[MPlayer-dev-eng] New patches coming

[Dan Oscarsson]

On 2010-05-20 at 09:23 +0200 Dan Oscarsson wrote:

> > >  It's especially useful if your 
> > > method could be extended painlessly for every video output.
> > 
> > I doubt there's anything that would be both painless (in the sense of
> > needing nothing VO-specific) and accurate, though some common handling
> > could be used for VOs that block until a frame can be switched (as
> > opposed to say some kind of buffering behavior).
> 
> My patches only have a few requirements on the VO. Basically queue frame
> for display on next vsync without blocking and supply information about
> time between vsyncs and when they occur. Using opengl and X I can get
> the vsync info, so I think it could be made to work fairly well with xv,
> x11, opengl, and a few more as long as opengl is available.

I have made a quick test and added code to get vsync info through opengl
calls, to the x11 code.
Testing with -vo xv, -vo gl to see what happens when the a movie is
played in same speed as refresh rate or faster.
In standard mplayer or in Uoti's git-version audio-video will get out of
sync or stall/jumpy frames.
With my quick test code it works much better.
So it looks like my frame scheduling code works quite well for other vos
than vdpau too.

  Dan