Re: What are ATI chips able to do in hardware (regarding DVD playback)? (was: Re: [xine-user] I get it now, all Xine is is a cheap immitation of Power DVD!)

[B. <rb...@in...>]

On May 06 2002, Thibaut Mattern wrote:
> On Monday 06 May 2002 21:49, Rog=E9rio Brito wrote:
> > 	I'm asking this because I'd love to see extra CPU time
> > 	available on this G3 600MHz so that it would be possible to
> > 	deinterlace some DVDs (frankly, interlacing destroys a good
> > 	part of fun of watching DVDs) under Linux.
>=20
> Have you tried onefield_xv deinterlace method ?

	Hi, and thanks for answering.

	In fact, no and I don't know if it would help, given the fact
	that the G3 is not able to play all the frames (i.e., it skips
	some of them -- read below), even when using XV.

	Would it help if I enabled onefield_xv?

> > 	I perhaps should have not bought this computer if I knew at
> > 	first that a G3 computer with 600MHz would play videos as well
> > 	as, say, a Celeron 466MHz with MMX instructions.

	Now, with more substantial data, I can say that xine plays
	DVDs on my iBook 600MHz skipping around 35/200 frames, while
	the Celeron 466MHz plays the same thing skipping around 40/200
	frames, that is, technically the same thing, given the
	imprecisions in measurements.

	Also, the Celeron has "only" 64MB, of which a good part is
	taken by other programs running in the background, while the
	iBook has 128MB. This means that the Celeron's VM subsystem is
	under higher pressure than the iBook's, which would lead me to
	think that the Celeron would be a tiny bit faster if it had
	more memory.

	So, perhaps compiling xine with a newer GCC and/or using X 4.2
	(with DRI) would help the G3 a bit.

	The options would be to use more hardware acceleration
	(probably not happening) and (perhaps) hand-tweaking xine to
	help the compiler generate better code (like, say, declaring
	relevant variables with register etc) or using assembly for
	some snippets of code (which I am not able to do right now, as
	I'd have to learn about that).


	[]s, Roger...

--=20
=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D=
-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D
  Rog=E9rio Brito - rb...@in... - http://www.ime.usp.br/~rbrito/
=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D=
-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D