[MPlayer-users] Encoding problems / tips needed

JiHO jo.irisson at noos.fr
Tue Aug 3 14:13:33 CEST 2004


Loren Merritt wrote:
> 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.

ok, so lavc rules! indeed I was around 6fps for lavc and 1fps for XviD 
(Duron 950MHz...). Just one question again: did you notice my email 
about the motion estimation prepass that you advised? I wondered when to 
do it: at first, second or both video passes?


> 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.
Martin Collins wrote:
 >>	Example: in order to fit the movie on two 700 Mb CDs
 >>		target_size = 1 400 000 000
 >
 >
 > Given that there are usually 703MB on a 700MB disc, and that 1MB is
 > 1024B I make target size here 1 474 297 856, which may go some way to
 > explaining why your encodes are 30MB smaller than a full disc.

Ok so both theses errors (bytes caculation and overhead percentage) 
could explain my 60Mo of error (yes it was in fact much more that 30Mo) 
on a 1,4 G file. Thanks for all this I will try a new encoding with new 
caculation and additional options.


>>mplayer dvd://1 -alang en -vc dummy -vo null -ao pcm -aofile
>>foo.wav oggenc -q 6 foo.wav -o foo.ogg
> 
> I use -q 0 and find it adequate. You may want to use more but 6 is
> probably way more than necessary.

I heard a difference between q 5 and 6 using headphones. The difference 
is mainly noticeable in the stereo effect: it is much better with q 6. 
and that seems to be what is advised on ogg-vorbis FAQ:
" For now, quality 0 is roughly equivalent to 64kbps average, 5 is 
roughly 160kbps, and 10 gives about 400kbps. Most people seeking 
very-near-CD-quality audio encode at a quality of 5 or, for lossless 
stereo coupling, 6."
in fact oggenc with "q 6" produces an audio quality near or equal to 
lame "preset standard" with 20% smaller files, that's why I used this 
instead of lame.

thanks again for this, after a new test, scripting and tutorial editing 
should begin soon.
-- 
JiHO
---
   Windows, c'est un peu comme le beaujolais nouveau :
à chaque nouvelle cuvée on sait que ce sera dégueulasse,
     mais on en prend quand même, par masochisme.
---




More information about the MPlayer-users mailing list