[mythtv-users] Frames per second?

Jeff Burricelli jeff at burstable.net
Mon Apr 21 21:34:14 UTC 2003


Looks like the gbuffers option in bttv has fixed the fps problem.   Thanks
again.

Jeff
----- Original Message -----
From: <jeff at burstable.net>
To: "Discussion about mythtv" <mythtv-users at snowman.net>
Sent: Monday, April 21, 2003 4:18 PM
Subject: Re: [mythtv-users] Frames per second?


> Thanks, I'll try the gbuffers option tonight when I get home.
>
> --Jeff
>
>
> On Mon, 21 Apr 2003, Ray Olszewski wrote:
>
> > At 03:32 PM 4/21/2003 -0400, jeff at burstable.net wrote:
> > >It doesnt appear to be 'torn'.  The best I can describe it would be if
> > >playing a PC game on a underpowered pc and the FPS drop under 20 fps or
> > >so.  Typical CPU usage is about 60% - 70% when watching live tv.  Oh
and
> > >the deinterlace filter is one.
> >
> > It is always hard to be certain from these sorts of descriptions (the
one
> > above and your earlier one), but what you describe sounds like the
capture
> > process is dropping frames for some reason. The usual one is that the
image
> > is sufficiently rich in complexity that the encoder can't work fast
enough
> > on one frame before it has to get the next one, so some are lost. In
this
> > context, "complexity" is anything that makes encoding harder to do,
> > including simple noise (I get a 50% drop rate in frames when there is
> > nothing connected to the Television input, for example).
> >
> > Panning motion is a good way to introduce this sort of complexity ... I
> > don't tape sporting events, but I've seen similar sorts of problems in
The
> > Simpsons, at the beginning where the "camera" (the animation PoV) pans
over
> > the field full of supporting characters, or a bit earlier where
sometimes
> > it pans and zooms to the school for a Bart Blackboard Quote.
> >
> > If I'm right, I think your only options are (A) get a faster computer or
> > (B) reduce the pixel size of the capture.
> >
> > Oh, one thing you should check: when you load the bttv module, are you
> > including a suitable gbuffers= argument? (I usually use gbuffers=32 with
my
> > non-Myth systems, and I expect I'll do the same when I finally get Myth
> > running here. If MythTV doesn't work the way I'm assuming here, I hope
> > someone will correct me.) This can help buffer against brief bursts of
> > complexity in the frame sequence.
> >
> >
> >
> > _______________________________________________
> > mythtv-users mailing list
> > mythtv-users at snowman.net
> > http://lists.snowman.net/cgi-bin/mailman/listinfo/mythtv-users
> >
>
> _______________________________________________
> mythtv-users mailing list
> mythtv-users at snowman.net
> http://lists.snowman.net/cgi-bin/mailman/listinfo/mythtv-users




More information about the mythtv-users mailing list