[MEncoder-users] Mencoder failure with -ss and -endpos

Pete Davis pete at petedavis.net
Wed Jul 20 14:15:15 CEST 2005


> 
> Hi,
> On Tue, Jul 19, 2005 at 04:42:38PM -0500, Pete Davis wrote:
> > This worked for a 43 second block, but for a 90 second block, the second
> > pass fails:
> [...]
> > It stops at 80.8 seconds into the second pass with:
> >
> > Pos:  80.8s   1956f (23%)  27fps Trem:   3min  49mb  A-V:0.037
> [1105:128]
> > 1 duplicate frame(s)!
> > mencoder: ratecontrol.c:644: ff_rate_estimate_qscale: Assertion
> > `picture_number<rcc->num_entries' failed.
> 
> This means that the firstpass log contains only data for 80 seconds. You
> have "duplicate frame" messages there, maybe you need to use -vf
> harddup?
> Or -mc 0 -noskip? or the right framerate for output with -ofps. Or some
> inverse telecince (sorry, don't remember what your input format was).
> 
> Greetings,
> Reimar Döffinger
> 

I got the latest build and it's not crashing anymore. The input source is a
DVD with mixed progressive and telecine, so as per 7.14.3.4 of the docs, I'm
using -vf pullup,softskip and -ofps 24000/1001

I still get quite a few duplicate frames and skipped frames, as well as
messages that: 30000/1001fps NTSC content detected and 24000/1001fps
progressive NTSC content detected along the way. Is this normal?

If I encode the entire episode, then go into Windows Media Player and get a
period I want to record and use that in -ss and -endpos, it doesn't line up
at all. For example, if the period is 00:07:38-00:8:08 when viewing it and I
use that for -ss -endpos, it really encodes a section more than minute
earlier. What would be causing this?

Pete






More information about the MEncoder-users mailing list