Qrobot開發總結之Android語音識別

Android sdk提供了語音識別的接口，有些人在網上找了例子發現運行不了（PS：網上的例子基本就那一個，都是各種轉載的），原因在於手機沒有安裝google語音搜索軟件！去網上下載一個安上就可以了，另外需保持手機網絡暢通。

第一種方法：這種方法會顯示一個語音對話框，各種提示信息會顯示的比較清晰，也是實現起來最簡單的。

觸發語音識別是調用

1. Intent intent = new Intent(RecognizerIntent.ACTION_RECOGNIZE_SPEECH);
2. intent.putExtra(RecognizerIntent.EXTRA_LANGUAGE_MODEL, RecognizerIntent.LANGUAGE_MODEL_FREE_FORM);
3. intent.putExtra(RecognizerIntent.EXTRA_PROMPT, "您的語音將轉化為文字");
4. startActivityForResult(intent, VOICE_RECOGNITION_REQUEST_CODE);

VOICE_RECOGNITION_REQUEST_CODE這個是一個靜態的全局變量

sdk解釋是：requestCode If>= 0, this code will be returned in onActivityResult() when the activity exits.

然後復寫onActivityResult()方法：

1. @Override
2. protected void onActivityResult(int requestCode, int resultCode, Intent data) {
3. if(requestCode == VOICE_RECOGNITION_REQUEST_CODE && resultCode == RESULT_OK){
4. ArrayList<String> results = data.getStringArrayListExtra(RecognizerIntent.EXTRA_RESULTS);
5. Toast.makeText(MainActivity.this, matches.get(0), Toast.LENGTH_LONG).show();
6. //to do sth
7. }
9. super.onActivityResult(requestCode, resultCode, data);
10. }

第2中方法：這種方法是不顯示對話框，完全在後台運行，給用戶的體驗會更好一些。

初始化：

1. SpeechRecognizer recognizer = SpeechRecognizer.createSpeechRecognizer(this);
2. recognizer.setRecognitionListener(new listener());

觸發語音識別時調用

1. Intent intent = new Intent(RecognizerIntent.ACTION_RECOGNIZE_SPEECH);
2. intent.putExtra("calling_package", "VoiceIME");
3. intent.putExtra(RecognizerIntent.EXTRA_MAX_RESULTS, 1);
4. recognizer.startListening(intent);

區別於第一種方法，這種方法要實現一個接口RecognitionListener類

1. class listener implements RecognitionListener {
2. public void onReadyForSpeech(Bundle params) {
4. }
6. public void onBeginningOfSpeech() {
8. }
10. public void onRmsChanged(float rmsdB) {
12. }
14. public void onBufferReceived(byte[] buffer) {
16. }
18. public void onEndOfSpeech() {
20. }
22. public void onError(int error) {
23. String s = "";
24. switch(error){
25. case SpeechRecognizer.ERROR_AUDIO:
26. s = "錄音設別錯誤";
27. break;
28. case SpeechRecognizer.ERROR_CLIENT:
29. s = "其他客戶端錯誤";
30. break;
31. case SpeechRecognizer.ERROR_INSUFFICIENT_PERMISSIONS:
32. s = "權限不足";
33. break;
34. case SpeechRecognizer.ERROR_NETWORK:
35. s = "網絡連接錯誤";
36. break;
37. case SpeechRecognizer.ERROR_NETWORK_TIMEOUT:
38. s = "網絡連接超時";
39. break;
40. case SpeechRecognizer.ERROR_NO_MATCH:
41. s = "沒有匹配項";
42. break;
43. case SpeechRecognizer.ERROR_RECOGNIZER_BUSY:
44. s = "識別服務繁忙";
45. break;
46. case SpeechRecognizer.ERROR_SERVER:
47. s = "識別服務器錯誤";
48. break;
49. case SpeechRecognizer.ERROR_SPEECH_TIMEOUT:
50. s = "無語音輸入";
51. break;
52. }
53. s += " 請重試";
54. Toast.makeText(MainActivity.this, s, Toast.LENGTH_SHORT).show();
56. }
58. public void onResults(Bundle results) {
60. recognizer_result = results.getStringArrayList(SpeechRecognizer.RESULTS_RECOGNITION).get(0);
61. Toast.makeText(MainActivity.this, recognizer_result, Toast.LENGTH_SHORT).show();
63. }
65. public void onPartialResults(Bundle partialResults) {
67. }
69. public void onEvent(int eventType, Bundle params) {
71. }
73. }

看方法的名字相信大家應該就知道要干什麼了，多看文檔。起初我覺得用第2中方法給用戶的提示信息太少，往往出錯了，用戶不知道是怎麼回事。後面看文檔才發現google android sdk已經定義了很多錯誤的錯誤號，只要根據不同錯誤給出恰當的提示就好了！