[MPlayer-users] Encoding problems / tips needed

[Loren Merritt]

On Mon, 2 Aug 2004, JiHO wrote:
> Loren Merritt wrote:
>> vmax_b_frames=1 (usually significant quality gain, but has some
>> side-effects, including increased cpu time to decode (+5% on my computer),
>> and reduced compatibility with some hardware players)
>> preme=2:predia=-4 (helps almost as much as 'dia', and is much faster)
>> and several other options with smaller gains for larger speed penalties.
>
> thanks for all this. I will try. I'm right now trying XviD and the
> result seems better than lavc but it's wayyyy longer to encode (at the
> moment, mencoder is planning 4114 minutes for a 2h45 movie....). so
> maybe with these options I can achieve XviD quality faster.

Well, it all depends on settings. I've seen anywhere from 6fps to 20fps
from xvid, and anywhere from 2fps to 50fps from lavc (Athlon 2500,
video=720x480). At equal speeds, I get better results from lavc.

>> What do you hope to gain from this alternate method? Just access to
>> mencoder's bitrate calc?
>
> yes exactly, because it seems that the formula I mentionned is not 100%
> accurate or at least that the header size is not exactly estimated. for
> example on my last encoding I have more that 30 empty Mo that could have
> been used to improve the video quality (a bit ;-) )
> I can't find a way to calculate accurately the header size of an AVI or
> OGM container. in both docs I only find estimates of 1% and 1,2%
> respectively. do you know a documentation ressource where I can find that?

30 Mo? that sounds excessively innacurate.
I don't mux avi (except as an intermediate format), nor ogm at all, but I
can say that at my normal rates (~1200kbps mpeg4, ~130kbps vorbis),
a video-only avi costs 0,4% overhead, and a mkv with one video and one
vorbis track costs about 0,2% overhead.

--Loren Merritt