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

Hack for interlaced video. We cannot know if video is interlaced from the

Browse Source 
current libavformat API. FFmpeg is a little inconsistent in this regard. For
interlaced video, it will report a frame rate which is the field rate, but
still decode both fields into a single frame. For the time being, we reduce
the frame rate from 50.0 to 25.0 and handle interlaced video transparently
for the clients. This needs to be fixed later on...


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@31995 a95241bf-73f2-0310-859d-f6bbb57e9c96
This commit is contained in:
Stephan Aßmus 2009-07-30 20:47:05 +00:00
parent 0876f8d08e
commit eeb4be0c7f
1 changed files with 12 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

15
src/add-ons/media/plugins/ffmpeg/AVFormatReader.cpp
View File

@ -502,10 +502,13 @@ AVFormatReader::StreamCookie::Init(int32 virtualIndex)
// format->u.encoded_video.forward_history = 0;
// format->u.encoded_video.backward_history = 0;
// TODO: Fix up for interlaced video
format->u.encoded_video.output.field_rate
= av_q2d(stream->r_frame_rate);
if (format->u.encoded_video.output.field_rate == 50.0f)
format->u.encoded_video.output.field_rate = 25.0f;
format->u.encoded_video.output.interlace = 1;
// TODO: Fix up for interlaced video
format->u.encoded_video.output.first_active = 0;
format->u.encoded_video.output.last_active
= codecContext->height - 1;
@ -724,7 +727,7 @@ AVFormatReader::StreamCookie::Seek(uint32 flags, int64* frame,
*frame, *time);
if ((flags & B_MEDIA_SEEK_TO_FRAME) != 0)
*time = *frame * 1000000LL / FrameRate();
*time = (bigtime_t)(*frame * 1000000LL / FrameRate());
double timeBase = av_q2d(fStream->time_base);
int64_t timeStamp;
@ -768,7 +771,7 @@ AVFormatReader::StreamCookie::FindKeyFrame(uint32 flags, int64* frame,
double frameRate = FrameRate();
if ((flags & B_MEDIA_SEEK_TO_FRAME) != 0)
*time = *frame * 1000000LL / frameRate;
*time = (bigtime_t)(*frame * 1000000LL / frameRate);
double timeBase = av_q2d(fStream->time_base);
int64_t timeStamp;
@ -823,6 +826,12 @@ AVFormatReader::StreamCookie::GetNextChunk(const void** chunkBuffer,
return ret;
}
// NOTE: AVPacket has a field called "convergence_duration", for which
// the documentation is quite interesting. It sounds like it could be
// used to know the time until the next I-Frame in streams that don't
// let you know the position of keyframes in another way (like through
// the index).
// According to libavformat documentation, fPacket is valid until the
// next call to av_read_frame(). This is what we want and we can share
// the memory with the least overhead.