使用pocketsphinx识别器进行语音反馈
仍然使用相同的项目，这是Run pocketSphinx和Google TTS的延续。 我已经根据Nikolay Shymyrev的指南做了修改并做了很多帮助。 但我想要实现的最终function仍然存在。 谷歌TTS现在运行得很好，但识别器有一些问题。
如果谷歌TTS有些话很长，那么识别器就无法启动
Speaker.speak("Drive mode now will be enabled. I will read your new messages for you now."); 然后我的onPartialResult如果条件不能满足就好
if (text.equals("exit")) { speaker.speak("yeah get in"); recognizer.cancel(); Intent i = new Intent(getApplicationContext(),PocketSphinxActivity.class); startActivity(i);
我认为识别器总是听，因为它在后台运行，然后它听取谷歌TTS句子，导致它后来不会识别我的演讲。 因为当我使用免提麦克风时， speaker.Speak的句子只是"Drive mode enabled" ，它会很好地识别我的下一个字，并在我说"exit"时执行上面的if条件。 但是，如果句子很长，就像"Drive mode now will be enabled, I will read bla bla bla"它不会听我的"exit"字。
我现在要做的是为识别器添加超时以超时几次，以便它不会识别任何不必要的声音。 我想放
startRecognition("search", timeout)
但我的Eclipse不会让我这样做。 它给了我错误。 我正在使用PocketSphinx for Android 5 pre alpha。
再次，我的代码构建只是为了测试并确保它只识别“退出”单词
SMSReaderMain.java
public class SMSReaderMain extends Activity implements RecognitionListener { private final int CHECK_CODE = 0x1; private final int LONG_DURATION = 5000; private final int SHORT_DURATION = 1200; private Speaker speaker; private TextView smsText; private TextView smsSender; private BroadcastReceiver smsReceiver; public static final String TURNON_SR = "drive mode"; public static final String TURNOFF_SR = "ok"; public static final String DESTROY_SR = "exit"; public SpeechRecognizer recognizer; public HashMap captions; @Override public void onCreate(Bundle savedInstanceState) { super.onCreate(savedInstanceState); setContentView(R.layout.main_sms); captions = new HashMap(); new AsyncTask() { @Override protected Exception doInBackground(Void... params) { try { Assets assets = new Assets(SMSReaderMain.this); File assetDir = assets.syncAssets(); setupRecognizer(assetDir); } catch (IOException e) { return e; } return null; } @Override protected void onPostExecute(Exception result) { if (result != null) { ((TextView) findViewById(R.id.caption_text)) .setText("Failed to init recognizer " + result); } else { switchSearch(TURNOFF_SR); } } }.execute(); //toggle = (ToggleButton)findViewById(R.id.speechToggle); smsText = (TextView)findViewById(R.id.sms_text); smsSender = (TextView)findViewById(R.id.sms_sender); checkTTS(); } private void startDriveMode(){ speaker.allow(true); //speaker.speak(getString(R.string.start_speaking)); speaker.speak("Drive mode enabled"); } private void checkTTS(){ Intent check = new Intent(); check.setAction(TextToSpeech.Engine.ACTION_CHECK_TTS_DATA); startActivityForResult(check, CHECK_CODE); } @Override protected void onActivityResult(int requestCode, int resultCode, Intent data) { if(requestCode == CHECK_CODE){ if(resultCode == TextToSpeech.Engine.CHECK_VOICE_DATA_PASS) { speaker = new Speaker(this); }else { Intent install = new Intent(); install.setAction(TextToSpeech.Engine.ACTION_INSTALL_TTS_DATA); startActivity(install); } } startDriveMode(); initializeSMSReceiver(); registerSMSReceiver(); } private void initializeSMSReceiver(){ smsReceiver = new BroadcastReceiver(){ @Override public void onReceive(Context context, Intent intent) { Bundle bundle = intent.getExtras(); if(bundle!=null){ Object[] pdus = (Object[])bundle.get("pdus"); for(int i=0;i<pdus.length;i++){ byte[] pdu = (byte[])pdus[i]; SmsMessage message = SmsMessage.createFromPdu(pdu); String text = message.getDisplayMessageBody(); String sender = getContactName(message.getOriginatingAddress()); speaker.pause(LONG_DURATION); speaker.speak("You have a new message from" + sender + "!"); speaker.pause(SHORT_DURATION); speaker.speak(text); smsSender.setText("Message from " + sender); smsText.setText(text); } } } }; } private void registerSMSReceiver() { IntentFilter intentFilter = new IntentFilter("android.provider.Telephony.SMS_RECEIVED"); registerReceiver(smsReceiver, intentFilter); } private String getContactName(String phone){ Uri uri = Uri.withAppendedPath(PhoneLookup.CONTENT_FILTER_URI, Uri.encode(phone)); String projection[] = new String[]{ContactsContract.Data.DISPLAY_NAME}; Cursor cursor = getContentResolver().query(uri, projection, null, null, null); if(cursor.moveToFirst()){ return cursor.getString(0); }else { return "unknown number"; } } @Override protected void onDestroy() { super.onDestroy(); unregisterReceiver(smsReceiver); speaker.destroy(); } public void onPartialResult(Hypothesis hypothesis) { //System.out.println("masuk coiii"); String text = hypothesis.getHypstr(); try { if (text.equals("exit")) { speaker.speak("yeah get in"); recognizer.cancel(); Intent i = new Intent(getApplicationContext(),PocketSphinxActivity.class); startActivity(i); } //Intent i= null; /**if (text.equals(TURNON_SR)) { recognizer.cancel(); popPicture(); startDriveMode(); } if (text.equals(TURNOFF_SR)) { //speaker = new Speaker(this); speaker.speak(getString(R.string.stop_speaking)); speaker.allow(false); //popPicture2(); } if (text.equals(DESTROY_SR)) { recognizer.cancel(); i = new Intent(getApplicationContext(),PocketSphinxActivity.class); startActivity(i); onDestroy(); //popPicture2(); } **/ } catch (Exception e) { e.printStackTrace(); } } public void popPicture() { LayoutInflater inflater = getLayoutInflater(); View layout = inflater.inflate(R.layout.toast_image,(ViewGroup) findViewById(R.id.toast_layout_id)); Toast toast = new Toast(getApplicationContext()); toast.setGravity(Gravity.CENTER_HORIZONTAL, 0, 0); toast.setDuration(Toast.LENGTH_SHORT); toast.setView(layout); toast.show(); } public void onResult(Hypothesis hypothesis) { ((TextView) findViewById(R.id.result_text)).setText(""); if (hypothesis != null) { String text = hypothesis.getHypstr(); makeText(getApplicationContext(), text, Toast.LENGTH_SHORT).show(); } } public void switchSearch(String searchName) { recognizer.stop(); recognizer.startListening(searchName); //taro timout disini biar mic ga denger suara hp sendiri ((TextView) findViewById(R.id.caption_text)).setText(searchName); } public void setupRecognizer(File assetsDir) { File modelsDir = new File(assetsDir, "models"); recognizer = defaultSetup() .setAcousticModel(new File(modelsDir, "hmm/en-us-semi")) .setDictionary(new File(modelsDir, "dict/cmu07a.dic")) .setRawLogDir(assetsDir).setKeywordThreshold(1e-10f) .getRecognizer(); recognizer.addListener(this); // Create grammar-based searches. // recognizer.addKeyphraseSearch(TURNOFF_SR, TURNON_SR); //recognizer.addGrammarSearch(TURNON_SR, new File(modelsDir, "grammar/sms.gram")); //recognizer.addGrammarSearch(TURNOFF_SR, new File(modelsDir, "grammar/sms.gram")); //recognizer.addGrammarSearch(DESTROY_SR, new File(modelsDir, "grammar/sms.gram")); File menuGrammar = new File(modelsDir, "grammar/sms.gram"); recognizer.addGrammarSearch(TURNOFF_SR, menuGrammar); //recognizer.addGrammarSearch(TURNON_SR, menuGrammar); //recognizer.addGrammarSearch(DESTROY_SR, menuGrammar); } @Override public void onBeginningOfSpeech() { // TODO Auto-generated method stub } @Override public void onEndOfSpeech() { // TODO Auto-generated method stub } }
Speaker.java
public class Speaker implements OnInitListener { private TextToSpeech tts; private boolean ready = false; private boolean prematureSpeak = false; private String ps; private boolean allowed = false; public Speaker(Context context){ tts = new TextToSpeech(context, this); } public boolean isAllowed(){ return allowed; } //public void allow(boolean allowed){ public void allow(boolean allowed){ this.allowed = allowed; } @Override public void onInit(int status) { if(status == TextToSpeech.SUCCESS){ // Change this to match your // locale tts.setLanguage(Locale.US); ready = true; if (prematureSpeak) { speak(ps); prematureSpeak = false; } }else{ ready = false; } } public void speak(String text){ // Speak only if the TTS is ready // and the user has allowed speech if(ready && allowed) { HashMap hash = new HashMap(); hash.put(TextToSpeech.Engine.KEY_PARAM_STREAM, String.valueOf(AudioManager.STREAM_NOTIFICATION)); tts.speak(text, TextToSpeech.QUEUE_ADD, hash); } else if(!ready) { prematureSpeak = true; ps = text; } } public void pause(int duration){ tts.playSilence(duration, TextToSpeech.QUEUE_ADD, null); } // Free up resources public void destroy(){ tts.shutdown(); } public boolean isSpeaking() { return tts.isSpeaking(); } }
您的代码有几个问题：
1）我告诉你使用关键字定位模式，你仍然使用语法模式
2）您需要在开始语音反馈之前取消识别器，而不是先说话然后取消
if (text.equals("exit")) { speaker.speak("yeah get in"); recognizer.cancel(); ....
你需要先取消然后说：
if (text.equals("exit")) { recognizer.cancel(); speaker.speak("yeah get in"); ....
3）一旦扬声器结束，您需要重新启动识别器，但不需要再次运行活动，请参阅详细信息如何知道TTS何时完成？
通过onUtteranceEnded中的这些更改，您再次启动识别器：
public void onUtteranceCompleted(String utteranceId) { recognizer.startSearch("search name"); }
不要在onPartialResult中重启识别器，等待TTS完成。