[MEncoder-users] [BUG] a-v sync problems with pullup, softskip (only on some DVDs)

Corey Hickey bugfood-ml at fatooh.org
Wed Mar 16 22:08:37 CET 2005


Oded Shimon wrote:
> On Wed, Mar 16, 2005 at 10:35:01AM -0500, D Richard Felker III wrote:
> 
>>back, but please telle me you command line in any case so that I don't
>>miss something critical!
> 
> 
> Didn't he give that in the first e-mail of this thread?...
> 

Yes, but I meant to re-send it anyway. I was bugreporting under the 
influence of lack of time. :)

Now, surprisingly(?), when I try encoding using your recent patch, the 
A-V desync seems to have gone away entirely - it looks-sounds perfect to 
me now. If the patch shouldn't have fixed the problem, then a bug may 
still be there, but it doesn't manifest in this clip any more. Pullup, 
however, still duplicates lots of frames when the star wars intro text 
fades into the distance and that is visible in the encoded clips.

------------------

Anyway, the sample command is:

mencoder starwars.vob -aid 128 \
-vf pullup,softskip,crop=714:352:2:62,scale=688:288 \
-ovc lavc -oac copy -ofps 24000/1001

Replace starwars.vob with buffer.vob, combined.vob, etc.

-------------------

There are now several vobs from different cuts; to summarize:

starwars.vob
The source. Way too big to upload entirely. :)

bigpart.vob
455 MB from the beginning of starwars.vob. Still quite big, so the other 
files are cuts of the relevant sections.
dd if=starwars.vob of=bigpart.vob bs=8192 count=58192

part1.vob:
The intro text fading into the distance. Pullup duplicates lots of 
frames, and (pre-patch) the a-v used to get out of sync.
dd if=starwars.vob of=part1.vob bs=8192 count=2560 skip=8192

part2.vob
There wasn't any way to tell that encoding from part1.vob got out of 
sync; part2.vob has speaking parts and can be combined with part1.vob to 
make the desync visible.
dd if=starwars.vob of=part2.vob bs=8192 count=8192 skip=50000

combined.vob
The combination of part1.vob and part2.vob, to make the desync visible. 
This behaves the same as encoding bigpart.vob; the data between 
part1.vob and part2.vob don't seem to affect the a-v sync.
cat part1.vob part2.vob > combined.vob

buffer.vob
The smallest cut I could find that still made pullup have buffer 
problems. No buffer problems after the patch, but still lots of 
duplicate frames.
dd if=starwars.vob of=buffer.vob bs=8192 skip=2172 count=8550

------------------

Thanks,
Corey




More information about the MEncoder-users mailing list