语音转文字

This commit is contained in:
18401019693
2022-10-24 18:18:39 +08:00
parent 6b7fd4d53e
commit 7fabc4f1b6
20 changed files with 532 additions and 403 deletions

View File

@@ -12,11 +12,12 @@ import androidx.annotation.NonNull;
import com.bumptech.glide.Glide;
import com.bumptech.glide.load.resource.bitmap.CircleCrop;
import com.bumptech.glide.request.RequestOptions;
import com.iflytek.cloud.SpeechConstant;
import com.iflytek.cloud.SpeechUtility;
import com.yunbao.common.CommonAppConfig;
import com.yunbao.common.bean.IMLoginModel;
import com.yunbao.common.event.RongIMConnectionStatusEvent;
import com.yunbao.common.manager.IMLoginManager;
import com.yunbao.common.views.weight.Recognizer;
import org.greenrobot.eventbus.EventBus;
@@ -61,7 +62,8 @@ public class RongcloudIMManager {
Log.e(CLASSNAME, "initRongIM:");
initPhotoGlide();
//初始化科大讯飞语音转文字
Recognizer.setAppId("671d2ae3");
String params = SpeechConstant.APPID + "=" + "671d2ae3";
SpeechUtility.createUtility(application.getApplicationContext(), params);
}
private static RongIMClient.OnReceiveMessageWrapperListener mListener;

View File

@@ -1,362 +0,0 @@
package com.yunbao.common.views.weight;
import android.content.Context;
import android.graphics.drawable.AnimationDrawable;
import android.os.Bundle;
import android.util.AttributeSet;
import android.util.Log;
import android.view.LayoutInflater;
import android.view.View;
import android.widget.ImageView;
import android.widget.RelativeLayout;
import android.widget.Toast;
import com.iflytek.cloud.ErrorCode;
import com.iflytek.cloud.InitListener;
import com.iflytek.cloud.RecognizerListener;
import com.iflytek.cloud.RecognizerResult;
import com.iflytek.cloud.SpeechConstant;
import com.iflytek.cloud.SpeechError;
import com.iflytek.cloud.SpeechRecognizer;
import com.iflytek.cloud.SpeechUtility;
import com.yunbao.common.R;
import org.json.JSONArray;
import org.json.JSONException;
import org.json.JSONObject;
import org.json.JSONTokener;
import java.util.Locale;
import java.util.Random;
public class Recognizer extends RelativeLayout {
private static final String TAG = "Recognizer";
private ImageView imgMic;
private Random random;
private IRecognizedResult mResultCallBack;
private SpeechRecognizer mIat = null;
private AnimationDrawable animStart;
private AnimationDrawable animEnd;
private static String mAppId;
/**
* 开发者可以通过此接口设置自己从科大讯飞官网申请的 appId。 此方法可以在 SDK init 之后调用。
*
* <p>注意: appid 必须和下载的SDK保持一致否则会出现10407错误
*
* @param appId 自定义的 appId
*/
public static void setAppId(String appId) {
mAppId = appId;
}
public Recognizer(Context context) {
super(context);
initViews();
}
public Recognizer(Context context, AttributeSet attrs) {
super(context, attrs);
initViews();
}
private void initViews() {
// setClickable(true);
// setBackgroundColor(getResources().getColor(R.color.white));
// RelativeLayout recognizerContainer =
// (RelativeLayout)
// LayoutInflater.from(getContext())
// .inflate(R.layout.rc_view_recognizer, null);
// View rlMic = recognizerContainer.findViewById(R.id.rl_mic);
// rlMic.setOnClickListener(
// new OnClickListener() {
// @Override
// public void onClick(View view) {
// if (mIat == null || !mIat.isListening()) {
// startRecognize();
// } else {
// reset();
// }
// }
// });
// imgMic = (ImageView) recognizerContainer.findViewById(R.id.img_mic);
//
//
// addView(recognizerContainer);
// random = new Random();
// String params = SpeechConstant.APPID + "=" + mAppId;
// SpeechUtility.createUtility(getContext().getApplicationContext(), params);
}
/**
* 初始化监听器。
*/
private static InitListener mInitListener =
new InitListener() {
@Override
public void onInit(int code) {
Log.e(TAG, "onInit " + code);
}
};
public void startRecognize() {
// if (null == mIat) {
// mIat = SpeechRecognizer.createRecognizer(getContext(), mInitListener);
// }
// if (mIat.isListening()) {
// return;
// }
// setParam();
// int ret = mIat.startListening(this);
// if (ret != ErrorCode.SUCCESS) {
// Log.e(TAG, "startRecognize ret error " + ret);
// }
}
/**
* 参数设置,设置听写参数详见《科大讯飞MSC API手册(Android)》SpeechConstant类
*
* @param
* @return
*/
private void setParam() {
// 清空参数
mIat.setParameter(SpeechConstant.PARAMS, null);
// 设置听写引擎
mIat.setParameter(SpeechConstant.ENGINE_TYPE, SpeechConstant.TYPE_CLOUD);
// 设置返回结果格式
mIat.setParameter(SpeechConstant.RESULT_TYPE, "json");
mIat.setParameter(SpeechConstant.DOMAIN, "iat");
if ("zh".equals(Locale.getDefault().getLanguage().toLowerCase())) {
mIat.setParameter(SpeechConstant.LANGUAGE, "zh_cn");
mIat.setParameter(SpeechConstant.ACCENT, "mandarin ");
} else {
mIat.setParameter(SpeechConstant.LANGUAGE, "en_us");
}
// 设置语音前端点:静音超时时间,即用户多长时间不说话则当做超时处理
mIat.setParameter(SpeechConstant.VAD_BOS, "4000");
// 设置语音后端点:后端点静音检测时间,即用户停止说话多长时间内即认为不再输入, 自动停止录音
mIat.setParameter(SpeechConstant.VAD_EOS, "1000");
// 设置标点符号,设置为"0"返回结果无标点,设置为"1"返回结果有标点
mIat.setParameter(SpeechConstant.ASR_PTT, "1");
}
private void setRandomImageResource() {
// int num = random.nextInt(3) + 1;
// switch (num) {
// case 1:
// imgMic.setImageResource(R.mipmap.rc_recognize_volume_01);
// break;
// case 2:
// imgMic.setImageResource(R.mipmap.rc_recognize_volume_02);
// break;
// default:
// imgMic.setImageResource(R.mipmap.rc_recognize_volume_03);
// break;
// }
}
private void changeVolume(int volume) {
// if (null != imgMic) {
// switch (volume / 2) {
// case 0:
// setRandomImageResource();
// break;
// case 1:
// imgMic.setImageResource(R.mipmap.rc_recognize_volume_02);
// break;
// case 2:
// imgMic.setImageResource(R.mipmap.rc_recognize_volume_03);
// break;
// case 3:
// imgMic.setImageResource(R.mipmap.rc_recognize_volume_04);
// break;
// case 4:
// imgMic.setImageResource(R.mipmap.rc_recognize_volume_05);
// break;
// case 5:
// imgMic.setImageResource(R.mipmap.rc_recognize_volume_06);
// break;
// case 6:
// imgMic.setImageResource(R.mipmap.rc_recognize_volume_07);
// break;
// case 7:
// imgMic.setImageResource(R.mipmap.rc_recognize_volume_08);
// break;
// case 8:
// imgMic.setImageResource(R.mipmap.rc_recognize_volume_09);
// break;
// case 9:
// imgMic.setImageResource(R.mipmap.rc_recognize_volume_10);
// break;
// case 10:
// imgMic.setImageResource(R.mipmap.rc_recognize_volume_11);
// break;
// case 11:
// imgMic.setImageResource(R.mipmap.rc_recognize_volume_12);
// break;
// case 12:
// imgMic.setImageResource(R.mipmap.rc_recognize_volume_13);
// break;
// default:
// imgMic.setImageResource(R.mipmap.rc_recognize_volume_14);
// break;
// }
// }
}
private void endOfSpeech() {
// if (null == imgMic) return;
// imgMic.setImageResource(R.drawable.rc_anim_speech_end);
// animEnd = (AnimationDrawable) imgMic.getDrawable();
// imgMic.clearAnimation();
// animEnd.start();
}
private void beginOfSpeech() {
// if (null == imgMic) return;
// imgMic.setImageResource(R.drawable.rc_anim_speech_start);
// animStart = (AnimationDrawable) imgMic.getDrawable();
// imgMic.clearAnimation();
// animStart.start();
}
@Override
protected void onDetachedFromWindow() {
super.onDetachedFromWindow();
if (null != mIat) {
mIat.cancel();
mIat.destroy();
mIat = null;
}
if (null != mResultCallBack) {
mResultCallBack = null;
}
if (animEnd != null) {
animEnd.stop();
animEnd = null;
}
if (animStart != null) {
animStart.stop();
animEnd = null;
}
mInitListener = null;
}
private void printResult(RecognizerResult result) {
String json = result.getResultString();
String text = parseRecognizeResult(result);
Log.e(TAG, "printResult " + text);
try {
JSONObject obj = new JSONObject(json);
boolean isLast = obj.getBoolean("ls");
if (isLast) {
endOfSpeech();
}
} catch (JSONException e) {
e.printStackTrace();
}
if (mResultCallBack != null) {
mResultCallBack.onResult(text);
}
}
//
// @Override
// public void onResult(RecognizerResult recognizerResult, boolean b) {
// printResult(recognizerResult);
// if (imgMic != null) {
// imgMic.setImageResource(R.mipmap.rc_recognize_disable);
// }
// }
//
// @Override
// public void onError(SpeechError speechError) {
// if (speechError.getErrorCode() == ErrorCode.ERROR_NO_NETWORK) {
// Toast.makeText(
// getContext(),
// getContext().getString(R.string.load_failure),
// Toast.LENGTH_SHORT)
// .show();
// }
// if (imgMic != null) {
// imgMic.setImageResource(R.mipmap.rc_recognize_disable);
// }
// }
//
// @Override
// public void onEvent(int eventType, int i1, int i2, Bundle bundle) {
// Log.e(TAG, "RecognizerView onEvent eventType: " + eventType);
// }
//
// @Override
// public void onVolumeChanged(int volume, byte[] bytes) {
// changeVolume(volume);
// }
//
// @Override
// public void onBeginOfSpeech() {
// Log.e(TAG, "RecognizerView onBeginOfSpeech");
// beginOfSpeech();
// }
//
// @Override
// public void onEndOfSpeech() {
// Log.e(TAG, "RecognizerView onEndOfSpeech");
// endOfSpeech();
// }
public void setCallBack(IRecognizedResult resultCallBack) {
mResultCallBack = resultCallBack;
}
private void reset() {
if (null != mIat) {
mIat.cancel();
mIat.destroy();
mIat = null;
}
if (null != animEnd) {
animEnd.stop();
animEnd = null;
}
if (animStart != null) {
animStart.stop();
animStart = null;
}
if (imgMic != null) {
// imgMic.setImageResource(R.mipmap.rc_recognize_disable);
}
}
public String parseRecognizeResult(RecognizerResult results) {
StringBuilder ret = new StringBuilder();
try {
JSONTokener jsonTokener = new JSONTokener(results.getResultString());
JSONObject jsonObject = new JSONObject(jsonTokener);
JSONArray words = jsonObject.getJSONArray("ws");
for (int i = 0; i < words.length(); i++) {
// 转写结果词,默认使用第一个结果
JSONArray items = words.getJSONObject(i).getJSONArray("cw");
JSONObject obj = items.getJSONObject(0);
ret.append(obj.getString("w"));
// 如果需要多候选结果,解析数组其他字段
// for(int j = 0; j < items.length(); j++)
// {
// JSONObject obj = items.getJSONObject(j);
// ret.append(obj.getString("w"));
// }
}
} catch (Exception e) {
e.printStackTrace();
}
return ret.toString();
}
}

Binary file not shown.

After

Width:  |  Height:  |  Size: 2.9 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 91 KiB