[MPlayer-DOCS] Mess in mencoder.xml's examples

Guillaume POIRIER poirierg at gmail.com
Wed Nov 23 10:02:20 CET 2005


Hi,

On 11/23/05, Diego Biurrun <diego at biurrun.de> wrote:
> On Wed, Nov 23, 2005 at 09:10:47AM +0100, Guillaume Poirier wrote:
> > Diego Biurrun wrote:
> > > On Mon, Nov 21, 2005 at 11:13:13AM +0100, Torinthiel wrote:
> > >
> > >>I've found that examples in mencoder.xml (and probably
> > >>encoding-guide.xml as well, but I haven't examined it yet) are in a bit
> > >>of mess. So:
> > >>
> > >>There are two examples of two-pass encoding. One with copying audio and
> > >>one with reencoding it. Can someone tell me why in the first output from
> > >>first pass goes to /dev/null and in the second to output.avi? IMHO it
> > >>should be /dev/null both.
> > >
> > > Can anybody with some encoding knowledge answer this?
> >
> > Well, the idea is too keep only the video from the last pass of
> > multipass encode. Why is that? That's because the last pass produces a
> > video with greater quality that all the previous ones.
> > Why do you want to write the previous pass to /dev/null? Mainly because
> > you won't keep those files anyway (but keep in mind that even if the
> > video is written to /dev/null, the stat file is still written on disk),
> > and because that way you don't generate needless I/O.
>
> IOW /dev/null is preferable since it speeds up the encoding process,
> right?  Then we should use /dev/null in our examples.

In _theory_, it should speed-up encoding process a bit. Now, I have
not made any benchmarking to back this up... but this is just common
sense IMHO. Also, in addition to possibly speed-up encoding, it can
also save some disk space if the first passes are done at low constant
qp for instance.

Guillaume
--
Remember, if you ever need a helping hand, it's
at the end of your arm, as you get older, remember
you have another hand:
The first is to help yourself,
the second is to help others.
-- Audrey Hepburn




More information about the MPlayer-DOCS mailing list