详解Android studio实现语音转文字功能

详解Android studio实现语音转文字功能

当今，语音转文字已经成为很多App的基础功能之一。本篇攻略将为大家介绍如何使用Android Studio实现此功能。

Step 1.配置环境依赖

在做任何事之前，我们需要为我们的工程添加依赖关系。在 build.gradle 文件中加入以下代码，引入 com.google.cloud:google-cloud-speech:1.23.0 依赖关系。

implementation 'com.google.cloud:google-cloud-speech:1.23.0'

Step 2.添加权限

使用语音转文字功能需要使用录音和网络连接，因此我们需要在 AndroidManifest.xml 文件中添加以下权限：

<uses-permission android:name="android.permission.RECORD_AUDIO"/>
<uses-permission android:name="android.permission.INTERNET"/>

Step 3.使用Google Cloud Speech API

Google Cloud Speech-to-Text API是开发人员使用的云端API，可帮助将音频转换为文本，支持更多比其他语音转换接口更广泛的音频文件格式。我们需要在代码中使用 SpeechClient 类，将音频上传到Google Cloud服务器，并获得转换后的文本。

以下示例代码展示了如何使用 SpeechClient 类。

try (SpeechClient speechClient = SpeechClient.create()) {
    // 从文件中创建RecognitionAudio
    try (InputStream inputStream = Files.newInputStream(Paths.get(fileName))) {
        byte[] audioContents = ByteStreams.toByteArray(inputStream);
        RecognitionAudio audio = RecognitionAudio.newBuilder()
            .setContent(ByteString.copyFrom(audioContents))
            .build();
    }

    // 设置配置模块
    RecognitionConfig config = RecognitionConfig.newBuilder()
        .setEncoding(AudioEncoding.LINEAR16)
        .setSampleRateHertz(16000)
        .setLanguageCode("en-US")
        .build();

    // 识别语音
    RecognizeResponse response = speechClient.recognize(config, audio);
    for (SpeechRecognitionResult result : response.getResultsList()) {
        SpeechRecognitionAlternative alternative = result.getAlternativesList().get(0);
        System.out.printf("Transcription: %s%n", alternative.getTranscript());
    }
}

Step 4.实现录音

在Android上，使用 AudioRecord 类来进行录音。在我们的代码中，我们需要创建一个 RecordTask 类来在后台线程中执行录音操作。以下是完整示例代码：

public class RecordTask extends AsyncTask<Void, Void, Void> {
    private static final String TAG = RecordTask.class.getSimpleName();

    private AudioRecord mAudioRecord;
    private byte[] mBuffer;
    private boolean mIsRecording;
    private int mBufferSize;
    private String mFileName;

    public RecordTask(String fileName) {
        mFileName = fileName;
        mBufferSize = AudioRecord.getMinBufferSize(16000, AudioFormat.CHANNEL_IN_MONO, AudioFormat.ENCODING_PCM_16BIT);
        mBuffer = new byte[mBufferSize];
        mAudioRecord = new AudioRecord(MediaRecorder.AudioSource.MIC, 16000, AudioFormat.CHANNEL_IN_MONO,
                AudioFormat.ENCODING_PCM_16BIT, mBufferSize);
    }

    @Override
    protected Void doInBackground(Void... params) {
        try (OutputStream outputStream = new FileOutputStream(mFileName)) {
            mIsRecording = true;
            mAudioRecord.startRecording();

            while (mIsRecording) {
                int readSize = mAudioRecord.read(mBuffer, 0, mBufferSize);

                if (readSize > 0) {
                    outputStream.write(mBuffer, 0, readSize);
                }
            }

            mAudioRecord.stop();
        } catch (IOException e) {
            Log.e(TAG, "Error recording audio", e);
        }

        return null;
    }

    public void stopRecording() {
        mIsRecording = false;
    }
}

Step 5.结合Google Cloud Speech API和录音

最后，我们需要将上述代码整合起来。首先，创建一个启动录音的按钮。然后，当录音完成时，利用Google Cloud Speech API将录音内容转换为文本，并将文本显示在界面上。

以下是完整示例代码：

public class MainActivity extends AppCompatActivity {
    private static final String TAG = MainActivity.class.getSimpleName();

    private RecordTask mRecordTask;
    private TextView mTextView;

    @Override
    protected void onCreate(Bundle savedInstanceState) {
        super.onCreate(savedInstanceState);
        setContentView(R.layout.activity_main);

        mTextView = findViewById(R.id.textView);
        Button recordButton = findViewById(R.id.recordButton);

        recordButton.setOnClickListener(new View.OnClickListener() {
            @Override
            public void onClick(View v) {
                if (mRecordTask == null) {
                    mRecordTask = new RecordTask(getExternalFilesDir(null).getAbsolutePath() + "/audio.raw");
                    mRecordTask.execute();
                    ((Button) v).setText(R.string.stop_recording);
                } else {
                    mRecordTask.stopRecording();
                    mRecordTask = null;
                    ((Button) v).setText(R.string.start_recording);

                    transcribe();
                }
            }
        });
    }

    private void transcribe() {
        AsyncTask.execute(new Runnable() {
            @Override
            public void run() {
                try (SpeechClient speechClient = SpeechClient.create()) {
                    try (InputStream inputStream = new FileInputStream(getExternalFilesDir(null).getAbsolutePath() + "/audio.raw")) {
                        byte[] audioContents = ByteStreams.toByteArray(inputStream);
                        RecognitionAudio audio = RecognitionAudio.newBuilder()
                                .setContent(ByteString.copyFrom(audioContents))
                                .build();

                        RecognitionConfig config = RecognitionConfig.newBuilder()
                                .setEncoding(AudioEncoding.LINEAR16)
                                .setSampleRateHertz(16000)
                                .setLanguageCode("en-US")
                                .build();

                        RecognizeResponse response = speechClient.recognize(config, audio);

                        final StringBuilder transcriptionBuilder = new StringBuilder();
                        for (SpeechRecognitionResult result : response.getResultsList()) {
                            SpeechRecognitionAlternative alternative = result.getAlternativesList().get(0);
                            transcriptionBuilder.append(alternative.getTranscript());
                        }

                        runOnUiThread(new Runnable() {
                            @Override
                            public void run() {
                                mTextView.setText(transcriptionBuilder.toString());
                            }
                        });
                    }
                } catch (IOException e) {
                    Log.e(TAG, "Error transcribing audio", e);
                }
            }
        });
    }
}

以上示例代码已经可以实现语音转文字的功能，你可以按下启动录音按钮，开始录制音频。录音完成后，按下按钮就可以将录音转换成文本，并显示在界面上。