[MPlayer-users] Mencoder 1.0pre4 and DVD rip

[D Richard Felker III]

On Mon, May 10, 2004 at 12:39:49AM +0300, Ville Saari wrote:
> On Sat, May 08, 2004 at 02:31:08PM -0700, Loren Merritt wrote:
> 
> > Also, mbd=2 is always better than mbd=1 (aka vhq).
> 
> Always? Are you sure?
> 
> I encode mostly using constant quantizer. I assume the quality should
> then be pretty much constant and better encoding options should produce
> smaller files.
> 
> I once compared vhq and mbd=2 by encoding ten full-length movies using
> both. Nine out of ten came out smaller with vhq! Some quite significantly,
> up to five per cent. I couldn't see any perceptible difference in quality.

Of course it comes out smaller. mbd=1 means always choose encoding
with smallest size. mbd=2 means optimize for minimal size*error. So
mbd=2 will give better quality at constant quantizer.

If you're encoding with a bitrate rather than constant quant, it's not
so clear which is better. But all PSRN tests have shown mbd=2 to give
better results.

> I've been stuck to vhq ever since. My lavcopts are usually something like
> this:
> 
> vcodec=mpeg4:vmax_b_frames=2:v4mv:vhq:qpel:keyint=120:lmin=2:lmax=2:vb_qfactor=2

All tests I've seen shown 2 B frames to hurt quality. 1 B frame
sometimes helps, but any more just hurt.

Rich