[MPlayer-dev-eng] Re: [PATCH] dvr-ms fixes for pts, key frame detection and seeking

[John Donaghy]

> About 0.4 seconds as in your example or are there bigger ones?

I only have one sample that display the problem so I'm not sure if there are
bigger ones or not.

> If it stays out of sync what does the status line show?

My machine isnt powerful enough to play these HD files properly without
-framedrop (it's aP4 2.8GHz with 1Gb RAM running WindowsXP with
hyper-threading) - so I'm getting testing help from Jon Elwood, who said in
response to this question:

"Before the switch to progressive, the A-V number was 0.0.  At the switch is
jumped to 0.5 and then settled to 0.3.  It stayed at 0.3 and never
recovered.  I can't really tell how much the sound is ahead of the video,
but I would say maybe ½ second (this is really a guess, I can't figure out a
good way to time it).

One other thing you should know.  It looks like this is only a problem
during the switch to progressive.  I have other progressive shows where it
is progressive from the start and the AV is in sync.  Because of this, I
really think it's just having a problem with the "switch"."

The other thing I can tell you is that he converted some samples that
exhibit the problem to ".mpg" files using DirectShow. The resulting files
played fine in WMP but, in MPlayer they had the exact same sync issue as the
dvr-ms version does with my proposed patch. This leads me to believe that
this is not an asf demuxing issue.

I posted a converted mpg sample that shows the problem to ftp.mplayerhq.hu a
while back (Ugly Betty_WPTA-DT_01_02_2007_19_56_07.mpg)

John