After some research, I find out this is possible to capture the audio data in libaudioflinger of Android.
I think the audio data is being written to HAL in here:
ssize_t framesWritten = mNormalSink->write((char *)mSinkBuffer + offset, count);
So, I would like to save the mSinkBuffer + offset to a file (which I expected it will be raw PCM audio file). I using those stream to write it to file:
std::ofstream audioData ("/data/audiodata.raw", std::fstream::app);
audioData.write((char *)mSinkBuffer + offset, count);
audioData.close();
The file is successfully written and it has data in it. But, when I play the PCM file (audiodata.raw) with aplay or ffplay, the only sound I got is noise.
aplay -t raw -c 2 -f S16_LE -r 48000 audiodata.raw
I was worry about the config of aplay. So I print some log of the libaudioflinger:
10-07 10:14:54.575 1300 1366 I AudioFlinger: I/O handle: 13
10-07 10:14:54.575 1300 1366 I AudioFlinger: Standby: no
10-07 10:14:54.575 1300 1366 I AudioFlinger: Sample rate: 48000 Hz
10-07 10:14:54.575 1300 1366 I AudioFlinger: HAL frame count: 512
10-07 10:14:54.575 1300 1366 I AudioFlinger: HAL format: 0x1 (AUDIO_FORMAT_PCM_16_BIT)
10-07 10:14:54.575 1300 1366 I AudioFlinger: HAL buffer size: 2048 bytes
10-07 10:14:54.575 1300 1366 I AudioFlinger: Channel count: 2
10-07 10:14:54.575 1300 1366 I AudioFlinger: Channel mask: 0x00000003 (front-left, front-right)
10-07 10:14:54.575 1300 1366 I AudioFlinger: Processing format: 0x5 (AUDIO_FORMAT_PCM_FLOAT)
10-07 10:14:54.576 1300 1366 I AudioFlinger: Processing frame size: 8 bytes
10-07 10:14:54.576 1300 1366 I AudioFlinger: Pending config events:
10-07 10:14:54.576 1300 1366 I AudioFlinger: none
10-07 10:14:54.576 1300 1366 I AudioFlinger: Output device: 0x2 (AUDIO_DEVICE_OUT_SPEAKER)
10-07 10:14:54.576 1300 1366 I AudioFlinger: Input device: 0 (AUDIO_DEVICE_NONE)
10-07 10:14:54.576 1300 1366 I AudioFlinger: Audio source: 0 (default)
I don't know what I did wrong. Please help me!
Thank you in advanced!
Open The file in append |binary mode
std::ofstream audioData ("/data/audiodata.raw", std::fstream::app | std::fstream::binary);
binary - binary - Operations are performed in binary mode rather than text.
Raw PCM buffer should be written in binary mode.
Please check the code for the difference between the bytes and count ( added some comments for the reference , I hope this will solve your problem)
if (mNormalSink != 0) {
/* Count is the number of Frames or sample written != bytes */
const size_t count = mBytesRemaining / mFrameSize;
ATRACE_BEGIN("write");
// update the setpoint when AudioFlinger::mScreenState changes
uint32_t screenState = AudioFlinger::mScreenState;
if (screenState != mScreenState) {
mScreenState = screenState;
MonoPipe *pipe = (MonoPipe *)mPipeSink.get();
if (pipe != NULL) {
pipe->setAvgFrames((mScreenState & 1) ?
(pipe->maxFrames() * 7) / 8 : mNormalFrameCount * 2);
}
}
ssize_t framesWritten = mNormalSink->write((char *)mSinkBuffer + offset, count);
ATRACE_END();
if (framesWritten > 0) {
bytesWritten = framesWritten * mFrameSize;
// std::ofstream audioData ("/data/audiodata.raw", std::fstream::binary);
/* File write or stream write is the number of bytes written to the file */
audioData.write((char *)mSinkBuffer + offset, bytesWritten);
// audioData.close();
} else {
bytesWritten = framesWritten;
}
// otherwise use the HAL / AudioStreamOut directly
}
Audacity - Open a raw file
File -> Import -> raw data
Select the raw file path
Based on your raw file attached.
Use these Setting
Encoding : 32 -bit float
Byte order: Little Endian
Channels : 2 channels ( stereo)
Start offset : 0
Amount to import : 100
Sample Rate : 48000
File plays properly without any glitches/noise.