657 lines
20 KiB
JavaScript
657 lines
20 KiB
JavaScript
import {
|
||
ref,
|
||
onUnmounted
|
||
} from 'vue'
|
||
import config from '@/config'
|
||
|
||
export function useRealtimeRecorderOnce() {
|
||
// --- 状态定义 ---
|
||
const isRecording = ref(false)
|
||
const isProcessing = ref(false)
|
||
const recordingDuration = ref(0)
|
||
const volumeLevel = ref(0) // 0-100
|
||
const recognizedText = ref('')
|
||
const audioData = ref(null)
|
||
const audioDataForDisplay = ref([])
|
||
|
||
// --- 内部变量 ---
|
||
let durationTimer = null
|
||
|
||
// --- APP/小程序 变量 ---
|
||
let recorderManager = null;
|
||
let appAudioChunks = [];
|
||
|
||
// --- H5 变量 ---
|
||
let audioContext = null;
|
||
let mediaRecorder = null;
|
||
let h5Stream = null;
|
||
let h5AudioChunks = [];
|
||
let analyser = null;
|
||
let dataArray = null;
|
||
|
||
// --- 配置项 ---
|
||
const RECORD_CONFIG = {
|
||
duration: 600000,
|
||
sampleRate: 16000,
|
||
numberOfChannels: 1,
|
||
format: 'wav',
|
||
encodeBitRate: 16000,
|
||
frameSize: 4096
|
||
}
|
||
|
||
// --- WAV文件头函数 ---
|
||
const encodeWAV = (samples, sampleRate = 16000, numChannels = 1, bitsPerSample = 16) => {
|
||
const bytesPerSample = bitsPerSample / 8;
|
||
const blockAlign = numChannels * bytesPerSample;
|
||
const byteRate = sampleRate * blockAlign;
|
||
const dataSize = samples.length * bytesPerSample;
|
||
const buffer = new ArrayBuffer(44 + dataSize);
|
||
const view = new DataView(buffer);
|
||
|
||
// RIFF chunk descriptor
|
||
writeString(view, 0, 'RIFF');
|
||
view.setUint32(4, 36 + dataSize, true);
|
||
writeString(view, 8, 'WAVE');
|
||
|
||
// fmt sub-chunk
|
||
writeString(view, 12, 'fmt ');
|
||
view.setUint32(16, 16, true); // Subchunk1Size (16 for PCM)
|
||
view.setUint16(20, 1, true); // AudioFormat (1 for PCM)
|
||
view.setUint16(22, numChannels, true);
|
||
view.setUint32(24, sampleRate, true);
|
||
view.setUint32(28, byteRate, true);
|
||
view.setUint16(32, blockAlign, true);
|
||
view.setUint16(34, bitsPerSample, true);
|
||
|
||
// data sub-chunk
|
||
writeString(view, 36, 'data');
|
||
view.setUint32(40, dataSize, true);
|
||
|
||
// Write audio samples
|
||
const volume = 1;
|
||
let offset = 44;
|
||
for (let i = 0; i < samples.length; i++) {
|
||
let sample = Math.max(-1, Math.min(1, samples[i]));
|
||
sample = sample * volume;
|
||
view.setInt16(offset, sample < 0 ? sample * 0x8000 : sample * 0x7FFF, true);
|
||
offset += 2;
|
||
}
|
||
|
||
return buffer;
|
||
}
|
||
|
||
const writeString = (view, offset, string) => {
|
||
for (let i = 0; i < string.length; i++) {
|
||
view.setUint8(offset + i, string.charCodeAt(i));
|
||
}
|
||
}
|
||
|
||
const floatTo16BitPCM = (output, offset, input) => {
|
||
for (let i = 0; i < input.length; i++, offset += 2) {
|
||
const s = Math.max(-1, Math.min(1, input[i]));
|
||
output.setInt16(offset, s < 0 ? s * 0x8000 : s * 0x7FFF, true);
|
||
}
|
||
}
|
||
|
||
// --- 音量计算函数 ---
|
||
const calculateVolumeFromFloat32 = (float32Array) => {
|
||
if (!float32Array || float32Array.length === 0) return 0;
|
||
|
||
let sum = 0;
|
||
const length = float32Array.length;
|
||
|
||
// 计算RMS (均方根)
|
||
for (let i = 0; i < length; i++) {
|
||
// 绝对值方法更敏感
|
||
const absValue = Math.abs(float32Array[i]);
|
||
sum += absValue * absValue;
|
||
}
|
||
|
||
const rms = Math.sqrt(sum / length);
|
||
|
||
// 调试：打印原始RMS值
|
||
// console.log('Float32 RMS:', rms);
|
||
|
||
// 转换为0-100的值
|
||
// 使用对数刻度，使小音量变化更明显
|
||
let volume = 0;
|
||
|
||
if (rms > 0) {
|
||
// 使用对数转换：-60dB到0dB映射到0-100
|
||
const db = 20 * Math.log10(rms);
|
||
// 静音阈值约-60dB
|
||
if (db > -60) {
|
||
volume = Math.min(100, Math.max(0, (db + 60) / 0.6));
|
||
}
|
||
}
|
||
|
||
// 如果没有计算到值，使用旧方法作为fallback
|
||
if (volume === 0 && rms > 0) {
|
||
volume = Math.min(100, Math.floor(rms * 500));
|
||
}
|
||
|
||
// 确保最小值为0
|
||
return Math.max(0, volume);
|
||
}
|
||
|
||
const calculateVolumeFromInt16 = (int16Array) => {
|
||
if (!int16Array || int16Array.length === 0) return 0;
|
||
|
||
let sum = 0;
|
||
const length = int16Array.length;
|
||
|
||
// 计算RMS
|
||
for (let i = 0; i < length; i++) {
|
||
const normalized = int16Array[i] / 32768; // 归一化到[-1, 1]
|
||
const absValue = Math.abs(normalized);
|
||
sum += absValue * absValue;
|
||
}
|
||
|
||
const rms = Math.sqrt(sum / length);
|
||
|
||
// 调试：打印原始RMS值
|
||
// console.log('Int16 RMS:', rms);
|
||
|
||
// 转换为0-100的值
|
||
// 使用对数刻度
|
||
let volume = 0;
|
||
|
||
if (rms > 0) {
|
||
const db = 20 * Math.log10(rms);
|
||
if (db > -60) {
|
||
volume = Math.min(100, Math.max(0, (db + 60) / 0.6));
|
||
}
|
||
}
|
||
|
||
// 如果没有计算到值，使用旧方法作为fallback
|
||
if (volume === 0 && rms > 0) {
|
||
volume = Math.min(100, Math.floor(rms * 500));
|
||
}
|
||
|
||
// 确保最小值为0
|
||
return Math.max(0, volume);
|
||
}
|
||
|
||
/**
|
||
* 开始录音 (入口)
|
||
*/
|
||
const startRecording = async () => {
|
||
if (isRecording.value) return
|
||
|
||
try {
|
||
recognizedText.value = ''
|
||
volumeLevel.value = 0
|
||
audioData.value = null
|
||
audioDataForDisplay.value = []
|
||
appAudioChunks = []
|
||
h5AudioChunks = []
|
||
|
||
// #ifdef H5
|
||
if (location.protocol !== 'https:' && location.hostname !== 'localhost') {
|
||
uni.showToast({
|
||
title: 'H5录音需要HTTPS环境',
|
||
icon: 'none'
|
||
});
|
||
return;
|
||
}
|
||
// #endif
|
||
|
||
// #ifdef H5
|
||
await startH5Recording();
|
||
// #endif
|
||
|
||
// #ifndef H5
|
||
startAppRecording();
|
||
// #endif
|
||
|
||
isRecording.value = true;
|
||
recordingDuration.value = 0;
|
||
durationTimer = setInterval(() => recordingDuration.value++, 1000);
|
||
|
||
// 启动波形显示更新
|
||
updateAudioDataForDisplay();
|
||
|
||
} catch (err) {
|
||
console.error('启动失败:', err);
|
||
uni.showToast({
|
||
title: '启动失败: ' + (err.message || ''),
|
||
icon: 'none'
|
||
});
|
||
cleanup();
|
||
}
|
||
}
|
||
|
||
/**
|
||
* H5录音实现 - 手动构建WAV文件
|
||
*/
|
||
const startH5Recording = async () => {
|
||
try {
|
||
// 1. 获取麦克风流
|
||
const stream = await navigator.mediaDevices.getUserMedia({
|
||
audio: {
|
||
sampleRate: 16000,
|
||
channelCount: 1,
|
||
echoCancellation: true,
|
||
noiseSuppression: true,
|
||
autoGainControl: false
|
||
}
|
||
});
|
||
h5Stream = stream;
|
||
|
||
// 2. 创建 AudioContext 用于处理音频
|
||
const AudioContext = window.AudioContext || window.webkitAudioContext;
|
||
audioContext = new AudioContext({
|
||
sampleRate: 16000,
|
||
latencyHint: 'interactive'
|
||
});
|
||
|
||
// 创建音频处理节点
|
||
const source = audioContext.createMediaStreamSource(stream);
|
||
|
||
// 创建分析器用于音量计算
|
||
analyser = audioContext.createAnalyser();
|
||
analyser.fftSize = 256;
|
||
analyser.smoothingTimeConstant = 0.8;
|
||
dataArray = new Float32Array(analyser.frequencyBinCount);
|
||
|
||
source.connect(analyser);
|
||
|
||
// 创建脚本处理器用于收集音频数据
|
||
const processor = audioContext.createScriptProcessor(4096, 1, 1);
|
||
|
||
// 存储所有音频样本
|
||
let audioSamples = [];
|
||
|
||
processor.onaudioprocess = (e) => {
|
||
if (!isRecording.value) return;
|
||
|
||
// 获取输入数据
|
||
const inputData = e.inputBuffer.getChannelData(0);
|
||
|
||
// 计算音量
|
||
analyser.getFloatTimeDomainData(dataArray);
|
||
const volume = calculateVolumeFromFloat32(dataArray);
|
||
volumeLevel.value = volume;
|
||
|
||
// 收集音频样本
|
||
for (let i = 0; i < inputData.length; i++) {
|
||
audioSamples.push(inputData[i]);
|
||
}
|
||
|
||
// 存储当前音频数据块
|
||
const buffer = new Float32Array(inputData.length);
|
||
buffer.set(inputData);
|
||
h5AudioChunks.push(buffer);
|
||
};
|
||
|
||
source.connect(processor);
|
||
processor.connect(audioContext.destination);
|
||
|
||
// console.log('H5 16kHz WAV录音已启动');
|
||
|
||
} catch (err) {
|
||
console.error('H5录音启动失败:', err);
|
||
throw err;
|
||
}
|
||
}
|
||
|
||
/**
|
||
* 停止H5录音资源
|
||
*/
|
||
const stopH5Resources = () => {
|
||
// 断开所有连接
|
||
if (audioContext && audioContext.state !== 'closed') {
|
||
audioContext.close();
|
||
}
|
||
|
||
// 停止音轨
|
||
if (h5Stream) {
|
||
h5Stream.getTracks().forEach(track => track.stop());
|
||
}
|
||
|
||
audioContext = null;
|
||
analyser = null;
|
||
dataArray = null;
|
||
h5Stream = null;
|
||
}
|
||
|
||
/**
|
||
* APP/小程序录音实现
|
||
*/
|
||
const startAppRecording = () => {
|
||
recorderManager = uni.getRecorderManager();
|
||
|
||
recorderManager.onFrameRecorded((res) => {
|
||
const { frameBuffer } = res;
|
||
|
||
if (frameBuffer && frameBuffer.byteLength > 0) {
|
||
// 计算音量
|
||
const int16Data = new Int16Array(frameBuffer);
|
||
const volume = calculateVolumeFromInt16(int16Data);
|
||
volumeLevel.value = volume;
|
||
|
||
// 保存音频数据
|
||
appAudioChunks.push(frameBuffer);
|
||
}
|
||
});
|
||
|
||
recorderManager.onStart(() => {
|
||
// console.log('APP 16kHz WAV录音已开始');
|
||
});
|
||
|
||
recorderManager.onError((err) => {
|
||
console.error('APP录音报错:', err);
|
||
uni.showToast({
|
||
title: '录音失败: ' + err.errMsg,
|
||
icon: 'none'
|
||
});
|
||
cleanup();
|
||
});
|
||
|
||
recorderManager.start(RECORD_CONFIG);
|
||
}
|
||
|
||
/**
|
||
* 停止录音 (通用)
|
||
*/
|
||
const stopRecording = async () => {
|
||
if (!isRecording.value) return;
|
||
|
||
isRecording.value = false;
|
||
clearInterval(durationTimer);
|
||
|
||
// 停止硬件录音
|
||
stopHardwareResource();
|
||
|
||
// 处理录音数据
|
||
await processAudioData();
|
||
}
|
||
|
||
/**
|
||
* 取消录音
|
||
*/
|
||
const cancelRecording = () => {
|
||
if (!isRecording.value) return;
|
||
|
||
// console.log('取消录音 - 丢弃结果');
|
||
|
||
// 1. 停止硬件录音
|
||
stopHardwareResource();
|
||
|
||
// 2. 清理状态
|
||
recognizedText.value = '';
|
||
audioData.value = null;
|
||
audioDataForDisplay.value = [];
|
||
appAudioChunks = [];
|
||
h5AudioChunks = [];
|
||
|
||
// 3. 清理资源
|
||
cleanup();
|
||
}
|
||
|
||
/**
|
||
* 停止硬件资源
|
||
*/
|
||
const stopHardwareResource = () => {
|
||
// APP/小程序停止
|
||
if (recorderManager) {
|
||
recorderManager.stop();
|
||
}
|
||
|
||
// H5停止
|
||
// #ifdef H5
|
||
stopH5Resources();
|
||
// #endif
|
||
}
|
||
|
||
/**
|
||
* 更新音频数据显示
|
||
*/
|
||
const updateAudioDataForDisplay = () => {
|
||
const updateInterval = setInterval(() => {
|
||
if (!isRecording.value) {
|
||
clearInterval(updateInterval);
|
||
audioDataForDisplay.value = [];
|
||
return;
|
||
}
|
||
|
||
// 获取当前音量值
|
||
const currentVolume = volumeLevel.value;
|
||
|
||
// 调试：打印音量值
|
||
// console.log('Current Volume:', currentVolume);
|
||
|
||
// 生成适合 WaveDisplay 的数据
|
||
const data = [];
|
||
const center = 15; // 中心索引
|
||
const timeFactor = Date.now() / 150; // 更快的动画
|
||
|
||
// 根据音量动态调整波形强度
|
||
const volumeFactor = currentVolume / 100;
|
||
|
||
// 添加基础噪声，使波形在安静时也有轻微活动
|
||
const baseNoise = Math.random() * 0.1;
|
||
|
||
for (let i = 0; i < 31; i++) {
|
||
// 距离中心的位置
|
||
const distanceFromCenter = Math.abs(i - center) / center;
|
||
|
||
// 基础波形模式
|
||
const basePattern = 1 - Math.pow(distanceFromCenter, 1.2);
|
||
|
||
// 动态效果
|
||
const dynamicEffect = Math.sin(timeFactor + i * 0.3) * 0.3;
|
||
|
||
// 计算基础值
|
||
let value;
|
||
|
||
if (volumeFactor > 0.1) {
|
||
// 有音量时：音量因子占主导
|
||
value = volumeFactor * 0.8 * basePattern +
|
||
volumeFactor * 0.4 +
|
||
dynamicEffect * volumeFactor * 0.5;
|
||
} else {
|
||
// 安静时：使用动态效果和基础噪声
|
||
value = basePattern * 0.2 +
|
||
dynamicEffect * 0.1 +
|
||
baseNoise;
|
||
}
|
||
|
||
// 确保值在有效范围内
|
||
value = Math.max(0.15, Math.min(1, value));
|
||
|
||
// 随机微调
|
||
const randomVariance = volumeFactor > 0.1 ? 0.15 : 0.05;
|
||
value += (Math.random() - 0.5) * randomVariance;
|
||
value = Math.max(0.15, Math.min(1, value));
|
||
|
||
data.push(value);
|
||
}
|
||
|
||
audioDataForDisplay.value = data;
|
||
|
||
// 调试：检查生成的数据范围
|
||
// const min = Math.min(...data);
|
||
// const max = Math.max(...data);
|
||
// console.log(`Data range: ${min.toFixed(3)} - ${max.toFixed(3)}`);
|
||
|
||
}, 50);
|
||
}
|
||
|
||
/**
|
||
* 处理录音数据并生成WAV文件
|
||
*/
|
||
const processAudioData = async () => {
|
||
if (isProcessing.value) return;
|
||
|
||
isProcessing.value = true;
|
||
|
||
try {
|
||
let audioBlob = null;
|
||
|
||
// #ifdef H5
|
||
// H5端：合并所有音频样本并生成WAV
|
||
if (h5AudioChunks.length > 0) {
|
||
// 合并所有Float32Array
|
||
const totalLength = h5AudioChunks.reduce((sum, chunk) => sum + chunk.length, 0);
|
||
const mergedSamples = new Float32Array(totalLength);
|
||
|
||
let offset = 0;
|
||
h5AudioChunks.forEach(chunk => {
|
||
mergedSamples.set(chunk, offset);
|
||
offset += chunk.length;
|
||
});
|
||
|
||
// 生成WAV文件
|
||
const wavBuffer = encodeWAV(mergedSamples, 16000, 1, 16);
|
||
audioBlob = new Blob([wavBuffer], { type: 'audio/wav' });
|
||
|
||
// console.log(`H5生成WAV文件: ${audioBlob.size} bytes, 时长: ${mergedSamples.length / 16000}秒`);
|
||
}
|
||
// #endif
|
||
|
||
// #ifndef H5
|
||
// APP/小程序端：合并Int16数据并生成WAV
|
||
if (appAudioChunks.length > 0) {
|
||
// 合并所有Int16Array
|
||
const totalLength = appAudioChunks.reduce((sum, chunk) => sum + chunk.byteLength / 2, 0);
|
||
const mergedInt16 = new Int16Array(totalLength);
|
||
|
||
let offset = 0;
|
||
appAudioChunks.forEach(chunk => {
|
||
const int16Data = new Int16Array(chunk);
|
||
mergedInt16.set(int16Data, offset);
|
||
offset += int16Data.length;
|
||
});
|
||
|
||
// 转换为Float32用于生成WAV
|
||
const floatSamples = new Float32Array(mergedInt16.length);
|
||
for (let i = 0; i < mergedInt16.length; i++) {
|
||
floatSamples[i] = mergedInt16[i] / 32768;
|
||
}
|
||
|
||
// 生成WAV文件
|
||
const wavBuffer = encodeWAV(floatSamples, 16000, 1, 16);
|
||
audioBlob = new Blob([wavBuffer], { type: 'audio/wav' });
|
||
|
||
// console.log(`APP生成WAV文件: ${audioBlob.size} bytes, 时长: ${floatSamples.length / 16000}秒`);
|
||
}
|
||
// #endif
|
||
|
||
if (audioBlob && audioBlob.size > 44) { // 确保至少包含WAV头部
|
||
audioData.value = audioBlob;
|
||
|
||
// 保存文件用于调试（可选）
|
||
// debugSaveWavFile(audioBlob);
|
||
|
||
// 发送到服务器进行识别
|
||
await sendToASR(audioBlob);
|
||
} else {
|
||
throw new Error('录音数据为空或无效');
|
||
}
|
||
|
||
} catch (error) {
|
||
console.error('处理音频数据失败:', error);
|
||
uni.showToast({
|
||
title: '音频处理失败，请重试',
|
||
icon: 'none'
|
||
});
|
||
} finally {
|
||
isProcessing.value = false;
|
||
appAudioChunks = [];
|
||
h5AudioChunks = [];
|
||
}
|
||
}
|
||
|
||
/**
|
||
* 调试用：保存WAV文件
|
||
*/
|
||
const debugSaveWavFile = (blob) => {
|
||
const url = URL.createObjectURL(blob);
|
||
const a = document.createElement('a');
|
||
a.href = url;
|
||
a.download = `recording_${Date.now()}.wav`;
|
||
document.body.appendChild(a);
|
||
a.click();
|
||
document.body.removeChild(a);
|
||
URL.revokeObjectURL(url);
|
||
// console.log('WAV文件已保存用于调试');
|
||
}
|
||
|
||
/**
|
||
* 发送音频到ASR服务器
|
||
*/
|
||
const sendToASR = async (audioBlob) => {
|
||
try {
|
||
// 创建FormData
|
||
const formData = new FormData();
|
||
formData.append('file', audioBlob, 'recording.wav');
|
||
|
||
// 添加Token
|
||
const token = uni.getStorageSync('token') || '';
|
||
|
||
const asrUrl = `${config.baseUrl}/app/speech/asr`
|
||
|
||
const response = await fetch(asrUrl, {
|
||
method: 'POST',
|
||
headers: {
|
||
'Authorization': `Bearer ${token}`
|
||
},
|
||
body: formData
|
||
});
|
||
|
||
if (response.ok) {
|
||
const result = await response.json();
|
||
if(result.code == 200){
|
||
isProcessing.value = false
|
||
recognizedText.value = result.data || ''
|
||
}else{
|
||
$api.msg(result.msg || '识别失败')
|
||
}
|
||
|
||
} else {
|
||
const errorText = await response.text();
|
||
throw new Error(`ASR请求失败: ${response.status} - ${errorText}`);
|
||
}
|
||
|
||
} catch (error) {
|
||
console.error('ASR识别失败:', error);
|
||
}
|
||
}
|
||
|
||
/**
|
||
* 清理状态
|
||
*/
|
||
const cleanup = () => {
|
||
clearInterval(durationTimer);
|
||
isRecording.value = false;
|
||
isProcessing.value = false;
|
||
recordingDuration.value = 0;
|
||
volumeLevel.value = 0;
|
||
audioDataForDisplay.value = [];
|
||
|
||
if (recorderManager) {
|
||
recorderManager = null;
|
||
}
|
||
}
|
||
|
||
onUnmounted(() => {
|
||
if (isRecording.value) {
|
||
stopRecording();
|
||
}
|
||
cleanup();
|
||
})
|
||
|
||
return {
|
||
isRecording,
|
||
isProcessing,
|
||
recordingDuration,
|
||
volumeLevel,
|
||
recognizedText,
|
||
audioData,
|
||
audioDataForDisplay,
|
||
startRecording,
|
||
stopRecording,
|
||
cancelRecording
|
||
}
|
||
} |