[MPlayer-users] Re: failed to compile mplayer(cvs) with xvid

[yong]

> Ok, here's try 3. This compiles and works for me on mingw. It appears
> that libavcodec xvid encoding (all xvid, not just the ratecontrol) was
> broken on MinGW. The patch fixes that as well.
> 
> For the record, the problem with my last patch was that, unlike
> mkstemp(), tempnam() doesn't actually create a file -- it just returns
> the name. I never bothered to put in an O_CREAT because the old code
> didn't have on either, which is why xvid+mingw didn't work before.
> 
> Please test this with whatever you want to throw at it. Bear in mind
> that:
> (a) vrc_strategy=1 is (for now) ignored for the first pass
> (b) xvid likes the first pass to be constant-quantizer (vqscale=2)
> 
> The existence of (b) makes (a) not much of a problem; at a constant
> quantizer, I think, the only real difference is the frame type
> decisions, but I haven't determined experimentally whether that is
> true or not, and, either way, how much is affected. Maybe this
> weekend...
> 
Ok i tested something like this:
-nosound -o nul: -ovc lavc -lavcopts vcodec=mpeg4:autoaspect:
vmax_b_frames=2:vpass=1:vb_qfactor=1.50:vb_qoffset=1.00:vbitrate=600
:vb_strategy=1
:turbo -passlogfile "fmp4.txt"

-audio-preload 2.0 -o "%userprofile%\desktop\%~n1.avi"
-ffourcc mp4v -oac copy -ovc lavc-lavcopts vcodec=mpeg4:autoaspect
:keyint=300:vmax_b_frames=2:vpass=2
:vb_qfactor=1.50:vb_qoffset=1.00:vbitrate=600:predia=2:dia=2:preme=2:trell
:mv0:mbd=2:qprd:cbp:v4mv:naq:lumi_mask=0.3:dark_mask=0.3:p_mask=0.1
:vrc_strategy=1 -passlogfile "fmp4.txt"

Since you said vrc_strategy doesnt work in 1st pass,
i put it at 2nd pass,
but mencoder ignored the vbitrate=600, resulted the 2ndpass video
encodes at 1800kbps++ for 320x240 resolution...
Am i doing something wrong?