[XviD-devel] [CVS commits] decoder speedups and fixed cbr

[Bryan Mayland]

Edouard Gomez wrote:

>On GNU/Linux box i use mplayer to measure decoding time:
>mplayer -benchmark -nosound -vo null -vc xvid sequence.avi
>  
>
    This seemed a lot easier than building mplayer on my windows box, 
but I may still try it later (hey tomorrow is a hurricane day, so if the 
power is on...).  Anyway, here's results from a Pentium4 2.4B on SuSE 9.1.
Source is 54244 frames of 512x384 material.  VBR 2 b-frames max
XviD-1.0.1  BENCHMARKs: VC: 318.849s VO:   0.294s A:   0.000s Sys:   
7.876s =  327.018s
XviD-CVS BENCHMARKs: VC: 153.663s VO:   0.318s A:   0.000s Sys:   7.394s 
=  161.375s

Ummm is it me or are my results a-typical?  Reproducable too.

>Yes, bvop qp is always the result of ratio*qp+offset.
>  
>
    Thanks for this.  I didn't know if I'd be able to understand the 
code enough to figure it out on my own.