[FFmpeg-user] inconsistent ffmpeg results when running same command on different machines

Frank Schwieterman fschwiet at gmail.com
Mon Aug 12 15:10:44 CEST 2013


  Ok, thanks for the info.

On Mon, Aug 12, 2013 at 12:45 AM, Anatol <anatol2002 at gmail.com> wrote:

> The same ffmpeg/Linux binary running on different hardwares or different
> Linux version might produce different output files.
> We had this with MPEG-TS decoding - there was bug in ffmpeg 0.10, it used
> multi-threaded decoding as a default. Some machines handled it properly,
> while others had nasty artifacts. It had to do with type of hw acceleration
> that each machine had and the drivers that it used.
>
>
> On Mon, Aug 12, 2013 at 1:57 AM, Frank Schwieterman <fschwiet at gmail.com
> >wrote:
>
> >   Thanks for the insight Moritz.  I suspect that the libx264 version is
> > determined by whoever puts together the build, and so picking up windows
> > and mac builds there isn't a straightforward way to keep those in sync.
> >  Getting a binary match is not critical, so this is ok.
> >   One concern though is that the result lengths are different.  I am
> trying
> > to extract a clip of a specific length, which is going to be concatenated
> > for a loop effect.  It is important that the concatenated portion is a
> > specific time duration because the loop establishes a musical tempo.
> >   The command I run should extract video from time 0.1 to 0.2, so I'd
> > expect the output result to be 0.1 seconds long.  From the output, it
> says
> > it is 0.11s or 0.12s instead (I see "frame= 3 fps=0.0 q=-1.0 Lsize= 133kB
> > time=00:00:00.12 bitrate=8851.7kbits/s dup=6 drop=0" in one). I suppose
> > 0.02 error on the duration is reasonable, if the error stays the same as
> > the clip is extended. Would you know if I could expect a similar
> fixed-size
> > error regardless of the duration of the extracted clip?
> >
> > On Sun, Aug 11, 2013 at 3:04 PM, Moritz Barsnick <barsnick at gmx.net>
> wrote:
> >
> > > On Sun, Aug 11, 2013 at 14:51:49 -0700, Frank Schwieterman wrote:
> > > > I suspect the ffmpeg configuration is different...
> > >
> > > The ffmpeg output (at the very top) will hint you towards that.
> > >
> > > > Is it possible to configure things such that I'd get the same result
> > > > from both machines?
> > >
> > > I guess if the identical versions of ffmpeg and all used libraries are
> > > identical (same versions, same configuration, not necessarily same
> > > compilers though), it should work.
> > >
> > > Your ffmpeg base code seems in sync (judging from the reported versions
> > > of the libav* libraries).
> > >
> > > But since these two differ version-wise:
> > > > [libx264 @ 0x7ff461908e00] 264 - core 125 - H.264/MPEG-4 AVC codec -
> > > Copyleft 2003-2012 - http://www.videolan.org/x264.html - options:
> > cabac=1
> > > ref=3 deblock=1:0:0 analyse=0x3:0x113 me=hex subme=7 psy=1
> > psy_rd=1.00:0.00
> > > mixed_ref=1 me_range=16 chroma_me=1 trellis=1 8x8dct=1 cqm=0
> > deadzone=21,11
> > > fast_pskip=1 chroma_qp_offset=-2 threads=6 lookahead_threads=1
> > > sliced_threads=0 nr=0 decimate=1 interlaced=0 bluray_compat=0
> > > constrained_intra=0 bframes=3 b_pyramid=2 b_adapt=1 b_bias=0 direct=1
> > > weightb=1 open_gop=0 weightp=2 keyint=250 keyint_min=25 scenecut=40
> > > intra_refresh=0 rc_lookahead=40 rc=crf mbtree=1 crf=23.0 qcomp=0.60
> > qpmin=0
> > > qpmax=69 qpstep=4 ip_ratio=1.40 aq=1:1.00
> > > > [libx264 @ 0000000002246da0] 264 - core 130 r2274 c832fe9 -
> > H.264/MPEG-4
> > > AVC codec - Copyleft 2003-2013 - http://www.videolan.org/x264.html -
> > > options: cabac=1 ref=3 deblock=1:0:0 analyse=0x3:0x113 me=hex subme=7
> > psy=1
> > > psy_rd=1.00:0.00 mixed_ref=1 me_range=16 chroma_me=1 trellis=1 8x8dct=1
> > > cqm=0 deadzone=21,11 fast_pskip=1 chroma_qp_offset=-2 threads=6
> > > lookahead_threads=1 sliced_threads=0 nr=0 decimate=1 interlaced=0
> > > bluray_compat=0 constrained_intra=0 bframes=3 b_pyramid=2 b_adapt=1
> > > b_bias=0 direct=1 weightb=1 open_gop=0 weightp=2 keyint=250
> keyint_min=25
> > > scenecut=40 intra_refresh=0 rc_lookahead=40 rc=crf mbtree=1 crf=23.0
> > > qcomp=0.60 qpmin=0 qpmax=69 qpstep=4 ip_ratio=1.40 aq=1:1.00
> > > you can suspect that they won't create bitwise identical output in the
> > > payload.
> > >
> > > Moritz
> > > _______________________________________________
> > > ffmpeg-user mailing list
> > > ffmpeg-user at ffmpeg.org
> > > http://ffmpeg.org/mailman/listinfo/ffmpeg-user
> > >
> > _______________________________________________
> > ffmpeg-user mailing list
> > ffmpeg-user at ffmpeg.org
> > http://ffmpeg.org/mailman/listinfo/ffmpeg-user
> >
> _______________________________________________
> ffmpeg-user mailing list
> ffmpeg-user at ffmpeg.org
> http://ffmpeg.org/mailman/listinfo/ffmpeg-user
>


More information about the ffmpeg-user mailing list