使用 android 自带回声消除

Android 在4.1 API leve 16 添加了回声消除处理 AcousticEchoCanceler

不过由于手机厂商太多，碎片化严重，实现效果也不一样，这里只做测试性使用

记录一下步骤：

检测系统是否支持回声消除

创建录制，并拿到sessionid

创建回声消除 关联录制的sessionid

创建播放，关联录制的sessionid

这三者由录制的sessionid 关联起来了，就能够实处理掉mic采集到的由喇叭播放出的声音了。

看代码：

    public static boolean chkNewDev()
    {
          return android.os.Build.VERSION.SDK_INT >= 16;
    }
    
	public static boolean isDeviceSupport()
	{
        return AcousticEchoCanceler.isAvailable();
	}

检测是否支持，居然存在不准确的情况？

int InitAudioRecord()
		{
			m_bufferSizeInBytes = AudioRecord.getMinBufferSize(m_sampleRateInHz,  m_channelConfig, m_audioFormat);  	
			if (chkNewDev())
			{
				m_audioRecord = new AudioRecord(MediaRecorder.AudioSource.VOICE_COMMUNICATION, m_sampleRateInHz, m_channelConfig, m_audioFormat, m_bufferSizeInBytes);
			}
			else
			{
				m_audioRecord = new AudioRecord(MediaRecorder.AudioSource.MIC, m_sampleRateInHz, m_channelConfig, m_audioFormat, m_bufferSizeInBytes);
			}
			
			int packSize = Math.min(960, m_bufferSizeInBytes*2);
			audioData = new short[packSize];
			return 0;
		}

如果系统支持回声消除，则用MediaRecorder.AudioSource.VOICE_COMMUNICATION

public int GetSessionId()
		{
			return m_audioRecord.getAudioSessionId();
		}

获取mic的sessionid，给回声消除和播放关联用

public boolean initAEC(int audioSession)
	{
		if (m_canceler != null)
		{
			return false;
		}
		m_canceler = AcousticEchoCanceler.create(audioSession);
		m_canceler.setEnabled(true);
		return m_canceler.getEnabled();
	}

创建回声消除 并使能

		int InitAudioTrack()
		{
			int m_sampleRateInHz = 48000;  
			int m_channelConfig  =  AudioFormat.CHANNEL_IN_STEREO;   
			int m_audioFormat    = AudioFormat.ENCODING_PCM_16BIT; 
			int m_bufferSizeInBytes = 1024*4;
			
			if (chkNewDev() && m_recorder != null)
			{
				m_audioTrack = new AudioTrack(AudioManager.STREAM_VOICE_CALL, m_sampleRateInHz, m_channelConfig, m_audioFormat, m_bufferSizeInBytes, AudioTrack.MODE_STREAM, m_recorder.GetSessionId());
			}
			else
			{
				m_audioTrack = new AudioTrack(AudioManager.STREAM_VOICE_CALL, m_sampleRateInHz, m_channelConfig, m_audioFormat, m_bufferSizeInBytes, AudioTrack.MODE_STREAM);
			}
			return 0;
		}

如果系统支持回声处理 则创建关联mic 的sessionid

ok，走通了

        getWindow().addFlags(WindowManager.LayoutParams.FLAG_KEEP_SCREEN_ON);
        AudioManager audioManager = (AudioManager) this.getSystemService(Context.AUDIO_SERVICE);
        audioManager.setMode(AudioManager.MODE_IN_COMMUNICATION);
        audioManager.setSpeakerphoneOn(true);

忘了说，在程序启动的时候需要设置一下 AudioManager.MODE_IN_COMMUNICATION 和 外放来验证回声处理效果