[MPlayer-dev-eng] Re: Re: PATCH [0/12] CoreAVCDecoder support

[Michael Niedermayer]

Hi

On Tue, Feb 20, 2007 at 12:28:04PM +0100, Reimar Döffinger wrote:
> Hello,
> On Tue, Feb 20, 2007 at 11:57:29AM +0100, Michael Niedermayer wrote:
> [...]
> > again the goal is to make ffh264 faster not figure out what coreavc does and
> > copy that blindly
> 
> In addition I'd like to mention that at least some of those extra-fast
> decoders (I think it was a GPU-based one) exhibit the same visual defects
> as ffh264 with skiploopfilter, i.e. they violate the standard.

AFIAK coreavc violates it too, at least if my memory doesnt fail it didnt
provide the same md5 as ffh264 (and ffh264 has binary identical output
for many conformance bitstreams, i didnt check too closely if all which we
claim to support give binary idetical output)


> And of course as always the main problem is the lack of people actually
> working on the code, though the size of the code admittedly does not
> make it easier in this case ;-)

well just take a random page of code put STRAT/STOP_TIMER around it and 
simplify it, as loren already said the majority of the code in h264.c is
executed once per mb in some videos
or just ask me ill give you some piece you can work on if you like :)
for example look at libavcodec/i386/h264dsp_mmx.c and turn the 
for() asm; into a normal pure asm loop, this generally makes the code
quite a bit faster even if you just write naive asm code without any
tricks


[...]
-- 
Michael     GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB

Let us carefully observe those good qualities wherein our enemies excel us
and endeavor to excel them, by avoiding what is faulty, and imitating what
is excellent in them. -- Plutarch
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
URL: <http://lists.mplayerhq.hu/pipermail/mplayer-dev-eng/attachments/20070220/13cf5f64/attachment.pgp>