,
        28324,
        6287,
        51069
      ],
      "temperature": 0,
      "avg_logprob": -0.26263707997847574,
      "compression_ratio": 1.3655913978494623,
      "no_speech_prob": 2.1979588288312257e-11,
      "source": "whisper"
    },
    {
      "id": 117,
      "seek": 27166,
      "start": 286.0,
      "end": 288.56,
      "text": "還有完整流程,那這個裡面當然就更詳細了",
      "tokens": [
        51082,
        15569,
        14128,
        27662,
        27854,
        29649,
        11,
        4184,
        6287,
        32399,
        21707,
        3111,
        19002,
        6716,
        111,
        41588,
        2289,
        51210
      ],
      "temperature": 0,
      "avg_logprob": -0.26263707997847574,
      "compression_ratio": 1.3655913978494623,
      "no_speech_prob": 2.1979588288312257e-11,
      "source": "whisper"
    },
    {
      "id": 118,
      "seek": 27166,
      "start": 288.82000000000005,
      "end": 291.88,
      "text": "比如說瀏覽器這邊設計到麥克風採集",
      "tokens": [
        51223,
        36757,
        4622,
        163,
        222,
        237,
        2888,
        121,
        34386,
        22821,
        39035,
        30114,
        4511,
        3994,
        98,
        24881,
        22713,
        6900,
        94,
        26020,
        51376
      ],
      "temperature": 0,
      "avg_logprob": -0.26263707997847574,
      "compression_ratio": 1.3655913978494623,
      "no_speech_prob": 2.1979588288312257e-11,
      "source": "whisper"
    },
    {
      "id": 119,
      "seek": 27166,
      "start": 292.14000000000004,
      "end": 292.66,
      "text": "對吧",
      "tokens": [
        51389,
        47905,
        51415
      ],
      "temperature": 0,
      "avg_logprob": -0.26263707997847574,
      "compression_ratio": 1.3655913978494623,
      "no_speech_prob": 2.1979588288312257e-11,
      "source": "whisper"
    },
    {
      "id": 120,
      "seek": 27166,
      "start": 294.18,
      "end": 294.96000000000004,
      "text": "重採樣",
      "tokens": [
        51491,
        12624,
        6900,
        94,
        6180,
        51530
      ],
      "temperature": 0,
      "avg_logprob": -0.26263707997847574,
      "compression_ratio": 1.3655913978494623,
      "no_speech_prob": 2.1979588288312257e-11,
      "source": "whisper"
    },
    {
      "id": 121,
      "seek": 27166,
      "start": 295.22,
      "end": 295.72,
      "text": "分幷",
      "tokens": [
        51543,
        6627,
        5873,
        251,
        51568
      ],
      "temperature": 0,
      "avg_logprob": -0.26263707997847574,
      "compression_ratio": 1.3655913978494623,
      "no_speech_prob": 2.1979588288312257e-11,
      "source": "whisper",
      "text_original": "分針",
      "correction_applied": true,
      "correction_source": "llm_correct_a_presegmentation"
    },
    {
      "id": 122,
      "seek": 27166,
      "start": 296.24,
      "end": 298.02000000000004,
      "text": "這個技術名稱我們不需要長那麼細",
      "tokens": [
        51594,
        6287,
        32502,
        9890,
        241,
        15940,
        49606,
        5884,
        1960,
        35748,
        15353,
        20504,
        41588,
        51683
      ],
      "temperature": 0,
      "avg_logprob": -0.26263707997847574,
      "compression_ratio": 1.3655913978494623,
      "no_speech_prob": 2.1979588288312257e-11,
      "source": "whisper"
    },
    {
      "id": 123,
      "seek": 27166,
      "start": 298.28000000000003,
      "end": 298.8,
      "text": "暫時",
      "tokens": [
        51696,
        23826,
        104,
        6611,
        51722
      ],
      "temperature": 0,
      "avg_logprob": -0.26263707997847574,
      "compression_ratio": 1.3655913978494623,
      "no_speech_prob": 2.1979588288312257e-11,
      "source": "whisper"
    },
    {
      "id": 124,
      "seek": 27166,
      "start": 299.3,
      "end": 300.58000000000004,
      "text": "然後二進制上行",
      "tokens": [
        51747,
        10213,
        11217,
        19117,
        22381,
        5708,
        8082,
        51811
      ],
      "temperature": 0,
      "avg_logprob": -0.26263707997847574,
      "compression_ratio": 1.3655913978494623,
      "no_speech_prob": 2.1979588288312257e-11,
      "source": "whisper",
      "text_original": "然後二金字上行",
      "correction_applied": true,
      "correction_source": "llm_correct_a_presegmentation"
    },
    {
      "id": 125,
      "seek": 30058,
      "start": 300.84,
      "end": 301.59999999999997,
      "text": "對吧",
      "tokens": [
        50378,
        47905,
        50416
      ],
      "temperature": 0,
      "avg_logprob": -0.2056061862266227,
      "compression_ratio": 1.3967611336032388,
      "no_speech_prob": 2.4404757728579973e-11,
      "source": "whisper"
    },
    {
      "id": 126,
      "seek": 30058,
      "start": 301.85999999999996,
      "end": 302.62,
      "text": "然後",
      "tokens": [
        50429,
        10213,
        50467
      ],
      "temperature": 0,
      "avg_logprob": -0.2056061862266227,
      "compression_ratio": 1.3967611336032388,
      "no_speech_prob": 2.4404757728579973e-11,
      "source": "whisper"
    },
    {
      "id": 127,
      "seek": 30058,
      "start": 302.88,
      "end": 303.9,
      "text": "這邊還會分發",
      "tokens": [
        50480,
        22821,
        7824,
        6236,
        6627,
        14637,
        50531
      ],
      "temperature": 0,
      "avg_logprob": -0.2056061862266227,
      "compression_ratio": 1.3967611336032388,
      "no_speech_prob": 2.4404757728579973e-11,
      "source": "whisper"
    },
    {
      "id": 128,
      "seek": 30058,
      "start": 304.94,
      "end": 307.74,
      "text": "那麼這邊有VAD檢測,這個VAD是幹什麼",
      "tokens": [
        50583,
        20504,
        22821,
        2412,
        53,
        6112,
        40741,
        95,
        9592,
        105,
        11,
        6287,
        53,
        6112,
        1541,
        29932,
        7598,
        50723
      ],
      "temperature": 0,
      "avg_logprob": -0.2056061862266227,
      "compression_ratio": 1.3967611336032388,
      "no_speech_prob": 2.4404757728579973e-11,
      "source": "whisper"
    },
    {
      "id": 129,
      "seek": 30058,
      "start": 308.26,
      "end": 309.53999999999996,
      "text": "你跟這個",
      "tokens": [
        50749,
        2166,
        9678,
        6287,
        50813
      ],
      "temperature": 0,
      "avg_logprob": -0.2056061862266227,
      "compression_ratio": 1.3967611336032388,
      "no_speech_prob": 2.4404757728579973e-11,
      "source": "whisper"
    },
    {
      "id": 130,
      "seek": 30058,
      "start": 310.06,
      "end": 310.56,
      "text": "AI",
      "tokens": [
        50839,
        48698,
        50864
      ],
      "temperature": 0,
      "avg_logprob": -0.2056061862266227,
      "compression_ratio": 1.3967611336032388,
      "no_speech_prob": 2.4404757728579973e-11,
      "source": "whisper"
    },
    {
      "id": 131,
      "seek": 30058,
      "start": 310.82,
      "end": 312.35999999999996,
      "text": "對話的時候你不經常可能會",
      "tokens": [
        50877,
        2855,
        11103,
        20643,
        2166,
        1960,
        15145,
        11279,
        16657,
        6236,
        50954
      ],
      "temperature": 0,
      "avg_logprob": -0.2056061862266227,
      "compression_ratio": 1.3967611336032388,
      "no_speech_prob": 2.4404757728579973e-11,
      "source": "whisper"
    },
    {
      "id": 132,
      "seek": 30058,
      "start": 312.62,
      "end": 313.9,
      "text": "打斷他嗎?比如說",
      "tokens": [
        50967,
        12467,
        4307,
        115,
        5000,
        7434,
        30,
        36757,
        4622,
        51031
      ],
      "temperature": 0,
      "avg_logprob": -0.2056061862266227,
      "compression_ratio": 1.3967611336032388,
      "no_speech_prob": 2.4404757728579973e-11,
      "source": "whisper"
    },
    {
      "id": 133,
      "seek": 30058,
      "start": 314.14,
      "end": 315.41999999999996,
      "text": "他正在回答你某一個",
      "tokens": [
        51043,
        5000,
        15789,
        3581,
        8350,
        28223,
        2166,
        17238,
        238,
        8990,
        51107
      ],
      "temperature": 0,
      "avg_logprob": -0.2056061862266227,
      "compression_ratio": 1.3967611336032388,
      "no_speech_prob": 2.4404757728579973e-11,
      "source": "whisper"
    },
    {
      "id": 134,
      "seek": 30058,
      "start": 315.68,
      "end": 317.47999999999996,
      "text": "某一段話的時候你突然間打斷他了",
      "tokens": [
        51120,
        17238,
        238,
        2257,
        28427,
        11103,
        20643,
        2166,
        40859,
        5823,
        11016,
        12467,
        4307,
        115,
        5000,
        2289,
        51210
      ],
      "temperature": 0,
      "avg_logprob": -0.2056061862266227,
      "compression_ratio": 1.3967611336032388,
      "no_speech_prob": 2.4404757728579973e-11,
      "source": "whisper"
    },
    {
      "id": 135,
      "seek": 30058,
      "start": 317.74,
      "end": 319.52,
      "text": "那如果是有打斷的話",
      "tokens": [
        51223,
        4184,
        13119,
        1541,
        2412,
        12467,
        4307,
        115,
        21358,
        51312
      ],
      "temperature": 0,
      "avg_logprob": -0.2056061862266227,
      "compression_ratio": 1.3967611336032388,
      "no_speech_prob": 2.4404757728579973e-11,
      "source": "whisper"
    },
    {
      "id": 136,
      "seek": 30058,
      "start": 319.78,
      "end": 320.29999999999995,
      "text": "對吧",
      "tokens": [
        51325,
        47905,
        51351
      ],
      "temperature": 0,
      "avg_logprob": -0.2056061862266227,
      "compression_ratio": 1.3967611336032388,
      "no_speech_prob": 2.4404757728579973e-11,
      "source": "whisper"
    },
    {
      "id": 137,
      "seek": 30058,
      "start": 320.53999999999996,
      "end": 321.32,
      "text": "有打斷的話",
      "tokens": [
        51363,
        2412,
        12467,
        4307,
        115,
        21358,
        51402
      ],
      "temperature": 0,
      "avg_logprob": -0.2056061862266227,
      "compression_ratio": 1.3967611336032388,
      "no_speech_prob": 2.4404757728579973e-11,
      "source": "whisper"
    },
    {
      "id": 138,
      "seek": 30058,
      "start": 321.58,
      "end": 322.08,
      "text": "那麼",
      "tokens": [
        51415,
        20504,
        51440
      ],
      "temperature": 0,
      "avg_logprob": -0.2056061862266227,
      "compression_ratio": 1.3967611336032388,
      "no_speech_prob": 2.4404757728579973e-11,
      "source": "whisper"
    },
    {
      "id": 139,
      "seek": 30058,
      "start": 322.86,
      "end": 324.38,
      "text": "這時候會丟給LM",
      "tokens": [
        51479,
        2664,
        14010,
        6236,
        940,
        253,
        17798,
        43,
        44,
        51555
      ],
      "temperature": 0,
      "avg_logprob": -0.2056061862266227,
      "compression_ratio": 1.3967611336032388,
      "no_speech_prob": 2.4404757728579973e-11,
      "source": "whisper"
    },
    {
      "id": 140,
      "seek": 30058,
      "start": 324.64,
      "end": 325.15999999999997,
      "text": "對吧",
      "tokens": [
        51568,
        47905,
        51594
      ],
      "temperature": 0,
      "avg_logprob": -0.2056061862266227,
      "compression_ratio": 1.3967611336032388,
      "no_speech_prob": 2.4404757728579973e-11,
      "source": "whisper"
    },
    {
      "id": 141,
      "seek": 30058,
      "start": 325.41999999999996,
      "end": 326.18,
      "text": "會切這個句子",
      "tokens": [
        51607,
        6236,
        23632,
        6287,
        34592,
        7626,
        51645
      ],
      "temperature": 0,
      "avg_logprob": -0.2056061862266227,
      "compression_ratio": 1.3967611336032388,
      "no_speech_prob": 2.4404757728579973e-11,
      "source": "whisper"
    },
    {
      "id": 142,
      "seek": 30058,
      "start": 326.44,
      "end": 327.46,
      "text": "然後又貼貼的再返回",
      "tokens": [
        51658,
        10213,
        17047,
        11561,
        120,
        11561,
        120,
        1546,
        8623,
        3316,
        242,
        8350,
        51709
      ],
      "temperature": 0,
      "avg_logprob": -0.2056061862266227,
      "compression_ratio": 1.3967611336032388,
      "no_speech_prob": 2.4404757728579973e-11,
      "source": "whisper"
    },
    {
      "id": 143,
      "seek": 32746,
      "start": 327.46,
      "end": 328.21999999999997,
      "text": "對吧",
      "tokens": [
        50365,
        47905,
        50403
      ],
      "temperature": 0,
      "avg_logprob": -0.3403220135232677,
      "compression_ratio": 1.2959183673469388,
      "no_speech_prob": 2.5257860039595847e-11,
      "source": "whisper"
    },
    {
      "id": 144,
      "seek": 32746,
      "start": 328.47999999999996,
      "end": 330.02,
      "text": "如果不打斷的話",
      "tokens": [
        50416,
        13119,
        1960,
        12467,
        4307,
        115,
        21358,
        50493
      ],
      "temperature": 0,
      "avg_logprob": -0.3403220135232677,
      "compression_ratio": 1.2959183673469388,
      "no_speech_prob": 2.5257860039595847e-11,
      "source": "whisper"
    },
    {
      "id": 145,
      "seek": 32746,
      "start": 330.28,
      "end": 331.03999999999996,
      "text": "不打的話他就",
      "tokens": [
        50506,
        1960,
        12467,
        21358,
        5000,
        3111,
        50544
      ],
      "temperature": 0,
      "avg_logprob": -0.3403220135232677,
      "compression_ratio": 1.2959183673469388,
      "no_speech_prob": 2.5257860039595847e-11,
      "source": "whisper"
    },
    {
      "id": 146,
      "seek": 32746,
      "start": 331.56,
      "end": 332.84,
      "text": "他就整個就是",
      "tokens": [
        50570,
        5000,
        3111,
        27662,
        3338,
        5620,
        50634
      ],
      "temperature": 0,
      "avg_logprob": -0.3403220135232677,
      "compression_ratio": 1.2959183673469388,
      "no_speech_prob": 2.5257860039595847e-11,
      "source": "whisper"
    },
    {
      "id": 147,
      "seek": 32746,
      "start": 333.85999999999996,
      "end": 335.14,
      "text": "朝下去走",
      "tokens": [
        50685,
        46610,
        34473,
        9575,
        50749
      ],
      "temperature": 0,
      "avg_logprob": -0.3403220135232677,
      "compression_ratio": 1.2959183673469388,
      "no_speech_prob": 2.5257860039595847e-11,
      "source": "whisper"
    },
    {
      "id": 148,
      "seek": 32746,
      "start": 335.4,
      "end": 336.15999999999997,
      "text": "這個意思",
      "tokens": [
        50762,
        6287,
        16697,
        50800
      ],
      "temperature": 0,
      "avg_logprob": -0.3403220135232677,
      "compression_ratio": 1.2959183673469388,
      "no_speech_prob": 2.5257860039595847e-11,
      "source": "whisper"
    },
    {
      "id": 149,
      "seek": 32746,
      "start": 336.68,
      "end": 337.44,
      "text": "整個朝下去走",
      "tokens": [
        50826,
        27662,
        3338,
        46610,
        34473,
        9575,
        50864
      ],
      "temperature": 0,
      "avg_logprob": -0.3403220135232677,
      "compression_ratio": 1.2959183673469388,
      "no_speech_prob": 2.5257860039595847e-11,
      "source": "whisper"
    },
    {
      "id": 150,
      "seek": 32746,
      "start": 337.7,
      "end": 338.46,
      "text": "直接從這",
      "tokens": [
        50877,
        43297,
        21068,
        2664,
        50915
      ],
      "temperature": 0,
      "avg_logprob": -0.3403220135232677,
      "compression_ratio": 1.2959183673469388,
      "no_speech_prob": 2.5257860039595847e-11,
      "source": "whisper"
    },
    {
      "id": 151,
      "seek": 32746,
      "start": 338.71999999999997,
      "end": 339.23999999999995,
      "text": "TTS",
      "tokens": [
        50928,
        51,
        11844,
        50954
      ],
      "temperature": 0,
      "avg_logprob": -0.3403220135232677,
      "compression_ratio": 1.2959183673469388,
      "no_speech_prob": 2.5257860039595847e-11,
      "source": "whisper",
      "text_original": "TDS",
      "correction_applied": true,
      "correction_source": "llm_correct_a_presegmentation"
    },
    {
      "id": 152,
      "seek": 32746,
      "start": 339.74,
      "end": 340.0,
      "text": "這走",
      "tokens": [
        50979,
        2664,
        9575,
        50992
      ],
      "temperature": 0,
      "avg_logprob": -0.3403220135232677,
      "compression_ratio": 1.2959183673469388,
      "no_speech_prob": 2.5257860039595847e-11,
      "source": "whisper"
    },
    {
      "id": 153,
      "seek": 32746,
      "start": 340.26,
      "end": 341.28,
      "text": "然後就返回",
      "tokens": [
        51005,
        10213,
        3111,
        3316,
        242,
        8350,
        51056
      ],
      "temperature": 0,
      "avg_logprob": -0.3403220135232677,
      "compression_ratio": 1.2959183673469388,
      "no_speech_prob": 2.5257860039595847e-11,
      "source": "whisper"
    },
    {
      "id": 154,
      "seek": 32746,
      "start": 341.53999999999996,
      "end": 342.56,
      "text": "再返回到前端",
      "tokens": [
        51069,
        8623,
        3316,
        242,
        8350,
        4511,
        8945,
        11957,
        107,
        51120
      ],
      "temperature": 0,
      "avg_logprob": -0.3403220135232677,
      "compression_ratio": 1.2959183673469388,
      "no_speech_prob": 2.5257860039595847e-11,
      "source": "whisper"
    },
    {
      "id": 155,
      "seek": 32746,
      "start": 343.34,
      "end": 344.09999999999997,
      "text": "然後進行播放",
      "tokens": [
        51159,
        10213,
        18214,
        8082,
        49993,
        12744,
        51197
      ],
      "temperature": 0,
      "avg_logprob": -0.3403220135232677,
      "compression_ratio": 1.2959183673469388,
      "no_speech_prob": 2.5257860039595847e-11,
      "source": "whisper"
    },
    {
      "id": 156,
      "seek": 32746,
      "start": 344.35999999999996,
      "end": 345.12,
      "text": "播放這個聲音",
      "tokens": [
        51210,
        49993,
        12744,
        6287,
        29454,
        18034,
        51248
      ],
      "temperature": 0,
      "avg_logprob": -0.3403220135232677,
      "compression_ratio": 1.2959183673469388,
      "no_speech_prob": 2.5257860039595847e-11,
      "source": "whisper"
    },
    {
      "id": 157,
      "seek": 32746,
      "start": 345.64,
      "end": 346.65999999999997,
      "text": "中間發生的過程就是",
      "tokens": [
        51274,
        5975,
        11016,
        14637,
        8244,
        1546,
        8816,
        29649,
        5620,
        51325
      ],
      "temperature": 0,
      "avg_logprob": -0.3403220135232677,
      "compression_ratio": 1.2959183673469388,
      "no_speech_prob": 2.5257860039595847e-11,
      "source": "whisper"
    },
    {
      "id": 158,
      "seek": 32746,
      "start": 346.91999999999996,
      "end": 347.67999999999995,
      "text": "STT code端",
      "tokens": [
        51338,
        6840,
        51,
        3089,
        11957,
        107,
        51376
      ],
      "temperature": 0,
      "avg_logprob": -0.3403220135232677,
      "compression_ratio": 1.2959183673469388,
      "no_speech_prob": 2.5257860039595847e-11,
      "source": "whisper"
    },
    {
      "id": 159,
      "seek": 32746,
      "start": 347.94,
      "end": 348.2,
      "text": "你要走",
      "tokens": [
        51389,
        32085,
        9575,
        51402
      ],
      "temperature": 0,
      "avg_logprob": -0.3403220135232677,
      "compression_ratio": 1.2959183673469388,
      "no_speech_prob": 2.5257860039595847e-11,
      "source": "whisper"
    },
    {
      "id": 160,
      "seek": 32746,
      "start": 348.46,
      "end": 349.21999999999997,
      "text": "STT的服務",
      "tokens": [
        51415,
        6840,
        51,
        1546,
        27408,
        39517,
        51453
      ],
      "temperature": 0,
      "avg_logprob": -0.3403220135232677,
      "compression_ratio": 1.2959183673469388,
      "no_speech_prob": 2.5257860039595847e-11,
      "source": "whisper"
    },
    {
      "id": 161,
      "seek": 34922,
      "start": 349.48,
      "end": 349.74,
      "text": "對吧",
      "tokens": [
        50378,
        47905,
        50391
      ],
      "temperature": 0,
      "avg_logprob": -0.3169225133107808,
      "compression_ratio": 1.3285198555956679,
      "no_speech_prob": 2.4271130244502004e-11,
      "source": "whisper"
    },
    {
      "id": 162,
      "seek": 34922,
      "start": 349.98,
      "end": 350.5,
      "text": "你比如說",
      "tokens": [
        50403,
        2166,
        36757,
        4622,
        50429
      ],
      "temperature": 0,
      "avg_logprob": -0.3169225133107808,
      "compression_ratio": 1.3285198555956679,
      "no_speech_prob": 2.4271130244502004e-11,
      "source": "whisper"
    },
    {
      "id": 163,
      "seek": 34922,
      "start": 351.02000000000004,
      "end": 351.78000000000003,
      "text": "阿里的呀",
      "tokens": [
        50455,
        13594,
        15759,
        1546,
        15348,
        50493
      ],
      "temperature": 0,
      "avg_logprob": -0.3169225133107808,
      "compression_ratio": 1.3285198555956679,
      "no_speech_prob": 2.4271130244502004e-11,
      "source": "whisper"
    },
    {
      "id": 164,
      "seek": 34922,
      "start": 352.04,
      "end": 352.54,
      "text": "Deepgram呀",
      "tokens": [
        50506,
        11089,
        595,
        1342,
        15348,
        50531
      ],
      "temperature": 0,
      "avg_logprob": -0.3169225133107808,
      "compression_ratio": 1.3285198555956679,
      "no_speech_prob": 2.4271130244502004e-11,
      "source": "whisper"
    },
    {
      "id": 165,
      "seek": 34922,
      "start": 352.8,
      "end": 353.06,
      "text": "之類的",
      "tokens": [
        50544,
        9574,
        40092,
        1546,
        50557
      ],
      "temperature": 0,
      "avg_logprob": -0.3169225133107808,
      "compression_ratio": 1.3285198555956679,
      "no_speech_prob": 2.4271130244502004e-11,
      "source": "whisper"
    },
    {
      "id": 166,
      "seek": 34922,
      "start": 353.32000000000005,
      "end": 353.82000000000005,
      "text": "那麼",
      "tokens": [
        50570,
        20504,
        50595
      ],
      "temperature": 0,
      "avg_logprob": -0.3169225133107808,
      "compression_ratio": 1.3285198555956679,
      "no_speech_prob": 2.4271130244502004e-11,
      "source": "whisper"
    },
    {
      "id": 167,
      "seek": 34922,
      "start": 354.6,
      "end": 355.62,
      "text": "這個如果是LLM的話",
      "tokens": [
        50634,
        6287,
        13119,
        1541,
        40,
        81,
        76,
        21358,
        50685
      ],
      "temperature": 0,
      "avg_logprob": -0.3169225133107808,
      "compression_ratio": 1.3285198555956679,
      "no_speech_prob": 2.4271130244502004e-11,
      "source": "whisper",
      "text_original": "這個如果是Irm的話",
      "correction_applied": true,
      "correction_source": "llm_correct_a_presegmentation"
    },
    {
      "id": 168,
      "seek": 34922,
      "start": 355.88000000000005,
      "end": 356.64000000000004,
      "text": "你可能接入OpenAI",
      "tokens": [
        50698,
        2166,
        16657,
        14468,
        14028,
        45569,
        48698,
        50736
      ],
      "temperature": 0,
      "avg_logprob": -0.3169225133107808,
      "compression_ratio": 1.3285198555956679,
      "no_speech_prob": 2.4271130244502004e-11,
      "source": "whisper"
    },
    {
      "id": 170,
      "seek": 34922,
      "start": 357.16,
      "end": 357.92,
      "text": "也可能接入Deepseek",
      "tokens": [
        50762,
        6404,
        16657,
        14468,
        14028,
        11089,
        595,
        405,
        916,
        50800
      ],
      "temperature": 0,
      "avg_logprob": -0.3169225133107808,
      "compression_ratio": 1.3285198555956679,
      "no_speech_prob": 2.4271130244502004e-11,
      "source": "whisper"
    },
    {
      "id": 171,
      "seek": 34922,
      "start": 358.18,
      "end": 358.94000000000005,
      "text": "有可能接入Milimax",
      "tokens": [
        50813,
        2412,
        16657,
        14468,
        14028,
        44,
        388,
        332,
        2797,
        50851
      ],
      "temperature": 0,
      "avg_logprob": -0.3169225133107808,
      "compression_ratio": 1.3285198555956679,
      "no_speech_prob": 2.4271130244502004e-11,
      "source": "whisper"
    },
    {
      "id": 172,
      "seek": 34922,
      "start": 359.20000000000005,
      "end": 359.46000000000004,
      "text": "Kimi",
      "tokens": [
        50864,
        42,
        10121,
        50877
      ],
      "temperature": 0,
      "avg_logprob": -0.3169225133107808,
      "compression_ratio": 1.3285198555956679,
      "no_speech_prob": 2.4271130244502004e-11,
      "source": "whisper"
    },
    {
      "id": 173,
      "seek": 34922,
      "start": 359.72,
      "end": 360.22,
      "text": "都可以",
      "tokens": [
        50890,
        7182,
        6723,
        50915
      ],
      "temperature": 0,
      "avg_logprob": -0.3169225133107808,
      "compression_ratio": 1.3285198555956679,
      "no_speech_prob": 2.4271130244502004e-11,
      "source": "whisper"
    },
    {
      "id": 174,
      "seek": 34922,
      "start": 360.48,
      "end": 360.74,
      "text": "好",
      "tokens": [
        50928,
        2131,
        50941
      ],
      "temperature": 0,
      "avg_logprob": -0.3169225133107808,
      "compression_ratio": 1.3285198555956679,
      "no_speech_prob": 2.4271130244502004e-11,
      "source": "whisper"
    },
    {
      "id": 175,
      "seek": 34922,
      "start": 361.0,
      "end": 361.26000000000005,
      "text": "那麼",
      "tokens": [
        50954,
        20504,
        50967
      ],
      "temperature": 0,
      "avg_logprob": -0.3169225133107808,
      "compression_ratio": 1.3285198555956679,
      "no_speech_prob": 2.4271130244502004e-11,
      "source": "whisper"
    },
    {
      "id": 176,
      "seek": 34922,
      "start": 361.5,
      "end": 362.28000000000003,
      "text": "那麼TTS這塊",
      "tokens": [
        50979,
        51,
        11844,
        2664,
        26268,
        51018
      ],
      "temperature": 0,
      "avg_logprob": -0.3169225133107808,
      "compression_ratio": 1.3285198555956679,
      "no_speech_prob": 2.4271130244502004e-11,
      "source": "whisper",
      "text_original": "TDS這塊",
      "correction_applied": true,
      "correction_source": "llm_correct_a_presegmentation"
    },
    {
      "id": 177,
      "seek": 34922,
      "start": 362.54,
      "end": 363.56,
      "text": "通過SSE的流",
      "tokens": [
        51031,
        19550,
        8816,
        50,
        5879,
        1546,
        27854,
        51082
      ],
      "temperature": 0,
      "avg_logprob": -0.3169225133107808,
      "compression_ratio": 1.3285198555956679,
      "no_speech_prob": 2.4271130244502004e-11,
      "source": "whisper"
    },
    {
      "id": 178,
      "seek": 34922,
      "start": 363.82000000000005,
      "end": 364.32000000000005,
      "text": "來進行播放",
      "tokens": [
        51095,
        3763,
        18214,
        8082,
        49993,
        12744,
        51120
      ],
      "temperature": 0,
      "avg_logprob": -0.3169225133107808,
      "compression_ratio": 1.3285198555956679,
      "no_speech_prob": 2.4271130244502004e-11,
      "source": "whisper"
    },
    {
      "id": 179,
      "seek": 34922,
      "start": 364.58000000000004,
      "end": 365.6,
      "text": "那麼SSE什麼呢",
      "tokens": [
        51133,
        20504,
        50,
        5879,
        7598,
        6240,
        51184
      ],
      "temperature": 0,
      "avg_logprob": -0.3169225133107808,
      "compression_ratio": 1.3285198555956679,
      "no_speech_prob": 2.4271130244502004e-11,
      "source": "whisper"
    },
    {
      "id": 180,
      "seek": 34922,
      "start": 365.86,
      "end": 367.90000000000003,
      "text": "這個是AI裡面的一個技術",
      "tokens": [
        51197,
        41427,
        48698,
        32399,
        1546,
        8990,
        32502,
        9890,
        241,
        51299
      ],
      "temperature": 0,
      "avg_logprob": -0.3169225133107808,
      "compression_ratio": 1.3285198555956679,
      "no_speech_prob": 2.4271130244502004e-11,
      "source": "whisper"
    },
    {
      "id": 181,
      "seek": 34922,
      "start": 368.16,
      "end": 369.70000000000005,
      "text": "就是一個流式的",
      "tokens": [
        51312,
        5620,
        8990,
        27854,
        27584,
        1546,
        51389
      ],
      "temperature": 0,
      "avg_logprob": -0.3169225133107808,
      "compression_ratio": 1.3285198555956679,
      "no_speech_prob": 2.4271130244502004e-11,
      "source": "whisper"
    },
    {
      "id": 182,
      "seek": 34922,
      "start": 370.22,
      "end": 370.98,
      "text": "返回的",
      "tokens": [
        51415,
        3316,
        242,
        8350,
        1546,
        51453
      ],
      "temperature": 0,
      "avg_logprob": -0.3169225133107808,
      "compression_ratio": 1.3285198555956679,
      "no_speech_prob": 2.4271130244502004e-11,
      "source": "whisper"
    },
    {
      "id": 183,
      "seek": 34922,
      "start": 371.24,
      "end": 371.5,
      "text": "一個",
      "tokens": [
        51466,
        8990,
        51479
      ],
      "temperature": 0,
      "avg_logprob": -0.3169225133107808,
      "compression_ratio": 1.3285198555956679,
      "no_speech_prob": 2.4271130244502004e-11,
      "source": "whisper"
    },
    {
      "id": 184,
      "seek": 34922,
      "start": 371.74,
      "end": 372.52000000000004,
      "text": "處理技術",
      "tokens": [
        51491,
        39289,
        13876,
        32502,
        9890,
        241,
        51530
      ],
      "temperature": 0,
      "avg_logprob": -0.3169225133107808,
      "compression_ratio": 1.3285198555956679,
      "no_speech_prob": 2.4271130244502004e-11,
      "source": "whisper"
    },
    {
      "id": 185,
      "seek": 34922,
      "start": 373.02000000000004,
      "end": 374.3,
      "text": "就是文字的一個流式",
      "tokens": [
        51555,
        5620,
        17174,
        22381,
        1546,
        8990,
        27854,
        27584,
        51619
      ],
      "temperature": 0,
      "avg_logprob": -0.3169225133107808,
      "compression_ratio": 1.3285198555956679,
      "no_speech_prob": 2.4271130244502004e-11,
      "source": "whisper"
    },
    {
      "id": 186,
      "seek": 34922,
      "start": 374.56,
      "end": 375.84000000000003,
      "text": "同時還有音頻的",
      "tokens": [
        51632,
        13089,
        6611,
        15569,
        18034,
        29940,
        1546,
        51696
      ],
      "temperature": 0,
      "avg_logprob": -0.3169225133107808,
      "compression_ratio": 1.3285198555956679,
      "no_speech_prob": 2.4271130244502004e-11,
      "source": "whisper"
    },
    {
      "id": 187,
      "seek": 34922,
      "start": 376.62,
      "end": 377.38000000000005,
      "text": "最後到",
      "tokens": [
        51735,
        20578,
        4511,
        51773
      ],
      "temperature": 0,
      "avg_logprob": -0.3169225133107808,
      "compression_ratio": 1.3285198555956679,
      "no_speech_prob": 2.4271130244502004e-11,
      "source": "whisper"
    },
    {
      "id": 188,
      "seek": 37738,
      "start": 377.38,
      "end": 380.7,
      "text": "網頁這段它有音頻的一個流式",
      "tokens": [
        50365,
        26352,
        8313,
        223,
        2664,
        28427,
        11284,
        2412,
        18034,
        29940,
        1546,
        8990,
        27854,
        27584,
        50531
      ],
      "temperature": 0,
      "avg_logprob": -0.2892406034153818,
      "compression_ratio": 1.371841155234657,
      "no_speech_prob": 2.0472621861666873e-11,
      "source": "whisper"
    },
    {
      "id": 189,
      "seek": 37738,
      "start": 381.48,
      "end": 383.78,
      "text": "所以這個是更詳細的一個完整流程",
      "tokens": [
        50570,
        7239,
        41427,
        19002,
        6716,
        111,
        41588,
        1546,
        8990,
        14128,
        27662,
        27854,
        29649,
        50685
      ],
      "temperature": 0,
      "avg_logprob": -0.2892406034153818,
      "compression_ratio": 1.371841155234657,
      "no_speech_prob": 2.0472621861666873e-11,
      "source": "whisper"
    },
    {
      "id": 190,
      "seek": 37738,
      "start": 384.04,
      "end": 385.82,
      "text": "這裡面涉及到好些技術",
      "tokens": [
        50698,
        18771,
        8833,
        35681,
        231,
        25703,
        4511,
        2131,
        13824,
        32502,
        9890,
        241,
        50787
      ],
      "temperature": 0,
      "avg_logprob": -0.2892406034153818,
      "compression_ratio": 1.371841155234657,
      "no_speech_prob": 2.0472621861666873e-11,
      "source": "whisper"
    },
    {
      "id": 191,
      "seek": 37738,
      "start": 386.6,
      "end": 386.86,
      "text": "好技術",
      "tokens": [
        50826,
        2131,
        32502,
        9890,
        241,
        50839
      ],
      "temperature": 0,
      "avg_logprob": -0.2892406034153818,
      "compression_ratio": 1.371841155234657,
      "no_speech_prob": 2.0472621861666873e-11,
      "source": "whisper"
    },
    {
      "id": 192,
      "seek": 37738,
      "start": 387.1,
      "end": 387.88,
      "text": "大家這些技術",
      "tokens": [
        50851,
        6868,
        29869,
        32502,
        9890,
        241,
        50890
      ],
      "temperature": 0,
      "avg_logprob": -0.2892406034153818,
      "compression_ratio": 1.371841155234657,
      "no_speech_prob": 2.0472621861666873e-11,
      "source": "whisper"
    },
    {
      "id": 193,
      "seek": 37738,
      "start": 388.14,
      "end": 388.9,
      "text": "我們會一步步拆解",
      "tokens": [
        50903,
        5884,
        6236,
        2257,
        31429,
        31429,
        6852,
        228,
        17278,
        50941
      ],
      "temperature": 0,
      "avg_logprob": -0.2892406034153818,
      "compression_ratio": 1.371841155234657,
      "no_speech_prob": 2.0472621861666873e-11,
      "source": "whisper"
    },
    {
      "id": 194,
      "seek": 37738,
      "start": 389.15999999999997,
      "end": 389.65999999999997,
      "text": "給大家講的",
      "tokens": [
        50954,
        17798,
        6868,
        11932,
        1546,
        50979
      ],
      "temperature": 0,
      "avg_logprob": -0.2892406034153818,
      "compression_ratio": 1.371841155234657,
      "no_speech_prob": 2.0472621861666873e-11,
      "source": "whisper"
    },
    {
      "id": 195,
      "seek": 37738,
      "start": 390.18,
      "end": 390.44,
      "text": "那麼",
      "tokens": [
        51005,
        20504,
        51018
      ],
      "temperature": 0,
      "avg_logprob": -0.2892406034153818,
      "compression_ratio": 1.371841155234657,
      "no_speech_prob": 2.0472621861666873e-11,
      "source": "whisper"
    },
    {
      "id": 196,
      "seek": 37738,
      "start": 390.7,
      "end": 392.21999999999997,
      "text": "我們對於一個",
      "tokens": [
        51031,
        5884,
        2855,
        19488,
        8990,
        51107
      ],
      "temperature": 0,
      "avg_logprob": -0.2892406034153818,
      "compression_ratio": 1.371841155234657,
      "no_speech_prob": 2.0472621861666873e-11,
      "source": "whisper"
    },
    {
      "id": 197,
      "seek": 37738,
      "start": 392.48,
      "end": 394.28,
      "text": "總體的架構以及完整流程",
      "tokens": [
        51120,
        26575,
        23987,
        1546,
        7360,
        114,
        43362,
        40282,
        14128,
        27662,
        27854,
        29649,
        51210
      ],
      "temperature": 0,
      "avg_logprob": -0.2892406034153818,
      "compression_ratio": 1.371841155234657,
      "no_speech_prob": 2.0472621861666873e-11,
      "source": "whisper"
    },
    {
      "id": 198,
      "seek": 37738,
      "start": 394.54,
      "end": 395.3,
      "text": "有所了解以後",
      "tokens": [
        51223,
        2412,
        5966,
        2289,
        17278,
        3588,
        5661,
        51261
      ],
      "temperature": 0,
      "avg_logprob": -0.2892406034153818,
      "compression_ratio": 1.371841155234657,
      "no_speech_prob": 2.0472621861666873e-11,
      "source": "whisper"
    },
    {
      "id": 199,
      "seek": 37738,
      "start": 395.82,
      "end": 396.58,
      "text": "你才能搞懂",
      "tokens": [
        51287,
        2166,
        18888,
        8225,
        41816,
        29624,
        51325
      ],
      "temperature": 0,
      "avg_logprob": -0.2892406034153818,
      "compression_ratio": 1.371841155234657,
      "no_speech_prob": 2.0472621861666873e-11,
      "source": "whisper"
    },
    {
      "id": 200,
      "seek": 37738,
      "start": 396.84,
      "end": 398.62,
      "text": "我們後續要給大家講的",
      "tokens": [
        51338,
        5884,
        5661,
        29583,
        4275,
        17798,
        6868,
        11932,
        1546,
        51427
      ],
      "temperature": 0,
      "avg_logprob": -0.2892406034153818,
      "compression_ratio": 1.371841155234657,
      "no_speech_prob": 2.0472621861666873e-11,
      "source": "whisper"
    },
    {
      "id": 201,
      "seek": 37738,
      "start": 399.4,
      "end": 399.9,
      "text": "比如說",
      "tokens": [
        51466,
        36757,
        4622,
        51491
      ],
      "temperature": 0,
      "avg_logprob": -0.2892406034153818,
      "compression_ratio": 1.371841155234657,
      "no_speech_prob": 2.0472621861666873e-11,
      "source": "whisper"
    },
    {
      "id": 202,
      "seek": 37738,
      "start": 401.96,
      "end": 404.0,
      "text": "從Web端到Server端的整個的一個",
      "tokens": [
        51594,
        21068,
        4360,
        65,
        11957,
        107,
        4511,
        31859,
        331,
        11957,
        107,
        1546,
        27662,
        3338,
        1546,
        8990,
        51696
      ],
      "temperature": 0,
      "avg_logprob": -0.2892406034153818,
      "compression_ratio": 1.371841155234657,
      "no_speech_prob": 2.0472621861666873e-11,
      "source": "whisper"
    },
    {
      "id": 203,
      "seek": 37738,
      "start": 404.78,
      "end": 406.82,
      "text": "Voice Agent的一個整個流程",
      "tokens": [
        51735,
        13339,
        27174,
        1546,
        8990,
        27662,
        3338,
        27854,
        29649,
        51837
      ],
      "temperature": 0,
      "avg_logprob": -0.2892406034153818,
      "compression_ratio": 1.371841155234657,
      "no_speech_prob": 2.0472621861666873e-11,
      "source": "whisper"
    },
    {
      "id": 204,
      "seek": 40682,
      "start": 406.82,
      "end": 408.09999999999997,
      "text": "那麼我們這個例子",
      "tokens": [
        50365,
        20504,
        5884,
        6287,
        17797,
        7626,
        50429
      ],
      "temperature": 0,
      "avg_logprob": NaN,
      "compression_ratio": 1.2715654952076678,
      "no_speech_prob": 2.1191064528136572e-11,
      "source": "whisper"
    },
    {
      "id": 205,
      "seek": 40682,
      "start": 408.36,
      "end": 409.12,
      "text": "寫的不是很大",
      "tokens": [
        50442,
        4510,
        104,
        1546,
        7296,
        4563,
        3582,
        50480
      ],
      "temperature": 0,
      "avg_logprob": NaN,
      "compression_ratio": 1.2715654952076678,
      "no_speech_prob": 2.1191064528136572e-11,
      "source": "whisper"
    },
    {
      "id": 206,
      "seek": 40682,
      "start": 409.38,
      "end": 410.65999999999997,
      "text": "但是通過這個例子",
      "tokens": [
        50493,
        11189,
        19550,
        8816,
        6287,
        17797,
        7626,
        50557
      ],
      "temperature": 0,
      "avg_logprob": NaN,
      "compression_ratio": 1.2715654952076678,
      "no_speech_prob": 2.1191064528136572e-11,
      "source": "whisper"
    },
    {
      "id": 207,
      "seek": 40682,
      "start": 410.92,
      "end": 411.18,
      "text": "你",
      "tokens": [
        50570,
        2166,
        50583
      ],
      "temperature": 0,
      "avg_logprob": NaN,
      "compression_ratio": 1.2715654952076678,
      "no_speech_prob": 2.1191064528136572e-11,
      "source": "whisper"
    },
    {
      "id": 208,
      "seek": 40682,
      "start": 411.42,
      "end": 413.21999999999997,
      "text": "能對Voice Agent有一個",
      "tokens": [
        50595,
        8225,
        2855,
        13339,
        27174,
        2412,
        8990,
        50685
      ],
      "temperature": 0,
      "avg_logprob": NaN,
      "compression_ratio": 1.2715654952076678,
      "no_speech_prob": 2.1191064528136572e-11,
      "source": "whisper"
    },
    {
      "id": 209,
      "seek": 40682,
      "start": 413.74,
      "end": 415.26,
      "text": "完整的流程的一個理解",
      "tokens": [
        50711,
        14128,
        27662,
        1546,
        27854,
        29649,
        1546,
        8990,
        13876,
        17278,
        50787
      ],
      "temperature": 0,
      "avg_logprob": NaN,
      "compression_ratio": 1.2715654952076678,
      "no_speech_prob": 2.1191064528136572e-11,
      "source": "whisper"
    },
    {
      "id": 210,
      "seek": 40682,
      "start": 415.52,
      "end": 416.04,
      "text": "並且",
      "tokens": [
        50800,
        35934,
        20334,
        50826
      ],
      "temperature": 0,
      "avg_logprob": NaN,
      "compression_ratio": 1.2715654952076678,
      "no_speech_prob": 2.1191064528136572e-11,
      "source": "whisper"
    },
    {
      "id": 211,
      "seek": 40682,
      "start": 416.3,
      "end": 417.82,
      "text": "它是正式辦理能跑起來的一個項目",
      "tokens": [
        50839,
        11284,
        1541,
        15789,
        27584,
        23626,
        13876,
        8225,
        32585,
        21670,
        1546,
        8990,
        8313,
        227,
        11386,
        50915
      ],
      "temperature": 0,
      "avg_logprob": NaN,
      "compression_ratio": 1.2715654952076678,
      "no_speech_prob": 2.1191064528136572e-11,
      "source": "whisper"
    },
    {
      "id": 212,
      "seek": 40682,
      "start": 418.34,
      "end": 420.14,
      "text": "我們首先是把理論先搞清楚",
      "tokens": [
        50941,
        5884,
        36490,
        1541,
        16075,
        13876,
        29340,
        10108,
        41816,
        38433,
        51031
      ],
      "temperature": 0,
      "avg_logprob": NaN,
      "compression_ratio": 1.2715654952076678,
      "no_speech_prob": 2.1191064528136572e-11,
      "source": "whisper"
    },
    {
      "id": 213,
      "seek": 40682,
      "start": 420.9,
      "end": 421.65999999999997,
      "text": "這裡面",
      "tokens": [
        51069,
        18771,
        8833,
        51107
      ],
      "temperature": 0,
      "avg_logprob": NaN,
      "compression_ratio": 1.2715654952076678,
      "no_speech_prob": 2.1191064528136572e-11,
      "source": "whisper"
    },
    {
      "id": 214,
      "seek": 40682,
      "start": 421.92,
      "end": 422.7,
      "text": "我們可以看到",
      "tokens": [
        51120,
        5884,
        6723,
        18032,
        51159
      ],
      "temperature": 0,
      "avg_logprob": NaN,
      "compression_ratio": 1.2715654952076678,
      "no_speech_prob": 2.1191064528136572e-11,
      "source": "whisper"
    },
    {
      "id": 215,
      "seek": 40682,
      "start": 422.94,
      "end": 423.98,
      "text": "這裡面我們涉及的各種服務",
      "tokens": [
        51171,
        18771,
        8833,
        5884,
        35681,
        231,
        25703,
        1546,
        17516,
        16516,
        27408,
        39517,
        51223
      ],
      "temperature": 0,
      "avg_logprob": NaN,
      "compression_ratio": 1.2715654952076678,
      "no_speech_prob": 2.1191064528136572e-11,
      "source": "whisper"
    },
    {
      "id": 216,
      "seek": 40682,
      "start": 424.21999999999997,
      "end": 425.0,
      "text": "比如",
      "tokens": [
        51235,
        36757,
        51274
      ],
      "temperature": 0,
      "avg_logprob": NaN,
      "compression_ratio": 1.2715654952076678,
      "no_speech_prob": 2.1191064528136572e-11,
      "source": "whisper"
    },
    {
      "id": 217,
      "seek": 40682,
      "start": 425.26,
      "end": 425.76,
      "text": "STT",
      "tokens": [
        51287,
        6840,
        7327,
        51312
      ],
      "temperature": 0,
      "avg_logprob": NaN,
      "compression_ratio": 1.2715654952076678,
      "no_speech_prob": 2.1191064528136572e-11,
      "source": "whisper",
      "text_original": "STTS",
      "correction_applied": true,
      "correction_source": "llm_correct_a_presegmentation"
    },
    {
      "id": 218,
      "seek": 40682,
      "start": 426.02,
      "end": 426.53999999999996,
      "text": "IM",
      "tokens": [
        51325,
        6324,
        51351
      ],
      "temperature": 0,
      "avg_logprob": NaN,
      "compression_ratio": 1.2715654952076678,
      "no_speech_prob": 2.1191064528136572e-11,
      "source": "whisper"
    },
    {
      "id": 219,
      "seek": 40682,
      "start": 426.78,
      "end": 427.03999999999996,
      "text": "這些",
      "tokens": [
        51363,
        29869,
        51376
      ],
      "temperature": 0,
      "avg_logprob": NaN,
      "compression_ratio": 1.2715654952076678,
      "no_speech_prob": 2.1191064528136572e-11,
      "source": "whisper"
    },
    {
      "id": 220,
      "seek": 40682,
      "start": 427.3,
      "end": 429.1,
      "text": "大家都可以去採用不同的",
      "tokens": [
        51389,
        46525,
        6723,
        6734,
        6900,
        94,
        9254,
        47123,
        1546,
        51479
      ],
      "temperature": 0,
      "avg_logprob": NaN,
      "compression_ratio": 1.2715654952076678,
      "no_speech_prob": 2.1191064528136572e-11,
      "source": "whisper"
    },
    {
      "id": 221,
      "seek": 40682,
      "start": 429.6,
      "end": 430.38,
      "text": "Provider",
      "tokens": [
        51504,
        12681,
        85,
        1438,
        51543
      ],
      "temperature": 0,
      "avg_logprob": NaN,
      "compression_ratio": 1.2715654952076678,
      "no_speech_prob": 2.1191064528136572e-11,
      "source": "whisper"
    },
    {
      "id": 222,
      "seek": 40682,
      "start": 430.62,
      "end": 431.4,
      "text": "有著供應商的服務",
      "tokens": [
        51555,
        2412,
        19382,
        3254,
        249,
        20481,
        45581,
        1546,
        27408,
        39517,
        51594
      ],
      "temperature": 0,
      "avg_logprob": NaN,
      "compression_ratio": 1.2715654952076678,
      "no_speech_prob": 2.1191064528136572e-11,
      "source": "whisper"
    },
    {
      "id": 223,
      "seek": 40682,
      "start": 431.65999999999997,
      "end": 432.68,
      "text": "不一定說是非要某一間的",
      "tokens": [
        51607,
        1960,
        24272,
        4622,
        1541,
        12107,
        4275,
        17238,
        238,
        2257,
        11016,
        1546,
        51658
      ],
      "temperature": 0,
      "avg_logprob": NaN,
      "compression_ratio": 1.2715654952076678,
      "no_speech_prob": 2.1191064528136572e-11,
      "source": "whisper"
    },
    {
      "id": 224,
      "seek": 40682,
      "start": 432.94,
      "end": 433.18,
      "text": "對吧",
      "tokens": [
        51671,
        47905,
        51683
      ],
      "temperature": 0,
      "avg_logprob": NaN,
      "compression_ratio": 1.2715654952076678,
      "no_speech_prob": 2.1191064528136572e-11,
      "source": "whisper"
    },
    {
      "id": 225,
      "seek": 40682,
      "start": 433.96,
      "end": 434.46,
      "text": "!",
      "tokens": [
        51722,
        0,
        51747
      ],
      "temperature": 0,
      "avg_logprob": NaN,
      "compression_ratio": 1.2715654952076678,
      "no_speech_prob": 2.1191064528136572e-11,
      "source": "whisper"
    },
    {
      "id": 226,
      "seek": 43446,
      "start": 434.46,
      "end": 437.02,
      "text": "比如說如果說我們英文的我建議大家用Deepgram",
      "tokens": [
        50365,
        36757,
        4622,
        13119,
        4622,
        5884,
        27869,
        17174,
        1546,
        1654,
        34157,
        24686,
        6868,
        9254,
        11089,
        595,
        1342,
        50493
      ],
      "temperature": 0,
      "avg_logprob": -0.2849373235935118,
      "compression_ratio": 1.2702702702702702,
      "no_speech_prob": 2.5017274710159576e-11,
      "source": "whisper"
    },
    {
      "id": 227,
      "seek": 43446,
      "start": 437.28,
      "end": 438.03999999999996,
      "text": "Eleven Lab",
      "tokens": [
        50506,
        36,
        306,
        553,
        10137,
        50544
      ],
      "temperature": 0,
      "avg_logprob": -0.2849373235935118,
      "compression_ratio": 1.2702702702702702,
      "no_speech_prob": 2.5017274710159576e-11,
      "source": "whisper"
    },
    {
      "id": 228,
      "seek": 43446,
      "start": 438.29999999999995,
      "end": 440.59999999999997,
      "text": "如果說是中文的建議大家用",
      "tokens": [
        50557,
        13119,
        4622,
        1541,
        5975,
        17174,
        1546,
        34157,
        24686,
        6868,
        9254,
        50672
      ],
      "temperature": 0,
      "avg_logprob": -0.2849373235935118,
      "compression_ratio": 1.2702702702702702,
      "no_speech_prob": 2.5017274710159576e-11,
      "source": "whisper"
    },
    {
      "id": 229,
      "seek": 43446,
      "start": 441.12,
      "end": 443.94,
      "text": "阿里的、騰訊的、火山引擎的，像這些",
      "tokens": [
        50698,
        13594,
        15759,
        1546,
        38168,
        108,
        5396,
        232,
        1546,
        26171,
        25941,
        41301,
        13167,
        236,
        1546,
        12760,
        29869,
        50839
      ],
      "temperature": 0,
      "avg_logprob": -0.2849373235935118,
      "compression_ratio": 1.2702702702702702,
      "no_speech_prob": 2.5017274710159576e-11,
      "source": "whisper",
      "text_original": "阿里的騰訊的火車引擎的像這些",
      "correction_applied": true,
      "correction_source": "llm_correct_a_presegmentation"
    },
    {
      "id": 230,
      "seek": 43446,
      "start": 444.18,
      "end": 445.21999999999997,
      "text": "包括MiniMax，像這些",
      "tokens": [
        50851,
        41828,
        44,
        373,
        332,
        2797,
        12760,
        29869,
        50903
      ],
      "temperature": 0,
      "avg_logprob": -0.2849373235935118,
      "compression_ratio": 1.2702702702702702,
      "no_speech_prob": 2.5017274710159576e-11,
      "source": "whisper",
      "text_original": "包括Millimax像這些",
      "correction_applied": true,
      "correction_source": "llm_correct_a_presegmentation"
    },
    {
      "id": 231,
      "seek": 43446,
      "start": 445.71999999999997,
      "end": 447.0,
      "text": "這樣的話是比較好的",
      "tokens": [
        50928,
        8377,
        21358,
        1541,
        25174,
        20715,
        50992
      ],
      "temperature": 0,
      "avg_logprob": -0.2849373235935118,
      "compression_ratio": 1.2702702702702702,
      "no_speech_prob": 2.5017274710159576e-11,
      "source": "whisper"
    },
    {
      "id": 232,
      "seek": 43446,
      "start": 447.26,
      "end": 447.78,
      "text": "因為",
      "tokens": [
        51005,
        11471,
        51031
      ],
      "temperature": 0,
      "avg_logprob": -0.2849373235935118,
      "compression_ratio": 1.2702702702702702,
      "no_speech_prob": 2.5017274710159576e-11,
      "source": "whisper"
    },
    {
      "id": 233,
      "seek": 43446,
      "start": 448.02,
      "end": 449.06,
      "text": "肯定是",
      "tokens": [
        51043,
        14356,
        107,
        12088,
        1541,
        51095
      ],
      "temperature": 0,
      "avg_logprob": -0.2849373235935118,
      "compression_ratio": 1.2702702702702702,
      "no_speech_prob": 2.5017274710159576e-11,
      "source": "whisper"
    },
    {
      "id": 234,
      "seek": 43446,
      "start": 450.08,
      "end": 450.84,
      "text": "各自的模型",
      "tokens": [
        51146,
        17516,
        9722,
        1546,
        41908,
        39823,
        51184
      ],
      "temperature": 0,
      "avg_logprob": -0.2849373235935118,
      "compression_ratio": 1.2702702702702702,
      "no_speech_prob": 2.5017274710159576e-11,
      "source": "whisper"
    },
    {
      "id": 235,
      "seek": 43446,
      "start": 451.09999999999997,
      "end": 454.17999999999995,
      "text": "它對中英文支持的出來的效果是不一樣的",
      "tokens": [
        51197,
        11284,
        2855,
        5975,
        27869,
        17174,
        35488,
        1546,
        29741,
        1546,
        43076,
        9319,
        1541,
        1960,
        22977,
        1546,
        51351
      ],
      "temperature": 0,
      "avg_logprob": -0.2849373235935118,
      "compression_ratio": 1.2702702702702702,
      "no_speech_prob": 2.5017274710159576e-11,
      "source": "whisper"
    },
    {
      "id": 236,
      "seek": 43446,
      "start": 454.94,
      "end": 457.24,
      "text": "當然同時它還支持方言的",
      "tokens": [
        51389,
        21707,
        13089,
        6611,
        11284,
        7824,
        35488,
        9249,
        12009,
        1546,
        51504
      ],
      "temperature": 0,
      "avg_logprob": -0.2849373235935118,
      "compression_ratio": 1.2702702702702702,
      "no_speech_prob": 2.5017274710159576e-11,
      "source": "whisper"
    },
    {
      "id": 237,
      "seek": 43446,
      "start": 457.5,
      "end": 458.52,
      "text": "比如說粵語",
      "tokens": [
        51517,
        36757,
        4622,
        18393,
        113,
        31348,
        51568
      ],
      "temperature": 0,
      "avg_logprob": -0.2849373235935118,
      "compression_ratio": 1.2702702702702702,
      "no_speech_prob": 2.5017274710159576e-11,
      "source": "whisper"
    },
    {
      "id": 238,
      "seek": 45852,
      "start": 459.03999999999996,
      "end": 459.79999999999995,
      "text": "對吧",
      "tokens": [
        50391,
        2855,
        6062,
        50429
      ],
      "temperature": 0,
      "avg_logprob": -0.33353091246329697,
      "compression_ratio": 1.2481203007518797,
      "no_speech_prob": 2.3687104361291844e-11,
      "source": "whisper"
    },
    {
      "id": 239,
      "seek": 45852,
      "start": 460.32,
      "end": 460.56,
      "text": "等等",
      "tokens": [
        50455,
        36000,
        50467
      ],
      "temperature": 0,
      "avg_logprob": -0.33353091246329697,
      "compression_ratio": 1.2481203007518797,
      "no_speech_prob": 2.3687104361291844e-11,
      "source": "whisper"
    },
    {
      "id": 240,
      "seek": 45852,
      "start": 460.82,
      "end": 462.09999999999997,
      "text": "像這些方言的四川話",
      "tokens": [
        50480,
        12760,
        29869,
        9249,
        12009,
        1546,
        19425,
        27905,
        11103,
        50544
      ],
      "temperature": 0,
      "avg_logprob": -0.33353091246329697,
      "compression_ratio": 1.2481203007518797,
      "no_speech_prob": 2.3687104361291844e-11,
      "source": "whisper"
    },
    {
      "id": 241,
      "seek": 45852,
      "start": 462.62,
      "end": 463.12,
      "text": "普通話",
      "tokens": [
        50570,
        29993,
        19550,
        11103,
        50595
      ],
      "temperature": 0,
      "avg_logprob": -0.33353091246329697,
      "compression_ratio": 1.2481203007518797,
      "no_speech_prob": 2.3687104361291844e-11,
      "source": "whisper"
    },
    {
      "id": 242,
      "seek": 45852,
      "start": 463.9,
      "end": 465.44,
      "text": "像方言普通話都支持的",
      "tokens": [
        50634,
        12760,
        9249,
        12009,
        29993,
        19550,
        11103,
        7182,
        35488,
        1546,
        50711
      ],
      "temperature": 0,
      "avg_logprob": -0.33353091246329697,
      "compression_ratio": 1.2481203007518797,
      "no_speech_prob": 2.3687104361291844e-11,
      "source": "whisper"
    },
    {
      "id": 243,
      "seek": 45852,
      "start": 466.46,
      "end": 467.21999999999997,
      "text": "所以說這個",
      "tokens": [
        50762,
        7239,
        4622,
        6287,
        50800
      ],
      "temperature": 0,
      "avg_logprob": -0.33353091246329697,
      "compression_ratio": 1.2481203007518797,
      "no_speech_prob": 2.3687104361291844e-11,
      "source": "whisper"
    },
    {
      "id": 244,
      "seek": 45852,
      "start": 467.47999999999996,
      "end": 468.76,
      "text": "你要不同的去",
      "tokens": [
        50813,
        32085,
        47123,
        1546,
        6734,
        50877
      ],
      "temperature": 0,
      "avg_logprob": -0.33353091246329697,
      "compression_ratio": 1.2481203007518797,
      "no_speech_prob": 2.3687104361291844e-11,
      "source": "whisper"
    },
    {
      "id": 246,
      "seek": 45852,
      "start": 469.02,
      "end": 469.52,
      "text": "另外還有一個",
      "tokens": [
        50890,
        26202,
        15569,
        8990,
        50915
      ],
      "temperature": 0,
      "avg_logprob": -0.33353091246329697,
      "compression_ratio": 1.2481203007518797,
      "no_speech_prob": 2.3687104361291844e-11,
      "source": "whisper"
    },
    {
      "id": 247,
      "seek": 45852,
      "start": 469.78,
      "end": 470.56,
      "text": "VAD打斷",
      "tokens": [
        50928,
        53,
        4731,
        12467,
        4307,
        115,
        50967
      ],
      "temperature": 0,
      "avg_logprob": -0.33353091246329697,
      "compression_ratio": 1.2481203007518797,
      "no_speech_prob": 2.3687104361291844e-11,
      "source": "whisper",
      "text_original": "VED打斷",
      "correction_applied": true,
      "correction_source": "llm_correct_a_presegmentation"
    },
    {
      "id": 248,
      "seek": 45852,
      "start": 470.79999999999995,
      "end": 471.58,
      "text": "VED打斷的",
      "tokens": [
        50979,
        53,
        4731,
        12467,
        4307,
        115,
        1546,
        51018
      ],
      "temperature": 0,
      "avg_logprob": -0.33353091246329697,
      "compression_ratio": 1.2481203007518797,
      "no_speech_prob": 2.3687104361291844e-11,
      "source": "whisper"
    },
    {
      "id": 249,
      "seek": 45852,
      "start": 471.84,
      "end": 472.34,
      "text": "不同的",
      "tokens": [
        51031,
        47123,
        1546,
        51056
      ],
      "temperature": 0,
      "avg_logprob": -0.33353091246329697,
      "compression_ratio": 1.2481203007518797,
      "no_speech_prob": 2.3687104361291844e-11,
      "source": "whisper"
    },
    {
      "id": 250,
      "seek": 45852,
      "start": 473.12,
      "end": 473.88,
      "text": "比如STT",
      "tokens": [
        51095,
        36757,
        6840,
        51,
        51133
      ],
      "temperature": 0,
      "avg_logprob": -0.33353091246329697,
      "compression_ratio": 1.2481203007518797,
      "no_speech_prob": 2.3687104361291844e-11,
      "source": "whisper"
    },
    {
      "id": 251,
      "seek": 45852,
      "start": 474.14,
      "end": 474.4,
      "text": "就是",
      "tokens": [
        51146,
        5620,
        51159
      ],
      "temperature": 0,
      "avg_logprob": -0.33353091246329697,
      "compression_ratio": 1.2481203007518797,
      "no_speech_prob": 2.3687104361291844e-11,
      "source": "whisper"
    },
    {
      "id": 252,
      "seek": 45852,
      "start": 474.64,
      "end": 475.15999999999997,
      "text": "它叫",
      "tokens": [
        51171,
        11284,
        19855,
        51197
      ],
      "temperature": 0,
      "avg_logprob": -0.33353091246329697,
      "compression_ratio": 1.2481203007518797,
      "no_speech_prob": 2.3687104361291844e-11,
      "source": "whisper"
    },
    {
      "id": 253,
      "seek": 45852,
      "start": 475.41999999999996,
      "end": 475.68,
      "text": "實際上叫",
      "tokens": [
        51210,
        10376,
        34837,
        5708,
        19855,
        51223
      ],
      "temperature": 0,
      "avg_logprob": -0.33353091246329697,
      "compression_ratio": 1.2481203007518797,
      "no_speech_prob": 2.3687104361291844e-11,
      "source": "whisper"
    },
    {
      "id": 254,
      "seek": 45852,
      "start": 475.91999999999996,
      "end": 476.96,
      "text": "ASR技術",
      "tokens": [
        51235,
        3160,
        17,
        32502,
        9890,
        241,
        51287
      ],
      "temperature": 0,
      "avg_logprob": -0.33353091246329697,
      "compression_ratio": 1.2481203007518797,
      "no_speech_prob": 2.3687104361291844e-11,
      "source": "whisper",
      "text_original": "AS2技術",
      "correction_applied": true,
      "correction_source": "llm_correct_a_presegmentation"
    },
    {
      "id": 255,
      "seek": 45852,
      "start": 477.46,
      "end": 478.47999999999996,
      "text": "不同的",
      "tokens": [
        51312,
        47123,
        1546,
        51363
      ],
      "temperature": 0,
      "avg_logprob": -0.33353091246329697,
      "compression_ratio": 1.2481203007518797,
      "no_speech_prob": 2.3687104361291844e-11,
      "source": "whisper"
    },
    {
      "id": 256,
      "seek": 45852,
      "start": 479.52,
      "end": 480.28,
      "text": "廠商提供出來的",
      "tokens": [
        51415,
        17527,
        254,
        45581,
        20949,
        3254,
        249,
        29741,
        1546,
        51453
      ],
      "temperature": 0,
      "avg_logprob": -0.33353091246329697,
      "compression_ratio": 1.2481203007518797,
      "no_speech_prob": 2.3687104361291844e-11,
      "source": "whisper"
    },
    {
      "id": 257,
      "seek": 45852,
      "start": 480.53999999999996,
      "end": 481.56,
      "text": "質量也是不一樣的",
      "tokens": [
        51466,
        43450,
        26748,
        22021,
        1960,
        22977,
        1546,
        51517
      ],
      "temperature": 0,
      "avg_logprob": -0.33353091246329697,
      "compression_ratio": 1.2481203007518797,
      "no_speech_prob": 2.3687104361291844e-11,
      "source": "whisper"
    },
    {
      "id": 258,
      "seek": 45852,
      "start": 481.82,
      "end": 482.84,
      "text": "這個都去接",
      "tokens": [
        51530,
        6287,
        7182,
        6734,
        14468,
        51581
      ],
      "temperature": 0,
      "avg_logprob": -0.33353091246329697,
      "compression_ratio": 1.2481203007518797,
      "no_speech_prob": 2.3687104361291844e-11,
      "source": "whisper"
    },
    {
      "id": 259,
      "seek": 45852,
      "start": 483.35999999999996,
      "end": 484.38,
      "text": "然後我們這裡面當然就接了",
      "tokens": [
        51607,
        10213,
        5884,
        18771,
        8833,
        21707,
        3111,
        14468,
        2289,
        51658
      ],
      "temperature": 0,
      "avg_logprob": -0.33353091246329697,
      "compression_ratio": 1.2481203007518797,
      "no_speech_prob": 2.3687104361291844e-11,
      "source": "whisper"
    },
    {
      "id": 260,
      "seek": 45852,
      "start": 484.64,
      "end": 485.14,
      "text": "有阿里的",
      "tokens": [
        51671,
        2412,
        13594,
        15759,
        1546,
        51696
      ],
      "temperature": 0,
      "avg_logprob": -0.33353091246329697,
      "compression_ratio": 1.2481203007518797,
      "no_speech_prob": 2.3687104361291844e-11,
      "source": "whisper"
    },
    {
      "id": 261,
      "seek": 45852,
      "start": 485.4,
      "end": 485.91999999999996,
      "text": "有Deepgram",
      "tokens": [
        51709,
        2412,
        11089,
        595,
        1342,
        51735
      ],
      "temperature": 0,
      "avg_logprob": -0.33353091246329697,
      "compression_ratio": 1.2481203007518797,
      "no_speech_prob": 2.3687104361291844e-11,
      "source": "whisper"
    },
    {
      "id": 262,
      "seek": 48592,
      "start": 485.92,
      "end": 486.44,
      "text": "這些",
      "tokens": [
        50365,
        29869,
        50391
      ],
      "temperature": 0,
      "avg_logprob": -0.23609568709034032,
      "compression_ratio": 1.15527950310559,
      "no_speech_prob": 1.672467893343299e-11,
      "source": "whisper"
    },
    {
      "id": 263,
      "seek": 48592,
      "start": 486.94,
      "end": 489.24,
      "text": "已經有兩個比較好用的給大家接進來了",
      "tokens": [
        50416,
        25338,
        2412,
        34623,
        25174,
        2131,
        9254,
        1546,
        17798,
        6868,
        14468,
        18214,
        25364,
        50531
      ],
      "temperature": 0,
      "avg_logprob": -0.23609568709034032,
      "compression_ratio": 1.15527950310559,
      "no_speech_prob": 1.672467893343299e-11,
      "source": "whisper"
    },
    {
      "id": 264,
      "seek": 48592,
      "start": 489.5,
      "end": 490.52000000000004,
      "text": "告訴大家怎麼去接的",
      "tokens": [
        50544,
        31080,
        6868,
        11614,
        6734,
        14468,
        1546,
        50595
      ],
      "temperature": 0,
      "avg_logprob": -0.23609568709034032,
      "compression_ratio": 1.15527950310559,
      "no_speech_prob": 1.672467893343299e-11,
      "source": "whisper"
    },
    {
      "id": 265,
      "seek": 48592,
      "start": 491.8,
      "end": 492.06,
      "text": "好",
      "tokens": [
        50659,
        2131,
        50672
      ],
      "temperature": 0,
      "avg_logprob": -0.23609568709034032,
      "compression_ratio": 1.15527950310559,
      "no_speech_prob": 1.672467893343299e-11,
      "source": "whisper"
    },
    {
      "id": 266,
      "seek": 48592,
      "start": 492.32,
      "end": 496.42,
      "text": "那麼關於它的這個總體的架構儀完整流程的我們就先講到這裡謝謝大家觀看",
      "tokens": [
        50685,
        20504,
        14899,
        19488,
        45224,
        6287,
        26575,
        23987,
        1546,
        7360,
        114,
        43362,
        12827,
        222,
        14128,
        27662,
        27854,
        29649,
        1546,
        5884,
        3111,
        10108,
        11932,
        4511,
        18771,
        11003,
        6868,
        25179,
        4200,
        50890
      ],
      "temperature": 0,
      "avg_logprob": -0.23609568709034032,
      "compression_ratio": 1.15527950310559,
      "no_speech_prob": 1.672467893343299e-11,
      "source": "whisper"
    }
  ],
  "transcript_selection": {
    "schema": "videonote.transcript_selection.v1",
    "status": "SELECTED",
    "requested_language": "zh",
    "duration": 497.580438,
    "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": 265,
      "multilingual": false,
      "first_start": 0.76,
      "final_end": 496.42,
      "duration_coverage": 0.997668,
      "maximum_gap_seconds": 2.06,
      "inverted_count": 0,
      "language": {
        "requested": "zh",
        "matched": true,
        "latin": 394,
        "han": 2061,
        "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": 497.58,
      "rejected_segment_count": 0,
      "rejected_segments": []
    }
  },
  "presegmentation_correction": {
    "applied": true,
    "correction_count": 17,
    "deterministic_fix_count": 0
  }
}