[MPlayer-users] recent regression in initial h264 decode/display

Reimar Döffinger Reimar.Doeffinger at gmx.de
Wed Aug 21 20:12:55 CEST 2013


On Sun, Aug 18, 2013 at 09:02:34PM +0200, Reimar Döffinger wrote:
> On 15.08.2013, at 23:46, Ilja Sekler <ilja_sekler_ at gmx.de> wrote:
> 
> > Am 15.08.2013 23:43, schrieb Ilja Sekler:
> >> An alternative sample could be "Doodling in Math Class: Snakes + Graphs"
> >> <http://www.youtube.com/watch?v=heKK95DAKms> by Victoria Hart, please
> >> specify format 18 (the size of the resulting mp4 file should be 12541447
> >> Bytes). Verbose MPlayer console output for this file is attached.
> > 
> > Forgot to attach the log. My apologies.
> 
> Ok, what seems to happen is that multi-threaded decode causes the aspect to be signalled delayed.
> On this aspect being set, MPlayer resets VDPAU and that causes the issue.
> No idea how to handle this, I think FFmpeg just should avoid using threading in this case, but it's a bit hard to actually implement.

I wouldn't call it a fix, but I committed an improvement that fixes some
other potential issues and mostly hides the effects of this one.


More information about the MPlayer-users mailing list