[MPlayer-dev-eng] check recommended MPEG2 encoding options

[Reimar Döffinger]

Hello,
On Mon, Mar 05, 2007 at 06:58:39PM -0500, Rich Felker wrote:
> On Tue, Mar 06, 2007 at 12:43:25AM +0100, Reimar Döffinger wrote:
> > http://www.mplayerhq.hu/DOCS/HTML/en/menc-feat-vcd-dvd.html
> > gives this for high quality encode:
> > -lavcopts vcodec=mpeg2video:vrc_buf_size=1835:vrc_maxrate=9800:vbitrate=8000:\
> > keyint=15:trell:mbd=2:precmp=2:subcmp=2:cmp=2:dia=-10:predia=-10:cbp:mv0:\
> > vqmin=1:lmin=1:dc=10:vstrict=0
> > 
> > Can someone double-check this a bit? First of all, the whole patch seems
> > to miss a reference to the absolutely critical two-pass.
> 
> 2pass is useless in extremely-restricted CBR.

Even assuming rate control properly respects those restrictions I'd say
they easily allow for bits to be allocated very badly between two
consecutive frames. Just theoretically though.

> > I have some doubts that it is a good idea to set vqmin/lmin that low,
> 
> It's necessary if you don't want the bitrate to drop in still scenes
> (i.e. if you really want CBR).

hm.. So that is supposed to be the "please waste bits if necessary"
option? I guess I should read up and see if there's a way to use
quant==1 only if all others are already at 2.

> > the man page says it is a bad idea. man page also suggests (by my
> > reading) that negative values for dia are fast but not exactly good.
> > And don't B-frames also exist for MPEG-2 and are allowed for DVD?
> 
> Required, iirc... On the other hand, lavf cannot mux them.

Though my impression was that the MPEG-muxer in lavf is suboptimal
anyway (and we recommend the for MPEG-4 that seems much more likely to
be muxed with lavf to me)...

Greetings,
Reimar Döffinger