feat: use VOICEVOX API for TTS
This commit is contained in:
@ -1,37 +1,69 @@
|
|||||||
import fs from "fs";
|
import fs from "fs";
|
||||||
import path from "path";
|
import path from "path";
|
||||||
import { PollyClient, SynthesizeSpeechCommand } from "@aws-sdk/client-polly";
|
import fetch from "node-fetch";
|
||||||
|
|
||||||
const polly = new PollyClient({ region: "ap-northeast-1" });
|
// VOICEVOX APIの設定
|
||||||
|
const VOICEVOX_HOST = "http://localhost:50021";
|
||||||
|
|
||||||
|
interface VoiceStyle {
|
||||||
|
speakerId: number;
|
||||||
|
styleId: number;
|
||||||
|
}
|
||||||
|
|
||||||
|
// 仮の声設定(例: サイドM = 3)
|
||||||
|
const defaultVoiceStyle: VoiceStyle = {
|
||||||
|
speakerId: 3,
|
||||||
|
styleId: 2,
|
||||||
|
};
|
||||||
|
|
||||||
export async function generateTTS(
|
export async function generateTTS(
|
||||||
itemId: string,
|
itemId: string,
|
||||||
scriptText: string,
|
scriptText: string,
|
||||||
): Promise<string> {
|
): Promise<string> {
|
||||||
const params = {
|
// 音声合成クエリの生成
|
||||||
OutputFormat: "mp3",
|
const queryResponse = await fetch(`${VOICEVOX_HOST}/audio_query`, {
|
||||||
Text: scriptText,
|
method: "POST",
|
||||||
VoiceId: "Mizuki",
|
headers: {
|
||||||
LanguageCode: "ja-JP",
|
"Content-Type": "application/json",
|
||||||
};
|
},
|
||||||
const command = new SynthesizeSpeechCommand(params);
|
body: JSON.stringify({
|
||||||
const response = await polly.send(command);
|
text: scriptText,
|
||||||
|
speaker: defaultVoiceStyle.speakerId,
|
||||||
|
}),
|
||||||
|
});
|
||||||
|
|
||||||
if (!response.AudioStream) {
|
if (!queryResponse.ok) {
|
||||||
throw new Error("TTSのAudioStreamが空です");
|
throw new Error("VOICEVOX 音声合成クエリ生成に失敗しました");
|
||||||
}
|
}
|
||||||
|
|
||||||
|
const audioQuery = await queryResponse.json();
|
||||||
|
|
||||||
|
// 音声合成
|
||||||
|
const audioResponse = await fetch(`${VOICEVOX_HOST}/synthesis`, {
|
||||||
|
method: "POST",
|
||||||
|
headers: {
|
||||||
|
"Content-Type": "application/json",
|
||||||
|
},
|
||||||
|
body: JSON.stringify({
|
||||||
|
audio_query: audioQuery,
|
||||||
|
speaker: defaultVoiceStyle.speakerId,
|
||||||
|
}),
|
||||||
|
});
|
||||||
|
|
||||||
|
if (!audioResponse.ok) {
|
||||||
|
throw new Error("VOICEVOX 音声合成に失敗しました");
|
||||||
|
}
|
||||||
|
|
||||||
|
const audioBuffer = await audioResponse.buffer();
|
||||||
|
|
||||||
|
// 出力ディレクトリの準備
|
||||||
const outputDir = path.join(__dirname, "../public/podcast_audio");
|
const outputDir = path.join(__dirname, "../public/podcast_audio");
|
||||||
if (!fs.existsSync(outputDir)) {
|
if (!fs.existsSync(outputDir)) {
|
||||||
fs.mkdirSync(outputDir, { recursive: true });
|
fs.mkdirSync(outputDir, { recursive: true });
|
||||||
}
|
}
|
||||||
const filePath = path.resolve(outputDir, `${itemId}.mp3`);
|
|
||||||
|
|
||||||
const chunks: Uint8Array[] = [];
|
const filePath = path.resolve(outputDir, `${itemId}.mp3`);
|
||||||
for await (const chunk of response.AudioStream as any) {
|
fs.writeFileSync(filePath, audioBuffer);
|
||||||
chunks.push(chunk);
|
|
||||||
}
|
|
||||||
const buffer = Buffer.concat(chunks);
|
|
||||||
fs.writeFileSync(filePath, buffer);
|
|
||||||
return filePath;
|
return filePath;
|
||||||
}
|
}
|
||||||
|
Reference in New Issue
Block a user