离线语音识别android采取不需要的声音

Posted 2023-04-19

【中文标题】离线语音识别android采取不需要的声音

【英文标题】：Offline voice recognition android taking unwanted voice

【发布时间】：2015-09-04 04:26:32

【问题描述】：

我做了很多研究并尝试了离线袖珍狮身人面像，但它正在接收周围的声音，并且对我的应用做出不同的反应。是否有任何适用于 kitkat 4.4 的 Google 离线应用程序。我从 2 周开始尝试它。感谢您的宝贵回答。

详细信息：当活动开始时，第一个文本必须在完成后阅读内容（文本到语音），阅读语音识别必须采取语音并按照命令例如：（下一个，上一个，前锋、选项、1、2、3、4 等）。根据命令，它必须在 onresult 方法中识别并对其做出反应。

错误：我听到一些声音后出现错误

06-18 19:54:00.159: V/onBeginningOfSpeech(3360): onBeginningOfSpeech
06-18 19:54:01.024: V/onPartialResult(3360): option
06-18 19:54:01.109: I/cmusphinx(3360): INFO: fsg_search.c(843): 105 frames, 5333 HMMs (50/fr), 7748 senones (73/fr), 371 history entries (3/fr)
06-18 19:54:01.110: I/SpeechRecognizer(3360): Stop recognition
06-18 19:54:01.110: E/cmusphinx(3360): ERROR: "fsg_search.c", line 913: Final result does not match the grammar in frame 105
06-18 19:54:01.111: V/onPartialResult-->(3360): option
06-18 19:54:01.111: V/onResult(3360): onResult

android Hive offline link

但如果我关闭互联网，它就无法离线工作。

在pocketsphinx 中，它对正确的单词没有反应。如果我说“下一个”，它会靠近另一个人的语音内容附加“密钥”，这给我带来了很多问题。是否有任何解决方案或库可以离线使用。任何离线谷歌离线语音支持。

下面是我尝试过的代码

package com.example.sample1;

import static edu.cmu.pocketsphinx.SpeechRecognizerSetup.defaultSetup;
import java.io.File;
import java.io.IOException;
import java.util.ArrayList;
import java.util.HashMap;
import java.util.Locale;
import edu.cmu.pocketsphinx.RecognitionListener;
import android.app.Activity;
import android.graphics.Color;
import android.os.AsyncTask;
import android.os.Bundle;
import android.os.Handler;
import android.speech.tts.TextToSpeech;
import android.speech.tts.TextToSpeech.OnUtteranceCompletedListener;
import android.util.Log;
import android.view.View;
import android.view.View.OnClickListener;
import android.widget.Button;
import android.widget.ExpandableListView;
import android.widget.ExpandableListView.OnChildClickListener;
import android.widget.ExpandableListView.OnGroupClickListener;
import android.widget.ExpandableListView.OnGroupExpandListener;
import android.widget.TextView;
import android.widget.Toast;
import edu.cmu.pocketsphinx.Assets;
import edu.cmu.pocketsphinx.Hypothesis;
import edu.cmu.pocketsphinx.SpeechRecognizer;

public class Sam extends Activity implements RecognitionListener, TextToSpeech.OnInitListener 

    /* Named searches allow to quickly reconfigure the decoder */

    private static final String DIGITS_SEARCH = "digits";

    private SpeechRecognizer recognizer;
    private HashMap<String, Integer> captions;
    private TextView caption_text;
    private TextView result_text;
    ArrayList<String> result1;
    private Button buttonLeft;
    private Button buttonRight;
    int count = 0;
    private ArrayList<DataAnswer> dummyListTemp;
    private ArrayList<DataAnswer> dummyList;
    AnswerDataAdapter listAdapter = null;
    int conteo = 0;
    Handler a = new Handler();
    private TextToSpeech tts; 
    String readIt ="";

    HashMap<String, String> params = new HashMap<String, String>();

    @Override
    public void onCreate(Bundle state) 
        super.onCreate(state);
        Log.v("onCreate", "onCreate");

        // Prepare the data for UI
        captions = new HashMap<String, Integer>();

        captions.put(DIGITS_SEARCH, R.string.digits_caption);

        setContentView(R.layout.quiz);
        caption_text = (TextView) findViewById(R.id.caption_text);
        result_text = (TextView) findViewById(R.id.result_text);
    //  listViewAnswer = (ExpandableListView) findViewById(R.id.listViewAnswer);
        buttonRight  = (Button) findViewById(R.id.buttonRight);
        buttonLeft  = (Button) findViewById(R.id.buttonLeft);
        result_text.setText("Result --->: ");

        tts = new TextToSpeech(this, this);  

        params.put(TextToSpeech.Engine.KEY_PARAM_UTTERANCE_ID,"stringId");

        String text = "World is full of chanllenge"; 
        //Speakes the text first and then after comple reading text voice recoginzation must start
        speakOut(text);




        buttonRight.setOnClickListener(new OnClickListener()           
            @Override
            public void onClick(View v) 
                try 
                    //onClickRight();

                    Toast.makeText(getApplicationContext(), "Right", Toast.LENGTH_SHORT).show();
                 catch (Exception e)
                    e.printStackTrace();
                


            
        );

        buttonLeft.setOnClickListener(new OnClickListener()            
            @Override
            public void onClick(View v) 
                //onClickLeft();
                Toast.makeText(getApplicationContext(), "Left", Toast.LENGTH_SHORT).show();
            
        );


        // Recognizer initialization is a time-consuming and it involves IO,
        // so we execute it in async task
        //if(!tts.isSpeaking()) 
        new AsyncTask<Void, Void, Exception>() 
            @Override
            protected Exception doInBackground(Void... params) 
                try 
                    Assets assets = new Assets(Sam.this);
                    File assetDir = assets.syncAssets();
                    Log.v("AsyncTask", "AsyncTask");
                    setupRecognizer(assetDir);

                 catch (IOException e) 
                    return e;
                
                return null;
            

            @Override
            protected void onPostExecute(Exception result) 
                Log.v("onPostExecute", "onPostExecute");
                try
                    if (result != null) 
                        //caption_text.setText("Failed to init recognizer " + result);
                        Toast.makeText(getApplicationContext(), "Failed to init recognizer ", Toast.LENGTH_SHORT).show();
                     else 
                        FireRecognition();
                        //switchSearch(DIGITS_SEARCH);

                    
                 catch (Exception e) 
                    e.printStackTrace();
                
            
        .execute();
    


    @Override
    public void onDestroy() 
        super.onDestroy();
        Log.v("onDestroy", "onDestroy");
        recognizer.cancel();
        recognizer.shutdown();

        if (tts != null) 
            tts.stop();
            tts.shutdown();
        
    

    public void FireRecognition()
        Log.d("Recognition","Recognition Started");
        //caption_text.setText("Recognition Started!");
        //Toast.makeText(getApplicationContext(), "Recognition Started!", Toast.LENGTH_SHORT).show();
        recognizer.stop();
        //recognizer.startListening("digits");

    

    /**
     * In partial result we get quick updates about current hypothesis. In
     * keyword spotting mode we can react here, in other modes we need to wait
     * for final result in onResult.
     */
    @Override
    public void onPartialResult(Hypothesis hypothesis) 
        try 
            //Log.v("onPartialResult", "onPartialResult");
            if (hypothesis == null)
                return;
            Log.v("onPartialResult", hypothesis.getHypstr().toString());

            String text = hypothesis.getHypstr();
            if(recognizer !=null)
                recognizer.stop();

            caption_text.setText("Partial result -->: " + text);

            Log.v("onPartialResult-->", text);

            //   Toast.makeText(getApplicationContext(), text, Toast.LENGTH_SHORT).show();
         catch (Exception e) 
            e.printStackTrace();
        
    

    /**
     * This callback is called when we stop the recognizer.
     */
    @Override
    public void onResult(Hypothesis hypothesis) 
        try
            Log.v("onResult", "onResult");
            //  result_text.setText("");
            if (hypothesis != null) 
                String text = hypothesis.getHypstr();
                //Toast.makeText(getApplicationContext(), text, Toast.LENGTH_SHORT).show();
                //  ((TextView) findViewById(R.id.result_text)).setText(text);

                if(text.toLowerCase().equals("next")) 
                    result_text.setText("Result --->: " + text);
                    Toast.makeText(getApplicationContext(), text, Toast.LENGTH_SHORT).show();
                 else if(text.toLowerCase().equals("previous")) 
                    result_text.setText("Result --->: " + text);
                    Toast.makeText(getApplicationContext(), text, Toast.LENGTH_SHORT).show();

                 else if(text.toLowerCase().trim().equals("option one".toLowerCase().trim())) 
                    result_text.setText("Result --->: " + text);


                    Toast.makeText(getApplicationContext(), text, Toast.LENGTH_SHORT).show();
                    result_text.setText("Result --->: " + text);
                 else if(text.toLowerCase().trim().equals("option two".toLowerCase().toString())) 

                    Toast.makeText(getApplicationContext(), text, Toast.LENGTH_SHORT).show();
                 else if(text.toLowerCase().trim().equals("option three".toLowerCase().toString())) 
                    result_text.setText("Result --->: " + text);
                    Toast.makeText(getApplicationContext(), text, Toast.LENGTH_SHORT).show();
                 else if(text.toLowerCase().trim().equals("option four".toLowerCase().toString())) 
                    result_text.setText("Result --->: " + text);
                    Toast.makeText(getApplicationContext(), text, Toast.LENGTH_SHORT).show();
                 else 
                    Toast.makeText(getApplicationContext(), " No Access:--" + text, Toast.LENGTH_SHORT).show();
                




                Log.v("onResult-->", text);
                if(recognizer != null)              
                    recognizer.startListening("digits");
            
         catch (Exception e) 
            e.printStackTrace();
        

    

    @Override
    public void onBeginningOfSpeech() 
        Log.v("onBeginningOfSpeech", "onBeginningOfSpeech");

    

    /**
     * We stop recognizer here to get a final result
     */
    @Override
    public void onEndOfSpeech() 
        Log.v("onEndOfSpeech", "onEndOfSpeech");

        if (!recognizer.getSearchName().equals(DIGITS_SEARCH))
            switchSearch(DIGITS_SEARCH);
    


    private void switchSearch(String searchName) 
        Log.v("switchSearch", "switchSearch--->" + searchName);
        recognizer.stop();

        // If we are not spotting, start listening with timeout (10000 ms or 10 seconds).
        if (searchName.equals(DIGITS_SEARCH))
            recognizer.startListening(searchName, 10000);
        /* else
            recognizer.startListening(searchName, 10000);*/

        /* String caption = getResources().getString(captions.get(searchName));
        caption_text.setText(caption);*/
    

    private void setupRecognizer(File assetsDir) throws IOException 

        // The recognizer can be configured to perform multiple searches
        // of different kind and switch between them
        Log.v("setupRecognizer", "setupRecognizer");
        recognizer = defaultSetup()
                .setAcousticModel(new File(assetsDir, "en-us-ptm"))
                .setDictionary(new File(assetsDir, "cmudict-en-us.dict"))

                // To disable logging of raw audio comment out this call (takes a lot of space on the device)
                .setRawLogDir(assetsDir)

                // Threshold to tune for keyphrase to balance between false alarms and misses
                .setKeywordThreshold(1e-20f) //1e-20f 1e-45f

                // Use context-independent phonetic search, context-dependent is too slow for mobile
                // .setBoolean("-allphone_ci", true)

                .getRecognizer();
        recognizer.addListener(this);

        /** In your application you might not need to add all those searches.
         * They are added here for demonstration. You can leave just one.
         */

        // Create keyword-activation search.
        // recognizer.addKeyphraseSearch(KWS_SEARCH, KEYPHRASE);



        // Create grammar-based search for digit recognition
        File digitsGrammar = new File(assetsDir, "digits.gram");
        recognizer.addGrammarSearch(DIGITS_SEARCH, digitsGrammar);

    



    @Override
    public void onError(Exception error) 
        Log.v("onError", "onError");
        //caption_text.setText(error.getMessage());
        Toast.makeText(getApplicationContext(), error.getMessage(), Toast.LENGTH_SHORT).show();
    

    @Override
    public void onTimeout() 
        Log.v("onTimeout", "onTimeout");
        switchSearch(DIGITS_SEARCH);
    


    @SuppressWarnings("deprecation")
    @Override
    public void onInit(int status) 

        tts.setOnUtteranceCompletedListener(new OnUtteranceCompletedListener() 

            @Override
            public void onUtteranceCompleted(String utteranceId) 
                runOnUiThread(new Runnable() 
                    @Override
                    public void run() 
                        if(recognizer != null)
                            recognizer.startListening("digits");
                        //Toast.makeText(getApplicationContext(), "Completed", Toast.LENGTH_LONG).show();
                    
                );
            
        );

        if (status == TextToSpeech.SUCCESS)   

            int result = tts.setLanguage(Locale.US);  


            Log.i("Success", "Completed");

            if (result == TextToSpeech.LANG_MISSING_DATA  || result == TextToSpeech.LANG_NOT_SUPPORTED)   
                Log.e("TTS", "This Language is not supported");  
             else   
                //buttonSpeak.setEnabled(true);  
                //  speakOut();  
                String text = " No Voice Found".toString();
                Log.i("else", "else");
                // speakOut(text);  
              

         else   
            Log.e("TTS", "Initilization Failed!");  
          

    


    @SuppressWarnings("deprecation")
    private void speakOut(String text) 
        if(tts.isSpeaking()) 
            //recognizer.stop();
        
        //  String text =  ((TextView) findViewById(R.id.caption_text)).getText().toString();
        tts.speak(text, TextToSpeech.QUEUE_FLUSH, params);

        Log.i("Speaking-->", "****" + tts.isSpeaking());

     

在 digit.gram 中

#JSGF V1.0;

grammar digits;

<digit> = option one |
          option two |
          option three |
          option four |
          back |
          previous |
          next;

public <digits> = <digit>+;

以上是我的代码，如果有人处理过，请告诉我解决方案 这是我的 xml quiz.xml

<?xml version="1.0" encoding="utf-8"?>
<RelativeLayout xmlns:android="http://schemas.android.com/apk/res/android"
    android:layout_
    android:layout_
    android:orientation="vertical" >

    <LinearLayout
        android:id="@+id/linearLayout2"
        android:layout_
        android:layout_
        android:layout_alignParentLeft="true"
        android:layout_below="@+id/linearLayout1"
        android:orientation="vertical" >

        <ExpandableListView
            android:id="@+id/listViewAnswer"
            android:layout_
            android:layout_
            android:focusable="false"
            android:focusableInTouchMode="false"
            android:childDivider="#334455" >
        </ExpandableListView>
    </LinearLayout>

    <RelativeLayout
        android:id="@+id/relativeLayout1"
        android:layout_
        android:layout_
        android:layout_alignParentLeft="true"
        android:layout_below="@+id/linearLayout2"
        android:layout_marginTop="10dip"
        android:orientation="horizontal" >

        <Button
            android:id="@+id/buttonRight"
            android:layout_
            android:layout_
            android:layout_alignParentRight="true"
            android:layout_marginRight="10dip"
            android:text="Next" />

        <Button
            android:id="@+id/buttonLeft"
            android:layout_
            android:layout_
            android:layout_marginLeft="10dip"
            android:text="Previous" />
    </RelativeLayout>

    <RelativeLayout
        android:id="@+id/relativeLayout2"
        android:layout_
        android:layout_
        android:layout_alignParentLeft="true"
        android:layout_below="@+id/relativeLayout1" >

        <TextView
            android:id="@+id/result_text"
            android:layout_
            android:layout_
            android:layout_alignParentLeft="true"
            android:textAppearance="?android:attr/textAppearanceMedium"
            android:textColor="#556677" />

        <TextView
            android:id="@+id/caption_text"
            android:layout_
            android:layout_
            android:layout_below="@+id/result_text"
            android:layout_alignParentLeft="true"
            android:textColor="#443399" />
    </RelativeLayout>

</RelativeLayout>

更新：

如果你正在使用 addKeywordSearch

File digitsGrammar = new File(context.getFilesDir(), "digits.gram");
recognizer.addKeywordSearch(DIGITS_SEARCH, digitsGrammar);

然后在 digit.gram 中只写

 option one /1e-1/
 option two /1e-1/
 option three /1e-1/
 option four /1e-1/
 back /1e-1/
 previous /1e-1/
 next /1e-1/

或

如果您使用的是 addGrammarSearch

 File digitsGrammar = new File(context.getFilesDir(), "digits.gram");
 recognizer.addGrammarSearch(DIGITS_SEARCH, digitsGrammar);

然后在 digit.gram 中

#JSGF V1.0;

grammar digits;

<digit> = option one |
      option two |
      option three  |
      option four |
      back  |
      previous |
      next;

     public <digits> = <digit>+;

【参考方案1】：

你需要使用关键词定位模式而不是语法模式来持续聆听

您可以在此处找到示例：

Recognizing multiple keywords using PocketSphinx

【讨论】：

怎么做。我可以得到任何相关的样本吗？

而不是识别器.addGrammarSearch(DIGITS_SEARCH, digitsGrammar);我必须添加识别器.addKeywordSearch(DIGITS_SEARCH, digitsGrammar);仪式。然后我需要在 digit.gram 中编辑什么

更新了上面的 digit.gram 是我必须为关键字搜索更改的写入方式

您可以在这里找到示例：***.com/questions/25748113/…

我已经尝试过该示例，但它在 digit.gram 和 jsgf 错误中显示错误。这是在 digit.gram #JSGF V1.0 中编辑的正确方法；语法数字； = 选项一 /1.0/ |选项二 /1.0/ |选项三 /1.0/ |选项四 /1.0/ |返回 /1.0/ |以前的 /1.0/ |下一个;公共 = +;

【参考方案2】：

从 Android M 开始，您可以在创建识别器意图时使用 EXTRA_PREFER_OFFLINE。