{
  "text": "做字幕這件事很奇妙。觀眾覺得它理所當然,做個人覺得它像是有氣圖形。Substar做的就是為字幕製作裡最重複、最耗時間的部分設計了一套開箱僅用的可視化工具包。\n從音视频聽寫開始,再到AI切分、校准、翻譯、省略、打轴,最後直接导出成品字幕。\n初次使用時,請在首頁配置好聽寫服務需要的阿里雲百鏈API Key和編輯服務需要的製补API Key。\n它們都有新用户的適用額度,長期使用也极具性價比。\n配置完服務後,我們就可以開始進行字幕工作了。\n先把視頻或音频拖進來,也可以一次處理多個素材。\n設置原文語言、目標語言以及不同語言的單條字幕字符數上限。\n若要提升專門領域的聽寫準確率,在這裡可以設置熱詞和prompt,\n讓模型提前了解內容背景和需要重點識別的詞語,還可以AI一鍵填寫。\n工作台提供兩種高級切分模式和對參考文稿的支持。\n在上傳視頻時,附上參考文稿,編輯器會智能匹配正確的支持,\n並保留參考文稿與聽寫結果之間的差異,共鄰定多。\n一般而言,推薦使用AI輔助切分,它會結合上下文和語法結構,正能選擇合適的句子缺點。\n再按照字符上限,生成可編輯的字幕工程。\n當然,你可以選擇按照參考文稿標點切分。\n進入編輯器後,分距列表,媒體預覽和時間軸會保持聯動。\n右鍵單機自辭間隙可以進行切分,右鍵單機行尾輔助點可以合併下一條。\n左鍵雙機磁源進行編輯,脫選多個磁源可以進行合併、刪除等操作。\n拖動時間軸邊界即可調整字幕時間。\n對於相互吸附的字幕,左鍵單機控制鍵即可在聯動和個別調整之間自由切換。\nAI校準負責處理,標點、英語大小寫、專有名詞和可能的清解錯誤。\n也可以補充要求,比如重點檢查品牌名、機構名或數字。\n它不會重新生成整篇稿子,而是在原文上留下具體的改動、標記。\n當然,側邊還有經典的查找與替換功能,供你批量對命中的詞源進行編輯,流控就位刪除,隨時可以撤銷。\n外部AI審閱會生成一份可以粘貼到網頁端AI的審閱材料,可以檢查全部當前分距或指定範圍,並自動附帶前後上下文。\n遇到拿不准的地方,也可以附上自己的疑問,直接交給你最襯手的網頁端AI進一步判斷。\nAI翻譯支持逐航映射和多對多映射。\n在多對多映射模式中,模型會先結合上下文理解完整的意思,再把一文放回合適的分句,讓一文既照顧上下文,也符合字幕的閱讀節奏。\n大部分工作完成了,還有一些實用小工具來收尾。\n標點處理功能可以分別設置上型字幕和下型字幕中哪些符號應該被刪除,哪些應該被替換為空格。\n自動吸附用來批量調整字幕時間,向前智能吸附會尋找可信的起因位置,並保留少量靜音與量。\n向後則按預值處理相鄰字幕間的小空隙,減少字幕剛消失又立刻出現的閃爍感。\n編輯器還提供減反轉換和收畫人設置,適合不同地區版本以及採訪、播客這類多人素材。\n大公告稱,點擊右上角即可導出,需要交給剪輯軟件就直接導出字幕文件。\n如果還需要換一台電腦繼續使用Substar編輯,也可以把完整工程一起帶走。\n這就是Substar,把聽寫、切分、校準、翻譯、審閱和打軸整合在一個不斷完勝的工作流當中,讓這部工作不再是一件苦差事。",
  "language": "zh",
  "duration": 230.458063,
  "segments": [
    {
      "id": 0,
      "seek": 0,
      "start": 0.0,
      "end": 22.7,
      "text": "做字幕這件事很奇妙。觀眾覺得它理所當然,做個人覺得它像是有氣圖形。Substar做的就是為字幕製作裡最重複、最耗時間的部分設計了一套開箱僅用的可視化工具包。",
      "tokens": [
        50365,
        10907,
        22381,
        40211,
        2664,
        20485,
        6973,
        4563,
        26907,
        5648,
        247,
        1543,
        39368,
        11274,
        11284,
        13876,
        5966,
        21707,
        11,
        10907,
        40104,
        11274,
        11284,
        12760,
        1541,
        2412,
        24090,
        2523,
        244,
        30900,
        1543,
        39582,
        9710,
        10907,
        1546,
        5620,
        6344,
        22381,
        40211,
        8083,
        121,
        11914,
        11066,
        8661,
        12624,
        164,
        97,
        229,
        1231,
        8661,
        4450,
        245,
        20788,
        1546,
        32174,
        39035,
        30114,
        2289,
        2257,
        1881,
        245,
        8949,
        11249,
        109,
        9502,
        227,
        9254,
        1546,
        4429,
        27333,
        23756,
        23323,
        39806,
        23305,
        1543,
        51500
      ],
      "temperature": 0,
      "avg_logprob": -0.20803077404315656,
      "compression_ratio": 1.1065989847715736,
      "no_speech_prob": 8.079542243577542e-12,
      "source": "whisper"
    },
    {
      "id": 1,
      "seek": 2270,
      "start": 22.7,
      "end": 30.7,
      "text": "從音视频聽寫開始,再到AI切分、校准、翻譯、省略、打轴,最後直接导出成品字幕。",
      "tokens": [
        50365,
        21068,
        18034,
        40656,
        39752,
        21524,
        4510,
        104,
        21017,
        11,
        8623,
        4511,
        48698,
        23632,
        6627,
        1231,
        49684,
        6336,
        228,
        1231,
        42716,
        11823,
        107,
        1231,
        2862,
        223,
        6904,
        98,
        1231,
        12467,
        17819,
        112,
        11,
        20578,
        43297,
        4510,
        120,
        7781,
        11336,
        30246,
        22381,
        40211,
        1543,
        50765
      ],
      "temperature": 0,
      "avg_logprob": -0.1297893070039295,
      "compression_ratio": 1.299093655589124,
      "no_speech_prob": 9.783684279396354e-12,
      "source": "whisper"
    },
    {
      "id": 2,
      "seek": 2270,
      "start": 30.7,
      "end": 37.7,
      "text": "初次使用時,請在首頁配置好聽寫服務需要的阿里雲百鏈API Key和編輯服務需要的製补API Key。",
      "tokens": [
        50765,
        28727,
        9487,
        22982,
        9254,
        6611,
        11,
        16302,
        3581,
        25444,
        8313,
        223,
        38846,
        34719,
        2131,
        21524,
        4510,
        104,
        27408,
        39517,
        35748,
        1546,
        13594,
        15759,
        6306,
        110,
        31906,
        46241,
        230,
        4715,
        40,
        12759,
        12565,
        12608,
        101,
        12290,
        107,
        27408,
        39517,
        35748,
        1546,
        8083,
        121,
        9890,
        98,
        4715,
        40,
        12759,
        1543,
        51115
      ],
      "temperature": 0,
      "avg_logprob": -0.1297893070039295,
      "compression_ratio": 1.299093655589124,
      "no_speech_prob": 9.783684279396354e-12,
      "source": "whisper"
    },
    {
      "id": 3,
      "seek": 2270,
      "start": 37.7,
      "end": 41.7,
      "text": "它們都有新用户的適用額度,長期使用也极具性價比。",
      "tokens": [
        51115,
        11284,
        4623,
        48121,
        12560,
        9254,
        1486,
        115,
        1546,
        3330,
        102,
        9254,
        10178,
        235,
        13127,
        11,
        15353,
        16786,
        22982,
        9254,
        6404,
        7360,
        223,
        39806,
        21686,
        45441,
        11706,
        1543,
        51315
      ],
      "temperature": 0,
      "avg_logprob": -0.1297893070039295,
      "compression_ratio": 1.299093655589124,
      "no_speech_prob": 9.783684279396354e-12,
      "source": "whisper"
    },
    {
      "id": 4,
      "seek": 2270,
      "start": 41.7,
      "end": 44.7,
      "text": "配置完服務後,我們就可以開始進行字幕工作了。",
      "tokens": [
        51315,
        38846,
        34719,
        14128,
        27408,
        39517,
        5661,
        11,
        5884,
        36781,
        21017,
        18214,
        8082,
        22381,
        40211,
        41315,
        2289,
        1543,
        51465
      ],
      "temperature": 0,
      "avg_logprob": -0.1297893070039295,
      "compression_ratio": 1.299093655589124,
      "no_speech_prob": 9.783684279396354e-12,
      "source": "whisper"
    },
    {
      "id": 5,
      "seek": 2270,
      "start": 44.7,
      "end": 48.7,
      "text": "先把視頻或音频拖進來,也可以一次處理多個素材。",
      "tokens": [
        51465,
        10108,
        16075,
        27333,
        29940,
        19780,
        18034,
        39752,
        6852,
        244,
        18214,
        3763,
        11,
        47935,
        27505,
        39289,
        13876,
        6392,
        3338,
        37204,
        4422,
        238,
        1543,
        51665
      ],
      "temperature": 0,
      "avg_logprob": -0.1297893070039295,
      "compression_ratio": 1.299093655589124,
      "no_speech_prob": 9.783684279396354e-12,
      "source": "whisper"
    },
    {
      "id": 6,
      "seek": 4870,
      "start": 48.7,
      "end": 53.7,
      "text": "設置原文語言、目標語言以及不同語言的單條字幕字符數上限。",
      "tokens": [
        50365,
        39035,
        34719,
        19683,
        17174,
        31348,
        12009,
        1231,
        11386,
        46299,
        31348,
        12009,
        40282,
        47123,
        31348,
        12009,
        1546,
        24227,
        31106,
        22381,
        40211,
        22381,
        5437,
        99,
        30622,
        5708,
        43446,
        1543,
        50615
      ],
      "temperature": 0,
      "avg_logprob": -0.09637519510034571,
      "compression_ratio": 1.292817679558011,
      "no_speech_prob": 1.7377354763481456e-11,
      "source": "whisper"
    },
    {
      "id": 7,
      "seek": 4870,
      "start": 53.7,
      "end": 58.7,
      "text": "若要提升專門領域的聽寫準確率,在這裡可以設置熱詞和prompt,",
      "tokens": [
        50615,
        13736,
        98,
        4275,
        20949,
        41670,
        46072,
        37232,
        8313,
        246,
        16262,
        253,
        1546,
        21524,
        4510,
        104,
        28220,
        24293,
        44866,
        11,
        3581,
        18771,
        6723,
        39035,
        34719,
        32745,
        6716,
        252,
        12565,
        28722,
        662,
        11,
        50865
      ],
      "temperature": 0,
      "avg_logprob": -0.09637519510034571,
      "compression_ratio": 1.292817679558011,
      "no_speech_prob": 1.7377354763481456e-11,
      "source": "whisper"
    },
    {
      "id": 8,
      "seek": 4870,
      "start": 58.7,
      "end": 64.7,
      "text": "讓模型提前了解內容背景和需要重點識別的詞語,還可以AI一鍵填寫。",
      "tokens": [
        50865,
        21195,
        41908,
        39823,
        20949,
        8945,
        2289,
        17278,
        28472,
        25750,
        46329,
        50218,
        12565,
        35748,
        12624,
        8216,
        43143,
        16158,
        1546,
        6716,
        252,
        31348,
        11,
        7824,
        6723,
        48698,
        2257,
        41668,
        113,
        13331,
        104,
        4510,
        104,
        1543,
        51165
      ],
      "temperature": 0,
      "avg_logprob": -0.09637519510034571,
      "compression_ratio": 1.292817679558011,
      "no_speech_prob": 1.7377354763481456e-11,
      "source": "whisper"
    },
    {
      "id": 9,
      "seek": 4870,
      "start": 64.7,
      "end": 68.7,
      "text": "工作台提供兩種高級切分模式和對參考文稿的支持。",
      "tokens": [
        51165,
        41315,
        15433,
        20949,
        3254,
        249,
        16313,
        16516,
        12979,
        35315,
        23632,
        6627,
        41908,
        27584,
        12565,
        2855,
        47375,
        26504,
        17174,
        10415,
        123,
        1546,
        35488,
        1543,
        51365
      ],
      "temperature": 0,
      "avg_logprob": -0.09637519510034571,
      "compression_ratio": 1.292817679558011,
      "no_speech_prob": 1.7377354763481456e-11,
      "source": "whisper"
    },
    {
      "id": 10,
      "seek": 4870,
      "start": 68.7,
      "end": 73.7,
      "text": "在上傳視頻時,附上參考文稿,編輯器會智能匹配正確的支持,",
      "tokens": [
        51365,
        3581,
        5708,
        40852,
        27333,
        29940,
        6611,
        11,
        8842,
        226,
        5708,
        47375,
        26504,
        17174,
        10415,
        123,
        11,
        12608,
        101,
        12290,
        107,
        34386,
        6236,
        5094,
        118,
        8225,
        9937,
        117,
        38846,
        15789,
        24293,
        1546,
        35488,
        11,
        51615
      ],
      "temperature": 0,
      "avg_logprob": -0.09637519510034571,
      "compression_ratio": 1.292817679558011,
      "no_speech_prob": 1.7377354763481456e-11,
      "source": "whisper"
    },
    {
      "id": 11,
      "seek": 4870,
      "start": 73.7,
      "end": 77.7,
      "text": "並保留參考文稿與聽寫結果之間的差異,共鄰定多。",
      "tokens": [
        51615,
        35934,
        24302,
        24456,
        47375,
        26504,
        17174,
        10415,
        123,
        34127,
        21524,
        4510,
        104,
        35181,
        9574,
        11016,
        1546,
        21679,
        6904,
        108,
        11,
        15408,
        165,
        26267,
        12088,
        6392,
        1543,
        51815
      ],
      "temperature": 0,
      "avg_logprob": -0.09637519510034571,
      "compression_ratio": 1.292817679558011,
      "no_speech_prob": 1.7377354763481456e-11,
      "source": "whisper"
    },
    {
      "id": 12,
      "seek": 7770,
      "start": 77.7,
      "end": 84.7,
      "text": "一般而言,推薦使用AI輔助切分,它會結合上下文和語法結構,正能選擇合適的句子缺點。",
      "tokens": [
        50365,
        2257,
        49640,
        11070,
        12009,
        11,
        33597,
        27746,
        99,
        22982,
        9254,
        48698,
        12290,
        242,
        37618,
        23632,
        6627,
        11,
        11284,
        6236,
        17144,
        14245,
        5708,
        4438,
        17174,
        12565,
        31348,
        11148,
        17144,
        43362,
        11,
        15789,
        8225,
        21625,
        13167,
        229,
        14245,
        3330,
        102,
        1546,
        34592,
        7626,
        38109,
        118,
        8216,
        1543,
        50715
      ],
      "temperature": 0,
      "avg_logprob": -0.09382529485793341,
      "compression_ratio": 1.1975806451612903,
      "no_speech_prob": 2.336704441052717e-11,
      "source": "whisper"
    },
    {
      "id": 13,
      "seek": 7770,
      "start": 84.7,
      "end": 88.7,
      "text": "再按照字符上限,生成可編輯的字幕工程。",
      "tokens": [
        50715,
        8623,
        26613,
        32150,
        22381,
        5437,
        99,
        5708,
        43446,
        11,
        8244,
        11336,
        4429,
        12608,
        101,
        12290,
        107,
        1546,
        22381,
        40211,
        23323,
        29649,
        1543,
        50915
      ],
      "temperature": 0,
      "avg_logprob": -0.09382529485793341,
      "compression_ratio": 1.1975806451612903,
      "no_speech_prob": 2.336704441052717e-11,
      "source": "whisper"
    },
    {
      "id": 14,
      "seek": 7770,
      "start": 88.7,
      "end": 92.7,
      "text": "當然,你可以選擇按照參考文稿標點切分。",
      "tokens": [
        50915,
        21707,
        11,
        42766,
        21625,
        13167,
        229,
        26613,
        32150,
        47375,
        26504,
        17174,
        10415,
        123,
        46299,
        8216,
        23632,
        6627,
        1543,
        51115
      ],
      "temperature": 0,
      "avg_logprob": -0.09382529485793341,
      "compression_ratio": 1.1975806451612903,
      "no_speech_prob": 2.336704441052717e-11,
      "source": "whisper"
    },
    {
      "id": 15,
      "seek": 7770,
      "start": 92.7,
      "end": 97.7,
      "text": "進入編輯器後,分距列表,媒體預覽和時間軸會保持聯動。",
      "tokens": [
        51115,
        18214,
        14028,
        12608,
        101,
        12290,
        107,
        34386,
        5661,
        11,
        6627,
        6563,
        251,
        43338,
        17571,
        11,
        44332,
        240,
        23987,
        30156,
        2888,
        121,
        12565,
        20788,
        14454,
        116,
        6236,
        24302,
        17694,
        43942,
        12572,
        1543,
        51365
      ],
      "temperature": 0,
      "avg_logprob": -0.09382529485793341,
      "compression_ratio": 1.1975806451612903,
      "no_speech_prob": 2.336704441052717e-11,
      "source": "whisper"
    },
    {
      "id": 16,
      "seek": 9770,
      "start": 97.7,
      "end": 104.7,
      "text": "右鍵單機自辭間隙可以進行切分,右鍵單機行尾輔助點可以合併下一條。",
      "tokens": [
        50365,
        22060,
        41668,
        113,
        24227,
        17543,
        9722,
        9830,
        255,
        11016,
        10673,
        247,
        6723,
        18214,
        8082,
        23632,
        6627,
        11,
        22060,
        41668,
        113,
        24227,
        17543,
        8082,
        1530,
        122,
        12290,
        242,
        37618,
        8216,
        6723,
        14245,
        1593,
        113,
        45271,
        31106,
        1543,
        50715
      ],
      "temperature": 0,
      "avg_logprob": -0.12838055827830097,
      "compression_ratio": 1.270718232044199,
      "no_speech_prob": 2.0135572029178483e-11,
      "source": "whisper"
    },
    {
      "id": 17,
      "seek": 9770,
      "start": 104.7,
      "end": 110.7,
      "text": "左鍵雙機磁源進行編輯,脫選多個磁源可以進行合併、刪除等操作。",
      "tokens": [
        50715,
        21975,
        41668,
        113,
        6306,
        247,
        17543,
        163,
        96,
        223,
        47402,
        18214,
        8082,
        12608,
        101,
        12290,
        107,
        11,
        27067,
        104,
        21625,
        6392,
        3338,
        163,
        96,
        223,
        47402,
        6723,
        18214,
        8082,
        14245,
        1593,
        113,
        1231,
        2437,
        103,
        32999,
        10187,
        13167,
        235,
        11914,
        1543,
        51015
      ],
      "temperature": 0,
      "avg_logprob": -0.12838055827830097,
      "compression_ratio": 1.270718232044199,
      "no_speech_prob": 2.0135572029178483e-11,
      "source": "whisper"
    },
    {
      "id": 18,
      "seek": 9770,
      "start": 110.7,
      "end": 113.7,
      "text": "拖動時間軸邊界即可調整字幕時間。",
      "tokens": [
        51015,
        6852,
        244,
        12572,
        20788,
        14454,
        116,
        14901,
        21185,
        39127,
        4429,
        32943,
        27662,
        22381,
        40211,
        20788,
        1543,
        51165
      ],
      "temperature": 0,
      "avg_logprob": -0.12838055827830097,
      "compression_ratio": 1.270718232044199,
      "no_speech_prob": 2.0135572029178483e-11,
      "source": "whisper"
    },
    {
      "id": 19,
      "seek": 11370,
      "start": 113.7,
      "end": 120.7,
      "text": "對於相互吸附的字幕,左鍵單機控制鍵即可在聯動和個別調整之間自由切換。",
      "tokens": [
        50365,
        2855,
        19488,
        15106,
        1369,
        240,
        46431,
        8842,
        226,
        1546,
        22381,
        40211,
        11,
        21975,
        41668,
        113,
        24227,
        17543,
        48707,
        25491,
        41668,
        113,
        39127,
        4429,
        3581,
        43942,
        12572,
        12565,
        3338,
        16158,
        32943,
        27662,
        9574,
        11016,
        45233,
        23632,
        36338,
        1543,
        50715
      ],
      "temperature": 0,
      "avg_logprob": -0.0948273108823456,
      "compression_ratio": 1.2062937062937062,
      "no_speech_prob": 1.7371297109103345e-11,
      "source": "whisper"
    },
    {
      "id": 20,
      "seek": 11370,
      "start": 120.7,
      "end": 126.7,
      "text": "AI校準負責處理,標點、英語大小寫、專有名詞和可能的清解錯誤。",
      "tokens": [
        50715,
        48698,
        49684,
        28220,
        11561,
        254,
        11561,
        105,
        39289,
        13876,
        11,
        46299,
        8216,
        1231,
        27869,
        31348,
        3582,
        7322,
        4510,
        104,
        1231,
        46072,
        2412,
        15940,
        6716,
        252,
        12565,
        16657,
        1546,
        21784,
        17278,
        13133,
        3549,
        97,
        1543,
        51015
      ],
      "temperature": 0,
      "avg_logprob": -0.0948273108823456,
      "compression_ratio": 1.2062937062937062,
      "no_speech_prob": 1.7371297109103345e-11,
      "source": "whisper"
    },
    {
      "id": 21,
      "seek": 11370,
      "start": 126.7,
      "end": 130.7,
      "text": "也可以補充要求,比如重點檢查品牌名、機構名或數字。",
      "tokens": [
        51015,
        47935,
        8083,
        250,
        2347,
        227,
        4275,
        32718,
        11,
        36757,
        12624,
        8216,
        40741,
        95,
        42623,
        30246,
        6935,
        234,
        15940,
        1231,
        17543,
        43362,
        15940,
        19780,
        30622,
        22381,
        1543,
        51215
      ],
      "temperature": 0,
      "avg_logprob": -0.0948273108823456,
      "compression_ratio": 1.2062937062937062,
      "no_speech_prob": 1.7371297109103345e-11,
      "source": "whisper"
    },
    {
      "id": 22,
      "seek": 11370,
      "start": 130.7,
      "end": 135.7,
      "text": "它不會重新生成整篇稿子,而是在原文上留下具體的改動、標記。",
      "tokens": [
        51215,
        11284,
        21121,
        12624,
        12560,
        8244,
        11336,
        27662,
        20878,
        229,
        10415,
        123,
        7626,
        11,
        11070,
        1541,
        3581,
        19683,
        17174,
        5708,
        24456,
        4438,
        39806,
        23987,
        1546,
        34490,
        12572,
        1231,
        46299,
        16958,
        1543,
        51465
      ],
      "temperature": 0,
      "avg_logprob": -0.0948273108823456,
      "compression_ratio": 1.2062937062937062,
      "no_speech_prob": 1.7371297109103345e-11,
      "source": "whisper"
    },
    {
      "id": 23,
      "seek": 13570,
      "start": 135.7,
      "end": 144.7,
      "text": "當然,側邊還有經典的查找與替換功能,供你批量對命中的詞源進行編輯,流控就位刪除,隨時可以撤銷。",
      "tokens": [
        50365,
        21707,
        11,
        7437,
        112,
        14901,
        15569,
        15145,
        2347,
        116,
        1546,
        42623,
        25085,
        34127,
        9531,
        123,
        36338,
        32311,
        8225,
        11,
        3254,
        249,
        2166,
        3416,
        117,
        26748,
        2855,
        25236,
        5975,
        1546,
        6716,
        252,
        47402,
        18214,
        8082,
        12608,
        101,
        12290,
        107,
        11,
        27854,
        48707,
        3111,
        11160,
        2437,
        103,
        32999,
        11,
        48890,
        6611,
        6723,
        20559,
        97,
        42041,
        115,
        1543,
        50815
      ],
      "temperature": 0,
      "avg_logprob": -0.08522655826588575,
      "compression_ratio": 1.2961876832844574,
      "no_speech_prob": 1.923594443453691e-11,
      "source": "whisper"
    },
    {
      "id": 24,
      "seek": 13570,
      "start": 144.7,
      "end": 154.7,
      "text": "外部AI審閱會生成一份可以粘貼到網頁端AI的審閱材料,可以檢查全部當前分距或指定範圍,並自動附帶前後上下文。",
      "tokens": [
        50815,
        12022,
        13470,
        48698,
        4510,
        102,
        29602,
        6236,
        8244,
        11336,
        2257,
        36266,
        6723,
        18393,
        246,
        11561,
        120,
        4511,
        26352,
        8313,
        223,
        11957,
        107,
        48698,
        1546,
        4510,
        102,
        29602,
        4422,
        238,
        33404,
        11,
        6723,
        40741,
        95,
        42623,
        38714,
        13118,
        8945,
        6627,
        6563,
        251,
        19780,
        25922,
        12088,
        20878,
        226,
        2523,
        235,
        11,
        35934,
        9722,
        12572,
        8842,
        226,
        32379,
        8945,
        5661,
        5708,
        4438,
        17174,
        1543,
        51315
      ],
      "temperature": 0,
      "avg_logprob": -0.08522655826588575,
      "compression_ratio": 1.2961876832844574,
      "no_speech_prob": 1.923594443453691e-11,
      "source": "whisper"
    },
    {
      "id": 25,
      "seek": 13570,
      "start": 154.7,
      "end": 159.7,
      "text": "遇到拿不准的地方,也可以附上自己的疑問,直接交給你最襯手的網頁端AI進一步判斷。",
      "tokens": [
        51315,
        3330,
        229,
        4511,
        24351,
        1960,
        6336,
        228,
        1546,
        30146,
        11,
        47935,
        8842,
        226,
        5708,
        17645,
        1546,
        18140,
        239,
        11361,
        11,
        43297,
        28455,
        17798,
        2166,
        8661,
        15663,
        107,
        11389,
        1546,
        26352,
        8313,
        223,
        11957,
        107,
        48698,
        18214,
        2257,
        31429,
        2437,
        97,
        4307,
        115,
        1543,
        51565
      ],
      "temperature": 0,
      "avg_logprob": -0.08522655826588575,
      "compression_ratio": 1.2961876832844574,
      "no_speech_prob": 1.923594443453691e-11,
      "source": "whisper"
    },
    {
      "id": 26,
      "seek": 13570,
      "start": 159.7,
      "end": 163.7,
      "text": "AI翻譯支持逐航映射和多對多映射。",
      "tokens": [
        51565,
        48698,
        42716,
        11823,
        107,
        35488,
        2215,
        238,
        10256,
        103,
        1431,
        254,
        1530,
        226,
        12565,
        6392,
        2855,
        6392,
        1431,
        254,
        1530,
        226,
        1543,
        51765
      ],
      "temperature": 0,
      "avg_logprob": -0.08522655826588575,
      "compression_ratio": 1.2961876832844574,
      "no_speech_prob": 1.923594443453691e-11,
      "source": "whisper"
    },
    {
      "id": 27,
      "seek": 16370,
      "start": 163.7,
      "end": 174.7,
      "text": "在多對多映射模式中,模型會先結合上下文理解完整的意思,再把一文放回合適的分句,讓一文既照顧上下文,也符合字幕的閱讀節奏。",
      "tokens": [
        50365,
        3581,
        6392,
        2855,
        6392,
        1431,
        254,
        1530,
        226,
        41908,
        27584,
        5975,
        11,
        41908,
        39823,
        6236,
        10108,
        17144,
        14245,
        5708,
        4438,
        17174,
        13876,
        17278,
        14128,
        27662,
        1546,
        16697,
        11,
        8623,
        16075,
        2257,
        17174,
        12744,
        8350,
        14245,
        3330,
        102,
        1546,
        6627,
        34592,
        11,
        21195,
        2257,
        17174,
        4479,
        95,
        32150,
        10178,
        100,
        5708,
        4438,
        17174,
        11,
        6404,
        5437,
        99,
        14245,
        22381,
        40211,
        1546,
        29602,
        7422,
        222,
        27694,
        1881,
        237,
        1543,
        50915
      ],
      "temperature": 0,
      "avg_logprob": -0.05477287488825181,
      "compression_ratio": 1.3104693140794224,
      "no_speech_prob": 1.8644733326689256e-11,
      "source": "whisper"
    },
    {
      "id": 28,
      "seek": 16370,
      "start": 174.7,
      "end": 177.7,
      "text": "大部分工作完成了,還有一些實用小工具來收尾。",
      "tokens": [
        50915,
        3582,
        32174,
        41315,
        41509,
        2289,
        11,
        15569,
        38515,
        10376,
        9254,
        7322,
        23323,
        39806,
        3763,
        18681,
        1530,
        122,
        1543,
        51065
      ],
      "temperature": 0,
      "avg_logprob": -0.05477287488825181,
      "compression_ratio": 1.3104693140794224,
      "no_speech_prob": 1.8644733326689256e-11,
      "source": "whisper"
    },
    {
      "id": 29,
      "seek": 16370,
      "start": 177.7,
      "end": 185.7,
      "text": "標點處理功能可以分別設置上型字幕和下型字幕中哪些符號應該被刪除,哪些應該被替換為空格。",
      "tokens": [
        51065,
        46299,
        8216,
        39289,
        13876,
        32311,
        8225,
        6723,
        6627,
        16158,
        39035,
        34719,
        5708,
        39823,
        22381,
        40211,
        12565,
        4438,
        39823,
        22381,
        40211,
        5975,
        17028,
        13824,
        5437,
        99,
        18616,
        26087,
        23238,
        2437,
        103,
        32999,
        11,
        17028,
        13824,
        26087,
        23238,
        9531,
        123,
        36338,
        6344,
        23322,
        30921,
        1543,
        51465
      ],
      "temperature": 0,
      "avg_logprob": -0.05477287488825181,
      "compression_ratio": 1.3104693140794224,
      "no_speech_prob": 1.8644733326689256e-11,
      "source": "whisper"
    },
    {
      "id": 30,
      "seek": 18570,
      "start": 185.7,
      "end": 194.7,
      "text": "自動吸附用來批量調整字幕時間,向前智能吸附會尋找可信的起因位置,並保留少量靜音與量。",
      "tokens": [
        50365,
        9722,
        12572,
        46431,
        8842,
        226,
        9254,
        3763,
        3416,
        117,
        26748,
        32943,
        27662,
        22381,
        40211,
        20788,
        11,
        24282,
        8945,
        5094,
        118,
        8225,
        46431,
        8842,
        226,
        6236,
        1530,
        233,
        25085,
        4429,
        17665,
        1546,
        9147,
        8698,
        11160,
        34719,
        11,
        35934,
        24302,
        24456,
        15686,
        26748,
        5363,
        250,
        18034,
        34127,
        26748,
        1543,
        50815
      ],
      "temperature": 0,
      "avg_logprob": -0.11664553611509261,
      "compression_ratio": 1.1597938144329898,
      "no_speech_prob": 1.85585036921454e-11,
      "source": "whisper"
    },
    {
      "id": 31,
      "seek": 18570,
      "start": 194.7,
      "end": 200.7,
      "text": "向後則按預值處理相鄰字幕間的小空隙,減少字幕剛消失又立刻出現的閃爍感。",
      "tokens": [
        50815,
        24282,
        5661,
        46225,
        26613,
        30156,
        40242,
        39289,
        13876,
        15106,
        165,
        26267,
        22381,
        40211,
        11016,
        1546,
        7322,
        23322,
        10673,
        247,
        11,
        9592,
        249,
        15686,
        22381,
        40211,
        16940,
        28837,
        28564,
        17047,
        24409,
        45500,
        7781,
        9581,
        1546,
        5196,
        225,
        8164,
        235,
        9709,
        1543,
        51115
      ],
      "temperature": 0,
      "avg_logprob": -0.11664553611509261,
      "compression_ratio": 1.1597938144329898,
      "no_speech_prob": 1.85585036921454e-11,
      "source": "whisper"
    },
    {
      "id": 32,
      "seek": 20070,
      "start": 200.7,
      "end": 208.7,
      "text": "編輯器還提供減反轉換和收畫人設置,適合不同地區版本以及採訪、播客這類多人素材。",
      "tokens": [
        50365,
        12608,
        101,
        12290,
        107,
        34386,
        7824,
        20949,
        3254,
        249,
        9592,
        249,
        22138,
        32739,
        36338,
        12565,
        18681,
        43822,
        4035,
        39035,
        34719,
        11,
        3330,
        102,
        14245,
        47123,
        10928,
        44526,
        42096,
        8802,
        40282,
        6900,
        94,
        5396,
        103,
        1231,
        49993,
        32316,
        2664,
        40092,
        6392,
        4035,
        37204,
        4422,
        238,
        1543,
        50765
      ],
      "temperature": 0,
      "avg_logprob": -0.10818852992416117,
      "compression_ratio": 1.2712328767123289,
      "no_speech_prob": 1.82114462399241e-11,
      "source": "whisper"
    },
    {
      "id": 33,
      "seek": 20070,
      "start": 208.7,
      "end": 215.7,
      "text": "大公告稱,點擊右上角即可導出,需要交給剪輯軟件就直接導出字幕文件。",
      "tokens": [
        50765,
        3582,
        13545,
        16846,
        49606,
        11,
        8216,
        37813,
        22060,
        5708,
        29389,
        39127,
        4429,
        25647,
        7781,
        11,
        35748,
        28455,
        17798,
        5935,
        103,
        12290,
        107,
        14454,
        253,
        20485,
        3111,
        43297,
        25647,
        7781,
        22381,
        40211,
        17174,
        20485,
        1543,
        51115
      ],
      "temperature": 0,
      "avg_logprob": -0.10818852992416117,
      "compression_ratio": 1.2712328767123289,
      "no_speech_prob": 1.82114462399241e-11,
      "source": "whisper"
    },
    {
      "id": 34,
      "seek": 20070,
      "start": 215.7,
      "end": 220.7,
      "text": "如果還需要換一台電腦繼續使用Substar編輯,也可以把完整工程一起帶走。",
      "tokens": [
        51115,
        13119,
        7824,
        35748,
        36338,
        2257,
        15433,
        20545,
        21184,
        99,
        38459,
        22982,
        9254,
        39582,
        9710,
        12608,
        101,
        12290,
        107,
        11,
        47935,
        16075,
        14128,
        27662,
        23323,
        29649,
        29567,
        32379,
        9575,
        1543,
        51365
      ],
      "temperature": 0,
      "avg_logprob": -0.10818852992416117,
      "compression_ratio": 1.2712328767123289,
      "no_speech_prob": 1.82114462399241e-11,
      "source": "whisper"
    },
    {
      "id": 35,
      "seek": 20070,
      "start": 220.7,
      "end": 229.7,
      "text": "這就是Substar,把聽寫、切分、校準、翻譯、審閱和打軸整合在一個不斷完勝的工作流當中,讓這部工作不再是一件苦差事。",
      "tokens": [
        51365,
        2664,
        5620,
        39582,
        9710,
        11,
        16075,
        21524,
        4510,
        104,
        1231,
        23632,
        6627,
        1231,
        49684,
        28220,
        1231,
        42716,
        11823,
        107,
        1231,
        4510,
        102,
        29602,
        12565,
        12467,
        14454,
        116,
        27662,
        14245,
        3581,
        8990,
        1960,
        4307,
        115,
        14128,
        43434,
        1546,
        41315,
        27854,
        13118,
        5975,
        11,
        21195,
        2664,
        13470,
        41315,
        1960,
        8623,
        1541,
        2257,
        20485,
        30300,
        21679,
        6973,
        1543,
        51815
      ],
      "temperature": 0,
      "avg_logprob": -0.10818852992416117,
      "compression_ratio": 1.2712328767123289,
      "no_speech_prob": 1.82114462399241e-11,
      "source": "whisper"
    }
  ],
  "transcript_selection": {
    "schema": "videonote.transcript_selection.v1",
    "status": "SELECTED",
    "requested_language": "zh",
    "duration": 230.458063,
    "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": 36,
      "multilingual": false,
      "first_start": 0.0,
      "final_end": 229.7,
      "duration_coverage": 0.996711,
      "maximum_gap_seconds": 0.0,
      "inverted_count": 0,
      "language": {
        "requested": "zh",
        "matched": true,
        "latin": 55,
        "han": 1073,
        "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": 230.458,
      "rejected_segment_count": 1,
      "rejected_segments": [
        {
          "index": 36,
          "start": 229.7,
          "end": 231.7,
          "reason": "material_end_timestamp_overrun"
        }
      ]
    }
  }
}