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

Commented calling _EncodeVideoFrame() with NULL.

Browse Source 
As Adrian suggested, it's not only inefficient to do that there, but also wrong.
Modified the TODO comment
Added a new TODO comment
This commit is contained in:
JackBurton79 2018-08-09 10:44:22 +02:00
parent f049e59f50
commit 373832ae60
1 changed files with 8 additions and 6 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

14
src/add-ons/media/plugins/ffmpeg/AVCodecEncoder.cpp
View File

@ -649,6 +649,7 @@ AVCodecEncoder::_EncodeAudio(const uint8* buffer, size_t bufferSize,
if (fCodecContext->coded_frame) {
// Store information about the coded frame in the context.
fCodecContext->coded_frame->pts = packet.pts;
// TODO: double "!" operator ?
fCodecContext->coded_frame->key_frame = !!(packet.flags & AV_PKT_FLAG_KEY);
}
@ -709,12 +710,13 @@ AVCodecEncoder::_EncodeVideo(const void* buffer, int64 frameCount,
}
}
// Pass a NULL AVFrame and enter "draining" mode, then flush buffers
// before restarting encoding, again.
// TODO: It's probably not very efficient to do it like this. We should
// do this only when there is no more data (when closing the "stream")
_EncodeVideoFrame(NULL, pkt, info);
avcodec_flush_buffers(fCodecContext);
// TODO: we should pass a NULL AVFrame and enter "draining" mode, then flush buffers
// when we have finished and there is no more data. We cannot do that here, though, since
// 1. It's not efficient
// 2. It's incorrect, since many codecs need the "next" frame to be able to do optimization.
// if we drain the codec, they cannot work with the "next" frame.
//_EncodeVideoFrame(NULL, pkt, info);
//avcodec_flush_buffers(fCodecContext);
av_packet_free(&pkt);
return ret;
}