[MPlayer-users] Re: -ss switch broken in current 3-pass-mode of mencoder

[Josef Wolf]

I am really impressed how fast your responses come! Thanks again!

>> With this patch the second and third passes succeed _if_ the first
>> pass succeed. But whether the first pass succeed depends on the values
> huh?

>> given to the -ss and the -endpos parameters.
>> For example:
>>   "mencoder -other_options_snipped -ss 1:48 -endpos 3:00" fails
>>   "mencoder -other_options_snipped -ss 1:48 -endpos 4:00" succeed

> why should the first pass fail at all? it should NOT! NEVER.

Well, it does not really fail. Actually, it creates a frameno.avi,
which seems to have (at least approximately) the correct size.
But it behaves differently than usual: it forgets to print the
vbitrate-recommendations. And it seems to me that the failure of the
second/third pass are caused by this different behavior of the first
pass. I conclude this from the fact that the second pass fails
_exaclty_ when the first pass forgets to print the recommendations.

Maybe the "mux_v->timer>100" condition in line 1240 of mencoder.c is
not met?

> it does nothing with this code nor the patch,

Hmmm, maybe there is some other problem, which was not triggered as
long as demux2 was at the wrong position?

> unless you forget to delete frameno.avi!?

I can definitely rule out this cause of failure, because I have a
script which does all those 3-pass-stuff (and some more things).

-- 
-- Josef Wolf -- jw at raven.inka.de --