[MPlayer-users] Re: Encoding "dark" movies

[Martin Pavon]

On Sat, Dec 07, 2002 at 12:09:10PM +0100, Rémi Guyomarch wrote:
> [Automatic answer: RTFM (read DOCS, FAQ), also read DOCS/bugreports.html]
> On Sat, Dec 07, 2002 at 11:31:40AM +0100, Stephan André Schmidt wrote:
> > [Automatic answer: RTFM (read DOCS, FAQ), also read DOCS/bugreports.html]
> > Hello,
> > 
> > when encoding "dark" movies like "Panic Room" or "From Hell", the assumed 
> > bitrate to end up in a divx about 1400 MB, i.e. about 1390 for "From 
> > Hell", does end in an avi of around only 900 MB (3-pass encoding).
> 
> Libavcodec is simply too powerful.
> 
> > Giving 
> > the bitrate of about 1200 to mencoder does not make any difference to the 
> > file size of the avi; it is also around 900 MB. Are there any special 
> > filters or options to libavcodec handling this kind of "dark" movies, so 
> > that the "calculated" avi size will be reached after encoding?
> 
> dd if=/dev/urandom bs=1m count=500 >> your_900mb.avi
> 
 You get a 1400mb file indeed, but ... :-)

 
> Every single frame is already encoded at the minimum quantizer (2) and
> libavcodec simply can't produce more bits.
> 
> Hints (choose one or more and retry) :
> - use a sharper rescaling method, like -sws 9 or 10
> - use mpeg quantizers
> - increase encoding resolution, try scale=704:304,crop=704:432:8:75
>   (only scale Y axis)
> - go for one CD
> 

I found that sometimes  one gets undersize files when the  fps is wrong; if
you dvd is ntsc try with -ofps=23.976  (but I don't think you can get 500mb
this way)

But three or four times, I found that fps is right and I got undersize movies
in this cases I could get the desire target size setting vqmin=1, but some
things must be said:

a) The libavc-options.txt prevent us about this value

vqmin 2-31 (minimum quantizer) for pass1/2
 1 is not recommended (much larger file, little quality difference (if u
 are lucky and other weird things (if u are less lucky))

b) In my case, I was a lucky one; I can't claim that the quality improves by
reach the target size, but the quality is not lower.

c) I had this  problem before I try to get good quality  files in one CD, I
think now, that the intersting test is  two cds with vqmin=1 against one cd
without vqmin=1.

Regards,

Martín
-- 
Martín Pavón
martin_199ar at yahoo.com.ar
http://mate.dm.uba.ar/~mpavon/