[mythtv] DRM VSync
doug at ties.org
Fri Apr 1 12:34:06 UTC 2005
Ivor Hewitt wrote:
> I have a drm vsync question....
> I'm currently getting myth working on a VIA CN400 unichrome-pro board...
> the current problem I've got is that the mythplayer is running away with
> itself and seems to have problems with the drm timing calculation.
> A log from a cn400 (@75Hz) playing back a 25fps PAL broadcast shows:-
> 2005-04-01 09:45:30.718 Video timing method: DRM
> 2005-04-01 09:45:30.718 Refresh rate: 40000, frame interval: 40000
> 'video_output' mean = '13661.89', std. dev. = '4427.53', fps = '73.20'
> 'video_output' mean = '13365.44', std. dev. = '578.30', fps = '74.82'
> The same video on a CLE266 (@60Hz) shows:-
> 2005-04-01 09:48:11.850 Video timing method: DRM
> 2005-04-01 09:48:11.850 Refresh rate: 16669, frame interval: 40000
> 'video_output' mean = '39239.88', std. dev. = '8932.58', fps = '25.48'
> 'video_output' mean = '40000.95', std. dev. = '8204.02', fps = '25.00'
> 'video_output' mean = '40002.02', std. dev. = '8206.91', fps = '25.00'
> i.e. on the cn400, the initial drm check seems to have decided that the
> refresh rate matches the video rate... but it's actually 75hz so it has
> run off and played at 75hz. With inevitable results.
> This is not a Myth problem and bound to be a driver problem, but I
> wondered if someone who knew how the DRM calculation code worked could
> suggest why it would have concluded that the refresh rate matched rather
> than just failing.
Looking at the code... refresh rate comes from the videoout method, not
the vsync method. Unichrome uses XvMC, right? I can't see anything
weird in Myth's computation. What does xvidtune say about your vertical
-------------- next part --------------
A non-text attachment was scrubbed...
Size: 256 bytes
Desc: OpenPGP digital signature
Url : http://mythtv.org/pipermail/mythtv-dev/attachments/20050401/ff181c56/signature.pgp
More information about the mythtv-dev