[MPlayer-dev-eng] [PATCH] autoq support for control()

David Holm david at realityrift.com
Sat Feb 9 17:26:18 CET 2002


> > res would need a swscale which I find unlikely would speedup playback,
> > or?
> swscale is a lot faster than mpeg encoding

I guess, I'll rethinkg the whole idea, disregard the patch...
 
> anyway, imho libavcodec could be even faster... at least compared to mp1e it
> is still ~50% slower, at least on my system.
> i've noticed that mp1e uses 2 nasty tricks: aligned globals, even for dct
> temp blocks and so on - it may save one register and extra indexing so it
> may speed up things. the other trick is cache optimization, similar to
> libmpeg2's slices, so instead of doing various steps (like dct coding,
> huffman coding etc) on small macroblocks, it group things, like dct for
> while slice, then huffman for whole slice etc.
> 

On my machine the switch to libavcodec made divx playback on the dxr3
alot faster but the output quality was not as high. I know you disagree
to both of those statements, but you are welcome to come see for
yourself. The difference is noticeable...
I was going to do some sse stuff to libavcodec but I noticed Michael
beat me to it, which was probably for the best anyway ;).

//David




More information about the MPlayer-dev-eng mailing list