@aituber-onair/core/dist/services/voice/engines/VoiceVoxEngine.js

Core library for AITuber OnAir providing voice synthesis and chat processing

import { VOICE_VOX_API_URL } from '../../../constants';
/**
 * VoiceVox voice synthesis engine
 */
export class VoiceVoxEngine {
    constructor() {
        this.apiEndpoint = VOICE_VOX_API_URL;
    }
    async fetchAudio(input, speaker) {
        const talk = input;
        // get emotion from talk.style
        const emotion = talk.style || 'neutral';
        const ttsQueryResponse = await fetch(`${this.apiEndpoint}/audio_query?speaker=${speaker}&text=${encodeURIComponent(talk.message)}`, { method: 'POST' });
        if (!ttsQueryResponse.ok) {
            throw new Error('Failed to fetch TTS query.');
        }
        const ttsQueryJson = await ttsQueryResponse.json();
        // adjust parameters according to emotion
        this.adjustEmotionParameters(ttsQueryJson, emotion);
        const synthesisResponse = await fetch(`${this.apiEndpoint}/synthesis?speaker=${speaker}`, {
            method: 'POST',
            headers: {
                'Content-Type': 'application/json',
                'Transfer-Encoding': 'chunked',
            },
            body: JSON.stringify(ttsQueryJson),
        });
        if (!synthesisResponse.ok) {
            throw new Error('Failed to fetch TTS synthesis result.');
        }
        const blob = await synthesisResponse.blob();
        return await blob.arrayBuffer();
    }
    /**
     * Adjust parameters according to emotion
     */
    adjustEmotionParameters(ttsQueryJson, emotion) {
        // default values
        ttsQueryJson.speedScale = 1.16;
        ttsQueryJson.pitchScale = -0.02;
        ttsQueryJson.intonationScale = 1.26;
        switch (emotion.toLowerCase()) {
            case 'happy':
                ttsQueryJson.speedScale = 1.25;
                ttsQueryJson.pitchScale = 0.05;
                ttsQueryJson.intonationScale = 1.4;
                break;
            case 'sad':
                ttsQueryJson.speedScale = 1.0;
                ttsQueryJson.pitchScale = -0.1;
                ttsQueryJson.intonationScale = 1.0;
                break;
            case 'angry':
                ttsQueryJson.speedScale = 1.2;
                ttsQueryJson.pitchScale = -0.05;
                ttsQueryJson.intonationScale = 1.5;
                break;
            case 'surprised':
                ttsQueryJson.speedScale = 1.3;
                ttsQueryJson.pitchScale = 0.1;
                ttsQueryJson.intonationScale = 1.4;
                break;
            // default: "neutral" etc. other than default values
        }
    }
    getTestMessage(textVoiceText) {
        return textVoiceText || 'ボイスボックスを使用します';
    }
    /**
     * Set custom API endpoint URL
     * @param apiUrl custom API endpoint URL
     */
    setApiEndpoint(apiUrl) {
        this.apiEndpoint = apiUrl;
    }
}
//# sourceMappingURL=VoiceVoxEngine.js.map