Aren/haiku
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

AVCodecDecoder: fix timing problems

Browse Source 
A combination of two problems made things go wrong with the timestamp of
decoded audio.

1) The output buffer size is too small to hold the complete input.
swresample handles this by buffering the input for use the next time it
is called, however repeatedly doing this results in lots of buffering,
and our way to compute the output timestamp from the input does not take
it into account so it does weird things. Moreover, we would need to
empty the buffer by calling swr_convert with NULL input in that case.

Fix: make sure to not feed more data to swr_convert than it can output
in our buffer. This way, no buffering occurs, only the matrixing
conversion.

2) When using planar audio, the "frame size" is a bit different. Instead
of adding sample size * channel count to 1 pointer, we need to add
sample size * 1 to each channel buffer.

Fix: add the "fInputFrameSize" which takes this into account, instead of
misusing fOutputFrameSize for the input.

Fixes #12460.
This commit is contained in:
Adrien Destugues 2016-02-14 11:45:13 +01:00
parent 45569721aa
commit da455572ed
2 changed files with 21 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

21
src/add-ons/media/plugins/ffmpeg/AVCodecDecoder.cpp
View File

@ -112,6 +112,7 @@ AVCodecDecoder::AVCodecDecoder()
fOutputFrameCount(0),
fOutputFrameRate(1.0),
fOutputFrameSize(0),
fInputFrameSize(0),
fChunkBuffer(NULL),
fChunkBufferSize(0),
@ -403,6 +404,11 @@ AVCodecDecoder::_NegotiateAudioOutputFormat(media_format* inOutFormat)
fOutputFrameSize = sampleSize * outputAudioFormat.channel_count;
fOutputFrameCount = outputAudioFormat.buffer_size / fOutputFrameSize;
fOutputFrameRate = outputAudioFormat.frame_rate;
if (av_sample_fmt_is_planar(fContext->sample_fmt))
fInputFrameSize = sampleSize;
else
fInputFrameSize = fOutputFrameSize;
fRawDecodedAudio->opaque
= av_realloc(fRawDecodedAudio->opaque, sizeof(avformat_codec_context));
if (fRawDecodedAudio->opaque == NULL)
@ -925,10 +931,19 @@ AVCodecDecoder::_MoveAudioFramesToRawDecodedAudioAndUpdateStartTimes()
ptr[i] = fDecodedDataBuffer->data[i] + fDecodedDataBufferOffset;
}
int32 result = swr_convert(fResampleContext, fRawDecodedAudio->data,
// When there are more input frames than space in the output buffer,
// we could feed everything to swr and it would buffer the extra data.
// However, there is no easy way to flush that data without feeding more
// input, and it makes our timestamp computations fail.
// So, we feed only as much frames as we can get out, and handle the
// buffering ourselves.
// TODO Ideally, we should try to size our output buffer so that it can
// always hold all the output (swr provides helper functions for this)
inFrames = frames;
frames = swr_convert(fResampleContext, fRawDecodedAudio->data,
outFrames, ptr, inFrames);
if (result < 0)
if (frames < 0)
debugger("resampling failed");
} else {
memcpy(fRawDecodedAudio->data[0], fDecodedDataBuffer->data[0]
@ -937,7 +952,7 @@ AVCodecDecoder::_MoveAudioFramesToRawDecodedAudioAndUpdateStartTimes()
inFrames = frames;
}
size_t remainingSize = inFrames * fOutputFrameSize;
size_t remainingSize = inFrames * fInputFrameSize;
size_t decodedSize = outFrames * fOutputFrameSize;
fDecodedDataBufferSize -= inFrames;

3
src/add-ons/media/plugins/ffmpeg/AVCodecDecoder.h
View File

@ -131,6 +131,9 @@ private:
float fOutputFrameRate;
int fOutputFrameSize;
// sample size * channel count
int fInputFrameSize;
// sample size * channel count
// or just sample size for planar formats
uint8_t* fChunkBuffer;
size_t fChunkBufferSize;