[FFmpeg-trac] #2062(avcodec:new): Playback of HD-PVR recordings corrupted when using VDPAU
FFmpeg
trac at avcodec.org
Tue Feb 12 10:21:54 CET 2013
#2062: Playback of HD-PVR recordings corrupted when using VDPAU
-------------------------------------+-----------------------------------
Reporter: jyavenard | Owner:
Type: defect | Status: new
Priority: normal | Component: avcodec
Version: unspecified | Resolution:
Keywords: vdpau h264 | Blocked By:
Blocking: | Reproduced by developer: 0
Analyzed by developer: 0 |
-------------------------------------+-----------------------------------
Comment (by Glidos):
I have a fix for this. Like jyavenard, I have swapped away from Michael
Niedermeyer's has_b_frames calculation to use the qatar version. I've then
corrected the qatar algorithm so that it handles BBC HD recordings
correctly. The two steps of that fix are on the fixes/0.26 branch of my
mythtv github fork: https://github.com/Glidos/mythtv/commits/fixes/0.26
To be honest, I think Michael's algorithm is the superior: it is simpler,
more logical and faster acting (if I understand it correctly), but I
believe it is less resiliant to sudden long backward steps in poc (albeit
that these happen only in samples that don't conform to the H264 spec). I
need to confirm this, but it true, I may be able to add the same
resiliance to Michael's algorithm.
The correction to the qatar algorithm is based on its monitoring poc
values coming OUT of the reordering process. Because of that, any required
delay detected needs to be added to the current delay, not max'd. Also, to
avoid multiple unnecessary delay increments, the recored of output pocs
needs resetting each time has_b_frames increases.
--
Ticket URL: <https://ffmpeg.org/trac/ffmpeg/ticket/2062#comment:5>
FFmpeg <http://ffmpeg.org>
FFmpeg issue tracker
More information about the FFmpeg-trac
mailing list