[mythtv] Video Output patch v13

[Robert Johnston]

On 4/20/05, Daniel Kristjansson <danielk at cat.nyu.edu> wrote:
> On Tue, 2005-04-19 at 22:46 -0400, Daniel Kristjansson wrote:
> This is an update on the Xv/XvMC patch.
> The patch is here:
>    http://www.mrl.nyu.edu/~danielk/mythtv/xv-xvmc-merge-v13.patch.bz2

<snip />

> On my early model Geforce FX card I can pop up a non-transparent OSD
> without problems, while the non-transparent OSD no jump forward and
> back causes some studdering. Transparent OSD's make the problem worse.
> I don't expect OSD to work with HDTV until we have chromakey OSD, but
> at least this does not seem to be losing frames.

As a thought, can we not use the "Chromakey" OSD with a partial decode
(So you only re-decode the OSD-Covered part of the video, and mix,
then use the Chromakey to overlay the Software-decoded partial Video
stream and OSD) to "Fake" a transparent OSD... And it's a real shame
you can't use the Alpha channel of an RGBA "Surface" with a Chromakey
OSD.

> I've also fixed the missing progress bar and bad end time on playing
> back recordings with XvMC, this was caused by SyncPositionMap() in
> DecoderBase not functioning on the second AVFD::OpenFile(),
> needed to reinit ffmpeg with a different video codec. The solution
> was to simply not call it on opening the same file a second time.
> 
> This does point to another problem in avformatdecoder, the 800-1000
> hour recording time estimations for one hour programs were based on
> the calculated bit rate, obviously the calculated bit rate is way
> off, this could cause serious problems with buffering.

There's more problems with avformatdecoder, it seems. MythCommFlag is
failing on my box with "Aspect Ratio changed from 1 to 0 at frame -1"
and "CommDetect: Width or Height is 0, unable to process frame" for
every frame in that program (Which gives 0 commercials detected,
obviously)
-- 
Robert "Anaerin" Johnston