griefith/test
1
0
Fork 0
Code Issues Pull Requests Actions 1 Packages Projects Releases Wiki Activity

Fix #132214: wrong Render Audio result when using Matroska container format

Browse Source 
Matroska container really really wants to internally keep on using
milliseconds as a time base. Even if at initialization time both the
m_stream and m_codecCtx are set to use {1, sample_rate} timebase,
later during the avformat_write_header call the audio stream gets set
to {1, 1000}.

Doing av_packet_rescale_ts for each packet after receiving it but
before writing it seems to do all the correct work. This is what is
being done inside Blender movie_write_audio.cc, and in official ffmpeg
mux.c example.

Pull Request: https://projects.blender.org/blender/blender/pulls/132872
This commit is contained in:
Aras Pranckevicius
2025-01-10 12:25:01 +01:00
committed by Aras Pranckevicius
parent 6a1ac88b8e
commit 6a865a2265
1 changed files with 2 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
extern/audaspace/plugins/ffmpeg/FFMPEGWriter.cpp vendored
View File

@@ -120,7 +120,7 @@ void FFMPEGWriter::encode()
while(avcodec_receive_packet(m_codecCtx, m_packet) == 0)
{
m_packet->stream_index = m_stream->index;
av_packet_rescale_ts(m_packet, m_codecCtx->time_base, m_stream->time_base);
if(av_write_frame(m_formatCtx, m_packet) < 0)
AUD_THROW(FileException, "Frame couldn't be writen to the file with ffmpeg.");
}
@@ -161,7 +161,7 @@ void FFMPEGWriter::close()
while(avcodec_receive_packet(m_codecCtx, m_packet) == 0)
{
m_packet->stream_index = m_stream->index;
av_packet_rescale_ts(m_packet, m_codecCtx->time_base, m_stream->time_base);
if(av_write_frame(m_formatCtx, m_packet) < 0)
AUD_THROW(FileException, "Frame couldn't be writen to the file with ffmpeg.");
}