[MPlayer-users] [BUG] ratecontrol.c:693: ff_rate_estimate_qscale: Assertion `pict_type == rce->new_pict_type' failed.

[Guillaume POIRIER]

Hi,

On 5/20/06, vitko <vitko at post.cz> wrote:
> Today I was reviewing encoding tips at
>
> http://www.mplayerhq.hu/DOCS/HTML/en/menc-feat-enc-libavcodec.html#menc-feat-dvd-mpeg4-lavc-encoding-options
>
> and was playing with "very high quality" options suggestion, albeit with 2 pass
> mode. Suggested options

Nice to see someone is reading the docs ;)


>     -lavcopts vcodec=mpeg4:mbd=2:mv0:trell:v4mv:cbp:last_pred=3:predia=2:dia=2: \
>     vmax_b_frames=2:vb_strategy=1:precmp=2:cmp=2:subcmp=2:preme=2:qns=2:vpass=2
>
> end with
>
>     ...
>     [VE_LAVC] High quality encoding selected (non-realtime)!
>     [mpeg4 @ 0x86a25b0]b_frame_strategy must be 0 on the second pass
>     Could not open codec.
>     ...
>
> Ok, let vb_strategy = 0, so with revised opts

I'd just not set vb_strategy at all on second pass.


>     -lavcopts vcodec=mpeg4:mbd=2:mv0:trell:v4mv:cbp:last_pred=3:predia=2:dia=2: \
>     vmax_b_frames=2:vb_strategy=0:precmp=2:cmp=2:subcmp=2:preme=2:qns=2:vpass=2
>
> I get
>
>     ...
>     videocodec: libavcodec (720x304 fourcc=34504d46 [FMP4])
>     [VE_LAVC] High quality encoding selected (non-realtime)!
>     Pos:   0.0s      2f ( 0%)  0.16fps Trem:   0min   0mb  A-V:0.004 [0:0]
>     1 duplicate frame(s)!
>     Writing header...
>     ODML: Aspect information not (yet?) available or unspecified, not writing
> vprp header.
>     Setting audio delay to 0.040s.
>     Writing header...
>     ODML: Aspect information not (yet?) available or unspecified, not writing
> vprp header.
>     Setting audio delay to 0.040s.
>     [mpeg4 @ 0x86a25b0]Error, Invalid timestamp=1, last=1  A-V:0.012 [0:0]
>     Writing header...
>     ODML: Aspect information not (yet?) available or unspecified, not writing
> vprp header.
>     Setting audio delay to 0.040s.
>     Pos:   0.1s      5f ( 0%)  0.40fps Trem:   0min   0mb  A-V:0.016 [0:0]
>     1 duplicate frame(s)!
>     mencoder: ratecontrol.c:693: ff_rate_estimate_qscale: Assertion `pict_type
> == rce->new_pict_type' failed.
>
> (and my bash script ends with SIGABRT).
>
> Is it a bug, or am I forcing mencoder to do something absurd?

It could be a bug, but I would just advise you to set vb_strategy to
whatever you want on 1st pass, and not set it in second pass.

Also, please make sure you are using the same number of bframes on both passes.

G

-- 
"Success consists of going from failure to failure without loss of enthusiasm."
 -- Winston Churchil