最近一直接调国标28181协议,有关音频的也调了很久,发现服务端下发的音频,在Android端用AudioTrack播放,过了好几秒后就渐渐地出现电流声,而且电流声也越来越强。
其实这个问题,之前我的文章也有阐述过,不过后来发现是治标不治本orz,传送门:
文章所说就是把minBufSize设为4倍,这样前十几秒声音还正常,但随着时间递增,还是会出现电流声。
解决办法也不是很难,就是每隔一段时间,清理下内存,并且释放不必要的未播放的数据audioTrack.flush(),此方法文档的解析如下:
//---------------------------------------------------------
// Audio data supply
//--------------------
/**
* Flushes the audio data currently queued for playback. Any data that has
* been written but not yet presented will be discarded. No-op if not stopped or paused,
* or if the track‘s creation mode is not {@link #MODE_STREAM}。
* 《BR》 Note that although data written but not yet presented is discarded, there is no
* guarantee that all of the buffer space formerly used by that data
* is available for a subsequent write.
* For example, a call to {@link #write(byte[], int, int)} with 《code》sizeInBytes《/code》
* less than or equal to the total buffer size
* may return a short actual transfer count.
*/
public void flush() {
if (mState == STATE_INITIALIZED) {
// flush the data in native layer
native_flush();
mAvSyncHeader = null;
mAvSyncBytesRemaining = 0;
}
}
最后贴上解决电流声的代码:
public void playSound(int encodeType, int channel, int samplerate, int bitrate, int length, byte[] data){
//Log.e(“TAG”,“AAAAA》》》》》》encodeType=” + encodeType +“nchannel=” + channel +“nsamplerate=” + samplerate +“nbitrate=” + bitrate + “nlength=” + length + “ndata=” + data);
if (audioTrack == null){
init();
frameCount = 0;
Log.e(“TAG”,“voice to voice init complete!”);
}
frameCount++;
if (frameCount % 5 == 0) {
Log.i(“writeAudio”, “gc+”);
audioTrack.flush();
System.gc();
Log.i(“writeAudio”, “gc-”);
}
if (length 》 0){
try{
audioTrack.write(data, 0, length);
}catch (Exception e){
e.printStackTrace();
}
}
}
顺便说下,服务端传过来的音频数据,是G77.a的音频,然后我再转为pcm,最后再放到AudioTrack进行播放。
最近一直接调国标28181协议,有关音频的也调了很久,发现服务端下发的音频,在Android端用AudioTrack播放,过了好几秒后就渐渐地出现电流声,而且电流声也越来越强。
其实这个问题,之前我的文章也有阐述过,不过后来发现是治标不治本orz,传送门:
文章所说就是把minBufSize设为4倍,这样前十几秒声音还正常,但随着时间递增,还是会出现电流声。
解决办法也不是很难,就是每隔一段时间,清理下内存,并且释放不必要的未播放的数据audioTrack.flush(),此方法文档的解析如下:
//---------------------------------------------------------
// Audio data supply
//--------------------
/**
* Flushes the audio data currently queued for playback. Any data that has
* been written but not yet presented will be discarded. No-op if not stopped or paused,
* or if the track‘s creation mode is not {@link #MODE_STREAM}。
* 《BR》 Note that although data written but not yet presented is discarded, there is no
* guarantee that all of the buffer space formerly used by that data
* is available for a subsequent write.
* For example, a call to {@link #write(byte[], int, int)} with 《code》sizeInBytes《/code》
* less than or equal to the total buffer size
* may return a short actual transfer count.
*/
public void flush() {
if (mState == STATE_INITIALIZED) {
// flush the data in native layer
native_flush();
mAvSyncHeader = null;
mAvSyncBytesRemaining = 0;
}
}
最后贴上解决电流声的代码:
public void playSound(int encodeType, int channel, int samplerate, int bitrate, int length, byte[] data){
//Log.e(“TAG”,“AAAAA》》》》》》encodeType=” + encodeType +“nchannel=” + channel +“nsamplerate=” + samplerate +“nbitrate=” + bitrate + “nlength=” + length + “ndata=” + data);
if (audioTrack == null){
init();
frameCount = 0;
Log.e(“TAG”,“voice to voice init complete!”);
}
frameCount++;
if (frameCount % 5 == 0) {
Log.i(“writeAudio”, “gc+”);
audioTrack.flush();
System.gc();
Log.i(“writeAudio”, “gc-”);
}
if (length 》 0){
try{
audioTrack.write(data, 0, length);
}catch (Exception e){
e.printStackTrace();
}
}
}
顺便说下,服务端传过来的音频数据,是G77.a的音频,然后我再转为pcm,最后再放到AudioTrack进行播放。
举报