[MPlayer-users] Playing mencoder's recorded AVI's onwindowsmedia player: AUDIO DESYNC

[D Richard Felker III]

On Fri, Sep 26, 2003 at 11:09:10PM +0200, Moritz Bunkus wrote:
> [Automatic answer: RTFM (read DOCS, FAQ), also read DOCS/bugreports.html]

> Hi.
> 
> One could argue that...
> 
> > I would strongly recommend not using matroska since it's inefficient,
> 
> Tests show that a typical Matroska file usually has lower overhead than
> the comparable AVI file, and sometimes more overhead. OGM always does
> much worse than both of them.

That's like saying Exxon's not evil because they don't hang people
like Shell does. Yes, Matroska has *less* overhead than AVI or OGM,
but still much more than is needed.

> > bloated to implement,
> 
> Which should not be a concern for users, only for coders.

It's a concern for users because coders may choose not to implement it
(or to only implement it by linking the c++ libs, which may not
compile on some systems since they're c++).

> > and just not designed very competently. Of
> > course, whatever format you choose to use now, you're free to convert
> > to the new NUT format once it's stabilized and supported by mplayer.
> 
> One could also argue that NUT's "design" consisted of about one week of
> developers sharing ideas about a year ago. Since then nothing has
> happened. And NUT will probably not spread any further than

No thanks to you...... :/

OK, enough of that gripe. Since then, it's been implemented by Alex
for ffmpeg (libavformat) and mplayer g2. ffmpeg can make .nut files
and ffplay can play them. There're still a few details up in the air
in the draft spec (which Alex is writing up in TeX), but it's pretty
much done.

And btw, about your comment about nut's "design" being simple and
quick, that's the whole point. The people who designed it are
competent enough to make a better standard in one week than the
matroska team could make in a year (or probably even their whole
lifetimes). Read the nut demuxer code and look how simple and elegant
it is.

> mplayer/mencoder itself as no one seemed to be interested in making it a
> widely used format (if it's going to be implemented at all, but I guess
> one day Rich will be fed up enough, sit down for a day and implement it
> ;)).

Well right now I'm fed up with mplayer g1's video filter layer and the
limitations of mencoder, so I'll probably work on some mencoder-g2
type stuff first. That will probably eventually include nut muxing
from mencoder-g2.

> Rich is contra Matroska, I'm pro Matrosa. I'd say you should take
> everything said be either of us with a good portion of skepticism. The
> truth is probably somewhere in between. Why don't you try Matroska for
> yourself and base your decision on your own experience?

Try it? How about read the spec, look how how incredibly ugly it is,
and then do the math, compute the overhead (in %), and see how
wasteful it is. There are lots of things that suck, which, on the
outside appear ok until you start doing research into the inner
workings.

Rich