stt-sdk/src/base/BaseSTTClientUtil.ts

基于 LLMs 的语音转文本 SDK

/**
 * STT 客户端工具合集
 */
export abstract class BaseSTTClientUtil {
    /**
     * 默认编码选择器
     */
    public pickSupportedMime(): string {
        const cands = [
            "audio/webm;codecs=opus",
            "audio/ogg;codecs=opus",
            "audio/webm",
        ];
        return cands.find((t) => MediaRecorder.isTypeSupported(t)) ?? "audio/webm";
    }

    /**
     * 默认音频流选择器
     */
    public defaultVoiceStream(): Promise<MediaStream> {
        return new Promise(async (resolve, reject) => {
            try {
                const stream = await navigator.mediaDevices.getUserMedia({
                    audio: {
                        deviceId: "default",
                    },
                    video: false
                });
                resolve(stream);
            }
            catch (e) {
                reject(e);
            }
        });
    }

    /**
     * 默认音频流切割器
     */
    public defaultSlice(stream: MediaStream, receive_handler: (data: Blob) => void, config: {
        mime: string,
        sliceTimeout: number
    }) {
        const mr = new MediaRecorder(stream, { mimeType: config.mime });
        mr.ondataavailable = async (e: BlobEvent) => {
            if (e.data && e.data.size > 0) {
                receive_handler(e.data);
            }
        };
        mr.start(config.sliceTimeout);
    }
}