GladiaのSTTモデル「Solaria」を試す

export GLADIA_API_KEY=XXXXXXXXXX

curl --request POST \
  --url https://api.gladia.io/v2/upload \
  --header 'Content-Type: multipart/form-data' \
  --header "x-gladia-key: $GLADIA_API_KEY" \
  --form audio=@voice_lunch_jp_5min.wav

{
    "audio_url":"https://api.gladia.io/file/XXXXXXXX-XXXX-XXXX-XXXX-XXXXXXXXXXXX",
    "audio_metadata":{
        "id":"XXXXXXXX-XXXX-XXXX-XXXX-XXXXXXXXXXXX",
        "filename":"voice_lunch_jp_5min.wav",
        "extension":"wav",
        "size":8600312,
        "audio_duration":268.757,
        "number_of_channels":1
    }
}

  curl --request POST \
    --url https://api.gladia.io/v2/pre-recorded \
    --header 'Content-Type: application/json' \
    --header "x-gladia-key: $GLADIA_API_KEY" \
    --data '{
    "audio_url": "https://api.gladia.io/file/XXXXXXXX-XXXX-XXXX-XXXX-XXXXXXXXXXXX",
    "diarization": true,
    "diarization_config": {
      "number_of_speakers": 3,
      "min_speakers": 1,
      "max_speakers": 5
    },
    "translation": true,
    "translation_config": {
      "model": "base",
      "target_languages": ["ja", "en"]
    },
    "subtitles": true,
    "subtitles_config": {
      "formats": ["srt", "vtt"]
    },
    "detect_language": true,
    "enable_code_switching": false
    }
  '

{
    "id":"YYYYYYYY-YYYY-YYYY-YYYY-YYYYYYYYYYYY",
    "result_url":"https://api.gladia.io/v2/pre-recorded/YYYYYYYY-YYYY-YYYY-YYYY-YYYYYYYYYYYY"
}

curl --request GET \
  --url https://api.gladia.io/v2/pre-recorded/YYYYYYYY-YYYY-YYYY-YYYY-YYYYYYYYYYYY \
    --header "x-gladia-key: $GLADIA_API_KEY" | jq -r .

{
  "id": "YYYYYYYY-YYYY-YYYY-YYYY-YYYYYYYYYYYY",
  "request_id": "G-dfb9a7fc",
  "version": 2,
  "status": "processing",
  "created_at": "2025-04-11T08:35:31.066Z",
  "completed_at": null,
  "custom_metadata": null,
  "error_code": null,
  "kind": "pre-recorded",
  "file": {
    "id": "XXXXXXXX-XXXX-XXXX-XXXX-XXXXXXXXXXXX",
    "filename": "voice_lunch_jp_5min.wav",
    "source": null,
    "audio_duration": 268.757,
    "number_of_channels": 1
  },
  "request_params": {
    "audio_url": "https://api.gladia.io/file/XXXXXXXX-XXXX-XXXX-XXXX-XXXXXXXXXXXX",
    "sentences": false,
    "subtitles": true,
    "moderation": false,
    "diarization": true,
    "translation": true,
    "audio_to_llm": false,
    "display_mode": false,
    "summarization": false,
    "audio_enhancer": true,
    "chapterization": false,
    "custom_spelling": false,
    "detect_language": true,
    "name_consistency": false,
    "subtitles_config": {
      "style": "default",
      "formats": [
        "srt",
        "vtt"
      ]
    },
    "diarization_config": {
      "enhanced": false,
      "max_speakers": 5,
      "min_speakers": 1,
      "number_of_speakers": 3
    },
    "sentiment_analysis": false,
    "translation_config": {
      "model": "base",
      "target_languages": [
        "ja",
        "en"
      ],
      "match_original_utterances": true
    },
    "diarization_enhanced": false,
    "punctuation_enhanced": false,
    "enable_code_switching": false,
    "named_entity_recognition": false,
    "speaker_reidentification": false,
    "accurate_words_timestamps": false,
    "skip_channel_deduplication": false,
    "structured_data_extraction": false
  },
  "result": null
}

{
  "id": "YYYYYYYY-YYYY-YYYY-YYYY-YYYYYYYYYYYY",
  "request_id": "G-dfb9a7fc",
  "version": 2,
  "status": "done",
  "created_at": "2025-04-11T08:35:31.066Z",
  "completed_at": "2025-04-11T08:36:26.986Z",
  "custom_metadata": null,
  "error_code": null,
  "kind": "pre-recorded",
  "file": {
    "id": "XXXXXXXX-XXXX-XXXX-XXXX-XXXXXXXXXXXX",
    "filename": "voice_lunch_jp_5min.wav",
    "source": null,
    "audio_duration": 268.757,
    "number_of_channels": 1
  },
  "request_params": {
    "audio_url": "https://api.gladia.io/file/XXXXXXXX-XXXX-XXXX-XXXX-XXXXXXXXXXXX",
    "sentences": false,
    "subtitles": true,
    "moderation": false,
    "diarization": true,
    "translation": true,
    "audio_to_llm": false,
    "display_mode": false,
    "summarization": false,
    "audio_enhancer": true,
    "chapterization": false,
    "custom_spelling": false,
    "detect_language": true,
    "name_consistency": false,
    "subtitles_config": {
      "style": "default",
      "formats": [
        "srt",
        "vtt"
      ]
    },
    "diarization_config": {
      "enhanced": false,
      "max_speakers": 5,
      "min_speakers": 1,
      "number_of_speakers": 3
    },
    "sentiment_analysis": false,
    "translation_config": {
      "model": "base",
      "target_languages": [
        "ja",
        "en"
      ],
      "match_original_utterances": true
    },
    "diarization_enhanced": false,
    "punctuation_enhanced": false,
    "enable_code_switching": false,
    "named_entity_recognition": false,
    "speaker_reidentification": false,
    "accurate_words_timestamps": false,
    "skip_channel_deduplication": false,
    "structured_data_extraction": false
  },
  "result": {
    "metadata": {
      "audio_duration": 268.757313,
      "number_of_distinct_channels": 1,
      "billing_time": 268.757313,
      "transcription_time": 55.92
    },
    "transcription": {
      "utterances": [
        {
          "words": [
            {
              "word": "はい",
              "start": 0.876,
              "end": 1.096,
              "confidence": 0.45
            },
            {
              "word": "じゃあ",
              "start": 1.356,
              "end": 1.696,
              "confidence": 0.46
            },
            {
              "word": "始め",
              "start": 1.776,
              "end": 1.976,
              "confidence": 0.43
            },
            {
              "word": "ます",
              "start": 2.056,
              "end": 2.336,
              "confidence": 0.86
            },
(snip)

(snip)
    "translation": true,
    "translation_config": {
      "model": "base",
      "target_languages": ["ja", "en"]
    },
    "subtitles": true,
    "subtitles_config": {
      "formats": ["srt", "vtt"]
    },
(snip)

curl --request GET \
    --url https://api.gladia.io/v2/pre-recorded/YYYYYYYY-YYYY-YYYY-YYYY-YYYYYYYYYYYY \
    --header "x-gladia-key: $GLADIA_API_KEY" > result.json

jq -r .result.transcription.full_transcript result.json

jq -r .result.transcription.subtitles[0].subtitles result.json

jq -r .result.transcription.utterances result.json

[
  {
    "words": [
      {
        "word": "はい",
        "start": 0.876,
        "end": 1.096,
        "confidence": 0.45
      },
      {
        "word": "じゃあ",
        "start": 1.356,
        "end": 1.696,
        "confidence": 0.46
      },
      {
        "word": "始め",
        "start": 1.776,
        "end": 1.976,
        "confidence": 0.43
      },
      {
        "word": "ます",
        "start": 2.056,
        "end": 2.336,
        "confidence": 0.86
      },
      {
        "word": "ちょっと",
        "start": 2.356,
(snip)

jq -r .result.translation.results[1].full_transcript result.json

jq -r .result.translation.results[1].subtitles[1].subtitles result.json

jq -r .result.translation.results[1].utterances result.json

[
  {
    "words": [
      {
        "word": "Okay,",
        "start": 0.876,
        "end": 1.096,
        "confidence": 0.45
      },
      {
        "word": " then",
        "start": 1.356,
        "end": 1.696,
        "confidence": 0.46
      },
      {
        "word": " let's",
        "start": 1.776,
        "end": 1.976,
        "confidence": 0.43
      },
      {
        "word": " begin.",
        "start": 2.056,
        "end": 2.336,
        "confidence": 0.86
      },
      {
        "word": " There",
        "start": 2.356,
(snip)

{
    "audio_url": "https://api.gladia.io/file/XXXXXXXX-XXXX-XXXX-XXXX-XXXXXXXXXXXX",
    "punctuation_enhanced": true
}

{
    "audio_url": "https://api.gladia.io/file/XXXXXXXX-XXXX-XXXX-XXXX-XXXXXXXXXXXX",
    "sentences": true
}

{
    "audio_url": "https://api.gladia.io/file/XXXXXXXX-XXXX-XXXX-XXXX-XXXXXXXXXXXX",
    "detect_language": true
}

{
    "audio_url": "https://api.gladia.io/file/XXXXXXXX-XXXX-XXXX-XXXX-XXXXXXXXXXXX",
    "detect_language": false,
    "language": "ja"
}

{
    "audio_url": "https://api.gladia.io/file/XXXXXXXX-XXXX-XXXX-XXXX-XXXXXXXXXXXX",
    "enable_code_switching": true
}

{
    "audio_url": "https://api.gladia.io/file/XXXXXXXX-XXXX-XXXX-XXXX-XXXXXXXXXXXX",
    "enable_code_switching": true,
    "code_switching_config": {
        "languages": ["en", "es", "fr"]
    }
}

{
    "audio_url": "https://api.gladia.io/file/XXXXXXXX-XXXX-XXXX-XXXX-XXXXXXXXXXXX",
    "subtitles": true,
    "subtitles_config": {
      "formats": ["srt", "vtt"],
      "minimum_duration": 1,
      "maximum_duration": 5,
      "maximum_characters_per_row": 42,
      "maximum_rows_per_caption": 2,
      "style": "compliance"
    }
}

{
    "audio_url": "https://api.gladia.io/file/XXXXXXXX-XXXX-XXXX-XXXX-XXXXXXXXXXXX",
    "context_prompt": "ゲーム・オブ・スローンズシリーズのサンサ・スタークとピーター・ベーリッシュの会話。"
}

{
    "audio_url": "https://api.gladia.io/file/XXXXXXXX-XXXX-XXXX-XXXX-XXXXXXXXXXXX",
    "custom_vocabulary": true,
    "custom_vocabulary_config": {
        "vocabulary": [
            "Westeros", 
            {"value": "Stark"}, 
            { 
                "value": "Night's Watch",
                "pronunciations": ["Nightz Vatch"],
                "intensity": 0.4,
                "language": "de"
            }
        ],
        "default_intensity": 0.6
    }
}

{
    "custom_spelling": true,
    "custom_spelling_config": {
        "spelling_dictionary": {
            "Gorish": ["ghorish", "gaurish", "gaureish"],
            "Data Science": ["data-science", "data science"],
            ".": ["period", "full stop"],
            "SQL": ["sequel"]
        }
    }
}

{
    "audio_url": "https://api.gladia.io/file/XXXXXXXX-XXXX-XXXX-XXXX-XXXXXXXXXXXX",
    "name_consistency": true
}

{
    "audio_url": "https://api.gladia.io/file/XXXXXXXX-XXXX-XXXX-XXXX-XXXXXXXXXXXX",
    "diarization": true
}

{
    "audio_url": "https://api.gladia.io/file/XXXXXXXX-XXXX-XXXX-XXXX-XXXXXXXXXXXX",
    "diarization": true,
    "diarization_config": {
        "enhanced": true
    }
}

{
    "diarization": true,
    "diarization_config": {
      "number_of_speakers": 3,
      "min_speakers": 1,
      "max_speakers": 5
    }
}

git clone https://github.com/gladiaio/gladia-samples && cd gladia-samples

cd python/src

uv init -p 3.12.9

uv add -r requirements.txt

 + certifi==2025.1.31
 + charset-normalizer==3.4.1
 + idna==3.10
 + pyaudio==0.2.14
 + python-dotenv==1.0.1
 + requests==2.32.3
 + urllib3==2.4.0
 + websockets==13.1

ls streaming/

live-from-file-with-resume.py		live-from-microphone-with-resume.py
live-from-file.py			live-from-microphone.py

import asyncio
import base64
import json
import signal
import sys
from datetime import time
from typing import Literal, TypedDict

import pyaudio
import requests
from websockets.asyncio.client import ClientConnection, connect
from websockets.exceptions import ConnectionClosedOK

## 定数
GLADIA_API_URL = "https://api.gladia.io"


## 型定義

class InitiateResponse(TypedDict):
    """初期化レスポンス"""
    id: str
    url: str

class LanguageConfiguration(TypedDict):
    """言語設定"""
    languages: list[str] | None
    code_switching: bool | None

class StreamingConfiguration(TypedDict):
    """ストリーミング設定
    以下は限定されたオプションの一部です。全てのオプションについては、APIドキュメントを参照してください。
    https://docs.gladia.io/ja/api-reference/v2/live/init
    """
    encoding: Literal["wav/pcm", "wav/alaw", "wav/ulaw"]
    bit_depth: Literal[8, 16, 24, 32]
    sample_rate: Literal[8_000, 16_000, 32_000, 44_100, 48_000]
    channels: int
    language_config: LanguageConfiguration | None


## ヘルパー関数

def get_gladia_key() -> str:
    """Gladia APIキーを取得"""
    if len(sys.argv) != 2 or not sys.argv[1]:
        print("You must provide a Gladia key as the first argument.")
        exit(1)
    return sys.argv[1]


# デフォルトのマイク設定を取得
def get_mic_settings():
    """デフォルトマイクの設定を取得する"""
    p = pyaudio.PyAudio()
    # デフォルト入力デバイスのインデックスを取得
    default_input_device_index = p.get_default_input_device_info()["index"]
    device_info = p.get_device_info_by_index(default_input_device_index)
    
    sample_rate = int(device_info["defaultSampleRate"])
    channels = int(device_info["maxInputChannels"])
    
    # チャンネル数をチェック
    if channels <= 0:
        raise ValueError("使用可能な入力チャンネルがありません。マイクが接続されているか確認してください。")
    
    # 情報を表示
    print(f"デフォルトマイク: {device_info['name']}")
    print(f"  最大入力チャンネル: {channels}")
    print(f"  デフォルトサンプルレート: {sample_rate} Hz")
    
    # フレームバッファサイズを計算（約100msの音声データ）
    frames_per_buffer = int(sample_rate * 0.1)
    
    return {
        "sample_rate": sample_rate,
        "channels": channels,
        "format": pyaudio.paInt16,  # 16ビット整数形式
        "frames_per_buffer": frames_per_buffer
    }


def init_live_session(config: StreamingConfiguration) -> InitiateResponse:
    """ライブセッションを開始"""
    gladia_key = get_gladia_key()
    response = requests.post(
        f"{GLADIA_API_URL}/v2/live",
        headers={"X-Gladia-Key": gladia_key},
        json=config,
        timeout=3,
    )
    if not response.ok:
        print(f"{response.status_code}: {response.text or response.reason}")
        exit(response.status_code)
    return response.json()


def format_duration(seconds: float) -> str:
    """秒をHH:MM:SS.mmm形式に変換"""
    milliseconds = int(seconds * 1_000)
    return time(
        hour=milliseconds // 3_600_000,
        minute=(milliseconds // 60_000) % 60,
        second=(milliseconds // 1_000) % 60,
        microsecond=milliseconds % 1_000 * 1_000,
    ).isoformat(timespec="milliseconds")


async def print_messages_from_socket(socket: ClientConnection) -> None:
    """WebSocketからメッセージを受信して、最終的なテキストを出力"""
    async for message in socket:
        content = json.loads(message)
        if content["type"] == "transcript" and content["data"]["is_final"]:
            start = format_duration(content["data"]["utterance"]["start"])
            end = format_duration(content["data"]["utterance"]["end"])
            text = content["data"]["utterance"]["text"].strip()
            print(f"{start} --> {end} | {text}")
        if content["type"] == "post_final_transcript":
            print("\n################ セッション終了 ################\n")
            print(json.dumps(content, indent=2, ensure_ascii=False))


async def stop_recording(websocket: ClientConnection) -> None:
    """録音を停止"""
    print(">>>>> 録音を終了します...")
    await websocket.send(json.dumps({"type": "stop_recording"}))
    await asyncio.sleep(0)


async def send_audio(socket: ClientConnection) -> None:
    """音声を送信"""
    stream = P.open(
        format=FORMAT,
        channels=CHANNELS,
        rate=SAMPLE_RATE,
        input=True,
        frames_per_buffer=FRAMES_PER_BUFFER,
    )

    while True:
        data = stream.read(FRAMES_PER_BUFFER)
        data = base64.b64encode(data).decode("utf-8")
        json_data = json.dumps({"type": "audio_chunk", "data": {"chunk": str(data)}})
        try:
            await socket.send(json_data)
            await asyncio.sleep(0.1)  # Send audio every 100ms
        except ConnectionClosedOK:
            return


## サンプルコード
P = pyaudio.PyAudio()

# マイク設定を取得
mic_settings = get_mic_settings()

# Gladiaがサポートするサンプルレートを確認
supported_rates = StreamingConfiguration.__annotations__["sample_rate"].__args__
sample_rate = mic_settings["sample_rate"]

# マイクのサンプルレートがサポートされているか確認
if sample_rate not in supported_rates:
    # 最も近いサポートされているサンプルレートを選択
    closest_rate = min(supported_rates, key=lambda x: abs(x - sample_rate))
    print(f"\u8b66告: マイクのサンプルレート {sample_rate} Hz はサポートされていません")
    print(f"  代わりに {closest_rate} Hz を使用します")
    sample_rate = closest_rate

# ビット深度は16ビットに固定
bit_depth = 16

# ビット深度がサポートされているか確認
supported_bit_depths = StreamingConfiguration.__annotations__["bit_depth"].__args__
if bit_depth not in supported_bit_depths:
    raise ValueError(f"ビット深度 {bit_depth} はサポートされていません。サポートされている値: {supported_bit_depths}")

# エンコーディングは固定
encoding = "wav/pcm"

# エンコーディングがサポートされているか確認
supported_encodings = StreamingConfiguration.__annotations__["encoding"].__args__
if encoding not in supported_encodings:
    raise ValueError(f"エンコーディング {encoding} はサポートされていません。サポートされている値: {supported_encodings}")

# チャンネル数はマイクの設定から取得
channels = mic_settings["channels"]

print(f"使用する設定:")
print(f"  サンプルレート: {sample_rate} Hz")
print(f"  ビット深度: {bit_depth} bit")
print(f"  エンコーディング: {encoding}")
print(f"  チャンネル数: {channels}")
print(f"  フレームバッファサイズ: {mic_settings['frames_per_buffer']}")

# PyAudio設定
FORMAT = mic_settings["format"]
CHANNELS = channels
SAMPLE_RATE = sample_rate
FRAMES_PER_BUFFER = mic_settings["frames_per_buffer"]

# Gladia API設定
STREAMING_CONFIGURATION: StreamingConfiguration = {
    "encoding": encoding,
    "sample_rate": sample_rate,
    "bit_depth": bit_depth,
    "channels": channels,
    "language_config": {
        "languages": [],
        "code_switching": True,
    }
}

async def main():
    """メイン処理"""
    response = init_live_session(STREAMING_CONFIGURATION)
    async with connect(response["url"]) as websocket:
        print("\n################ セッション開始 ################\n")
        loop = asyncio.get_running_loop()
        loop.add_signal_handler(
            signal.SIGINT,
            loop.create_task,
            stop_recording(websocket),
        )

        send_audio_task = asyncio.create_task(send_audio(websocket))
        print_messages_task = asyncio.create_task(print_messages_from_socket(websocket))
        await asyncio.wait(
            [send_audio_task, print_messages_task],
        )


if __name__ == "__main__":
    asyncio.run(main())

uv run streaming/live-from-microphone.py XXXXXXXXXX

デフォルトマイク: Jabra SPEAK 510 USB
  最大入力チャンネル: 1
  デフォルトサンプルレート: 16000 Hz
使用する設定:
  サンプルレート: 16000 Hz
  ビット深度: 16 bit
  エンコーディング: wav/pcm
  チャンネル数: 1
  フレームバッファサイズ: 1600

################ セッション開始 ################

00:00:02.556 --> 00:00:06.564 | 今日も新幹線をご利用くださいましてありがとうございます。
00:00:07.392 --> 00:00:10.408 | この電車はのぞみ号東京行きです
00:00:11.108 --> 00:00:13.484 | 途中の停車駅は京都
00:00:13.608 --> 00:00:16.944 | 名古屋、新横浜、品川です
^C>>>>> 録音を終了します...

################ セッション終了 ################

{
  "type": "post_final_transcript",
  "session_id": "c7db9295-3f54-46f3-beba-c3802c7e9b30",
  "created_at": "2025-04-11T12:07:34.313Z",
  "data": {
    "metadata": {
      "audio_duration": 20,
      "billing_time": 20,
      "number_of_distinct_channels": 1,
      "transcription_time": 22.518
    },
    "transcription": {
      "languages": [
        "ja"
      ],
      "utterances": [
        {
          "text": "今日も新幹線をご利用くださいましてありがとうございます。",
          "start": 2.556,
          "end": 6.564,
          "language": "ja",
          "confidence": 1,
          "channel": 0,
          "words": [
            {
              "word": "今日も新幹線をご利用くださいましてありがとうございます。",
              "start": 2.556,
              "end": 6.564,
              "confidence": 1
            }
          ]
        },
        {
          "text": "この電車はのぞみ号東京行きです",
          "start": 7.392,
          "end": 10.408,
          "language": "ja",
          "confidence": 1,
          "channel": 0,
          "words": [
            {
              "word": "この電車はのぞみ号東京行きです",
              "start": 7.392,
              "end": 10.408,
              "confidence": 1
            }
          ]
        },
        {
          "text": "途中の停車駅は京都",
          "start": 11.108,
          "end": 13.484,
          "language": "ja",
          "confidence": 1,
          "channel": 0,
          "words": [
            {
              "word": "途中の停車駅は京都",
              "start": 11.108,
              "end": 13.484,
              "confidence": 1
            }
          ]
        },
        {
          "text": "名古屋、新横浜、品川です",
          "start": 13.608,
          "end": 16.944,
          "language": "ja",
          "confidence": 1,
          "channel": 0,
          "words": [
            {
              "word": "名古屋、新横浜、品川です",
              "start": 13.608,
              "end": 16.944,
              "confidence": 1
            }
          ]
        }
      ],
      "full_transcript": "今日も新幹線をご利用くださいましてありがとうございます。この電車はのぞみ号東京行きです途中の停車駅は京都名古屋、新横浜、品川です"
    }
  }
}

(snip)

class TranslationConfiguration(TypedDict):
    target_languages: list[str] | None

class RealtimeProcessing(TypedDict):
    translation: bool | None
    translation_config: TranslationConfiguration | None

class StreamingConfiguration(TypedDict):
    encoding: Literal["wav/pcm", "wav/alaw", "wav/ulaw"]
    bit_depth: Literal[8, 16, 24, 32]
    sample_rate: Literal[8_000, 16_000, 32_000, 44_100, 48_000]
    channels: int
    language_config: LanguageConfiguration | None
    realtime_processing: RealtimeProcessing | None

(snip)

async def print_messages_from_socket(socket: ClientConnection) -> None:
    async for message in socket:
        content = json.loads(message)
        if content["type"] == "transcript" and content["data"]["is_final"]:
            start = format_duration(content["data"]["utterance"]["start"])
            end = format_duration(content["data"]["utterance"]["end"])
            text = content["data"]["utterance"]["text"].strip()
            print(f"{start} --> {end} | {text}")
        # 翻訳の場合を追加
        if content["type"] == "translation":
            start = format_duration(content["data"]["utterance"]["start"])
            end = format_duration(content["data"]["utterance"]["end"])
            language = content["data"]["target_language"]
            translation = content["data"]["translated_utterance"]["text"].strip()
            print(f"{start} --> {end} | ({language}) {translation}")
        # ここまで
        if content["type"] == "post_final_transcript":
            print("\n################ セッション終了 ################\n")
            print(json.dumps(content, indent=2, ensure_ascii=False))

(snip)

STREAMING_CONFIGURATION: StreamingConfiguration = {
    "encoding": encoding,
    "sample_rate": sample_rate,
    "bit_depth": bit_depth,
    "channels": channels,
    "language_config": {
        "languages": ["ja"],
        "code_switching": True,
    },
    "realtime_processing": {
        "translation": True,
        "translation_config": {
          "target_languages": ["en", "ko"]
        }
    }
}
(snip)

デフォルトマイク: Jabra SPEAK 510 USB
  最大入力チャンネル: 1
  デフォルトサンプルレート: 16000 Hz
使用する設定:
  サンプルレート: 16000 Hz
  ビット深度: 16 bit
  エンコーディング: wav/pcm
  チャンネル数: 1
  フレームバッファサイズ: 1600

################ セッション開始 ################

00:00:04.224 --> 00:00:08.135 | 今日も新幹線をご利用くださいましてありがとうございます。
00:00:04.224 --> 00:00:08.135 | (en) Thank you for using the Shinkansen today.
00:00:04.224 --> 00:00:08.135 | (ko) 오 늘 도 칸 센 을 용 해 셔 서 사 합 니 다.
00:00:09.793 --> 00:00:12.649 | この電車はのぞみ号東京行きです。
00:00:09.793 --> 00:00:12.649 | (en) This train is the Nozomi express bound for Tokyo.
00:00:09.793 --> 00:00:12.649 | (ko) こ の 電 車 は の ぞ み 号 東 京 行 き で す。
00:00:13.725 --> 00:00:16.229 | 途中の停車駅は京都
00:00:13.725 --> 00:00:16.229 | (en) The stop along the way is Kyoto.
00:00:13.725 --> 00:00:16.229 | (ko) 중 간 에 차 하 는 은 토 입 니 다.
00:00:16.513 --> 00:00:19.881 | 名古屋、新横浜、品川です
00:00:16.513 --> 00:00:19.881 | (en) Nagoya, Shin-Yokohama, Shinagawa.
00:00:16.513 --> 00:00:19.881 | (ko) 名 古 屋 、 横 浜 、 川 で す。
^C>>>>> 録音を終了します...

################ セッション終了 ################

{
  "type": "post_final_transcript",
  "session_id": "9da04472-1dfe-4eeb-9248-b9e7bdb9745c",
  "created_at": "2025-04-11T12:39:59.607Z",
  "data": {
    "metadata": {
      "audio_duration": 25,
      "billing_time": 25,
      "number_of_distinct_channels": 1,
      "transcription_time": 27.431
    },
    "transcription": {
      "languages": [
        "ja"
      ],
      "utterances": [
        {
          "text": "今日も新幹線をご利用くださいましてありがとうございます。",
          "start": 4.224,
          "end": 8.136,
          "language": "ja",
          "confidence": 1,
          "channel": 0,
          "words": [
            {
              "word": "今日も新幹線をご利用くださいましてありがとうございます。",
              "start": 4.224,
              "end": 8.136,
              "confidence": 1
            }
          ]
        },
        {
          "text": "この電車はのぞみ号東京行きです。",
          "start": 9.793,
          "end": 12.649,
          "language": "ja",
          "confidence": 1,
          "channel": 0,
          "words": [
            {
              "word": "この電車はのぞみ号東京行きです。",
              "start": 9.793,
              "end": 12.649,
              "confidence": 1
            }
          ]
        },
        {
          "text": "途中の停車駅は京都",
          "start": 13.725,
          "end": 16.229,
          "language": "ja",
          "confidence": 1,
          "channel": 0,
          "words": [
            {
              "word": "途中の停車駅は京都",
              "start": 13.725,
              "end": 16.229,
              "confidence": 1
            }
          ]
        },
        {
          "text": "名古屋、新横浜、品川です",
          "start": 16.513,
          "end": 19.881,
          "language": "ja",
          "confidence": 1,
          "channel": 0,
          "words": [
            {
              "word": "名古屋、新横浜、品川です",
              "start": 16.513,
              "end": 19.881,
              "confidence": 1
            }
          ]
        }
      ],
      "full_transcript": "今日も新幹線をご利用くださいましてありがとうございます。この電車はのぞみ号東京行きです。途中の停車駅は京都名古屋、新横浜、品川です"
    }
  }
}