open whole
Before my WeChat applet called Xunfei’s interface, I still sent a piece of audio to the background and then connected to Xunfei’s websocket. It was really slow. If it takes two or three seconds, it’s better to be a little longer. GG recently suddenly discovered that the WeChat applet has PCM format
. So I connected directly with the front websocket of the mini program.
The code is below.
I won’t talk about the application for the Xunfei account.
First, the WeChat mini program has to record and set variables.
const app = getApp()
const recorderManager = wx.getRecorderManager();
var wxst; //语音websocket
var status = 0; // 音频的状态
var iatResult = [] // 识别结果
const searchoptions = {
duration: 60000,//指定录音的时长，单位 ms
sampleRate: 8000,//采样率
numberOfChannels: 1,//录音通道数
encodeBitRate: 48000,//编码码率
format: 'PCM',//音频格式
frameSize: 5,//指定帧大小，单位 KB
}
Then there is a click event to start our recording and get Xunfei authentication. It is an interface because it was written in the background before and it is used. At the same time, it can also record the number of times used in the background. The java code
here
public static String getAuthUrl(String hostUrl, String apiKey, String apiSecret) throws Exception {
URL url = new URL(hostUrl);
SimpleDateFormat format = new SimpleDateFormat("EEE, dd MMM yyyy HH:mm:ss z", Locale.US);
format.setTimeZone(TimeZone.getTimeZone("GMT"));
String date = format.format(new Date());
StringBuilder builder = new StringBuilder("host: ").append(url.getHost()).append("\n").//
append("date: ").append(date).append("\n").//
append("GET ").append(url.getPath()).append(" HTTP/1.1");
//System.out.println(builder);
Charset charset = Charset.forName("UTF-8");
Mac mac = Mac.getInstance("hmacsha256");
SecretKeySpec spec = new SecretKeySpec(apiSecret.getBytes(charset), "hmacsha256");
mac.init(spec);
byte[] hexDigits = mac.doFinal(builder.toString().getBytes(charset));
String sha = Base64.getEncoder().encodeToString(hexDigits);
//System.out.println(sha);
String authorization = String.format("api_key=\"%s\", algorithm=\"%s\", headers=\"%s\", signature=\"%s\"", apiKey, "hmac-sha256", "host date request-line", sha);
//System.out.println(authorization);
HttpUrl httpUrl = HttpUrl.parse("https://" + url.getHost() + url.getPath()).newBuilder().//
addQueryParameter("authorization", Base64.getEncoder().encodeToString(authorization.getBytes(charset))).//
addQueryParameter("date", date).//
addQueryParameter("host", url.getHost()).//
build();
return httpUrl.toString();
}
Click event to start recording
start_say: function (e) {
//开始录音按钮
var that = this;
wx.getSetting({
//查看用户有没有开启语音权限
success(res) {
if (res.authSetting['scope.record']) {
wx.authorize({
scope: 'scope.record',
success() {
var xfurl = "";
wx.request({
//请求接口 获取讯飞语音鉴权
url: 接口地址,
method: "get",
header: {
'content-type': 'application/json' // 默认值
},
success: function (res) {
if (res.data.code == "200" && res.data.data) {
xfurl = res.data.data;
wxst = wx.connectSocket({
// 开启websocket连接
url: xfurl,
method: 'GET',
success: function (res) {
that.setData({
//我这里是个遮罩层 开启他
shows: true,
})
recorderManager.start(searchoptions);//开始录音
}
});
} else {
wx.showToast({
title: '获取语音鉴权失败',
icon: 'none',
mask: true,
duration: 3000
})
}
},
fail: function () {
wx.showToast({
title: '获取语音鉴权失败',
icon: 'none',
mask: true,
duration: 3000
})
}
})
},
fail() {
wx.showModal({
title: '微信授权',
content: '您当前未开启语音权限，请在右上角设置(···)中开启“录音功能”',
showCancel: false,
success(res) {
if (res.confirm) {
console.log('用户点击确定')
}
}
})
}
})
}else{
wx.showModal({
title: '微信授权',
content: '您当前未开启语音权限，请在右上角设置(···)中开启“录音功能”',
showCancel: false,
success(res) {
if (res.confirm) {
console.log('用户点击确定')
}
}
})
}
}
})
}
Click event to close recording
end_say: function () {
//结束录音按钮
var that = this;
recorderManager.stop();
that.setData({
//关闭遮罩层
shows: false,
})
}
After that, it is about recording and monitoring of websocket. I first open websocket and then record and
monitor the recording.
onShow: function() {
var that = this;
recorderManager.onStart(() => {
//开始录音时触发
status = 0;
iatResult = []
console.log('recorder start')
});
recorderManager.onError((res) => {
//错误回调
console.log(res);
});
recorderManager.onStop((res) => {
//结束录音时触发
console.log('recorder stop', res)
status = 2;
var sendsty = '{"data":{"status":2,"audio":"","format":"audio/L16;rate=8000","encoding":"raw"}}'
wxst.send({
data: sendsty
})
});
recorderManager.onFrameRecorded((res) => {
//每帧触发
const {
frameBuffer } = res
var int16Arr = new Int8Array(res.frameBuffer);
const base64 = wx.arrayBufferToBase64(int16Arr)
switch (status) {
case 0:
status = 1;
var sendsty = '{"common":{"app_id":"讯飞的appid"},"business":{"language":"zh_cn","domain":"iat","accent":"mandarin","dwa":"wpgs","vad_eos":1000},"data":{"status":0,"format":"audio/L16;rate=8000","encoding":"raw","audio":"' + base64 + '"}}'
wxst.send({
data: sendsty
})
break;
case 1:
var sendsty = '{"data":{"status":1,"format":"audio/L16;rate=8000","encoding":"raw","audio":"' + base64 + '"}}'
wxst.send({
data: sendsty
})
break;
default:
console.log("default");
}
})
}
About websocket monitoring
onLoad: function(options) {
var that = this;
wx.onSocketOpen((res) => {
// websocket打开
console.log('监听到 WebSocket 连接已打开' + res);
})
wx.onSocketError((err) => {
//连接失败
console.log('websocket连接失败', err);
wx.showToast({
title: 'websocket连接失败',
icon: 'none',
duration: 2000,
mask: false
})
})
wx.onSocketMessage((res) => {
//接收返回值
var data = JSON.parse(res.data)
if (data.code != 0) {
console.log("error code " + data.code + ", reason " + data.message)
return
}
let str = ""
if (data.data.status == 2) {
//最终识别结果
// data.data.status ==2 说明数据全部返回完毕，可以关闭连接，释放资源
wxst.close();
} else {
//中间识别结果
}
iatResult[data.data.result.sn] = data.data.result
if (data.data.result.pgs == 'rpl') {
data.data.result.rg.forEach(i => {
iatResult[i] = null
})
}
iatResult.forEach(i => {
if (i != null) {
i.ws.forEach(j => {
j.cw.forEach(k => {
str += k.w
})
})
}
})
that.setData({
searchKey: str //这个是中间的语音识别结果
})
})
wx.onSocketClose((res) => {
//WebSocket连接已关闭！
var that = this;
recorderManager.stop();
that.setData({
//把之前开开的遮罩层关上
shows: false,
})
var str = that.data.searchKey;
console.log(str);
str = str.replace(/\s*/g, "");//去除空格
if (str.substr(str.length - 1, 1) == "。") {
//去除句号
str = str.substr(0, str.length - 1);
}
that.setData({
searchKey: str//这个是最后确定的语音识别结果
})
console.log('WebSocket连接已关闭！')
})
}
ok. I tried it like this. Generally speaking, all the words will be recognized in about 1 second. It’s still pretty good. The user experience is also good. What’s wrong with
writing an article for the first time? Let’s communicate and welcome to leave a message. Aoli