[MPlayer-users] H.264 and advice on new hardware.

[Dominik 'Rathann' Mierzejewski]

On Wednesday, 22 March 2006 at 09:18, Mathieu Monnier wrote:
> >> And there are still a lot of optimizations needed in order to do so
> 
> It didn't come out as I wanted. I didn't want to say "there are lots of 
> parts that can be mmxed", but "there are lots of parts that must be 
> speeded up". And since I don't believe that speeding up,  let's say, 
> cabac decoding and pixel processing will bring the missing speed ( they 
> are already fast ), and since CoreAvc is so much faster, I can only 
> deduce that the core of the issue lies in the architecture of the decoder.

I wonder if the output is bit-exact with a reference decoder.
Otherwise it might just be sacrificing visual quality for speed.

> > 1. I have an overclocked Newcastle Athlon64 3400+ running at 2556
> > MHz, and it can almost, just-barely-not-quite decode those two videos
> > in realtime. There are scattered scenes where the video drops behind
> > by up to 0.4 seconds, but it catches back up after a moment; overall,
> > it never gets left behind.
> 
> But it's not acceptable. Video must be played realtime.
>
> > Every once in a while Loren Merritt makes a slight performance 
> improvement to lavc's h.264 decoder. Over time, these add up.
> 
> Cf previously, these speed ups won't ever add up to 50%

Never say never. 

> > Using '-lavdopts skiploopfilter=nonref' will speed up decoding 
> significantly without any readily-visible differences.
> 
> Not on a P-only video.

Who in their right frame of mind would make a P-only video?

R.

-- 
MPlayer developer and RPMs maintainer: http://rpm.greysector.net/mplayer/
There should be a science of discontent. People need hard times and oppression
to develop psychic muscles.
	-- from "Collected Sayings of Muad'Dib" by the Princess Irulan