On date Wednesday 2014-11-19 11:47:53 +0100, Nicolas George encoded: > Le nonidi 29 brumaire, an CCXXIII, Stefano Sabatini a écrit : > > I'm concatenating samples with different time bases. Apparently > > concatdec is setting the first stream timebase as the overall > > timebase, and always assuming the timebase is shared for all the input > > files. > > Concatenating segments with different settings is not currently supported, > and that includes the time base. But inconsistencies are not detected > either. > > For the time base specifically, I suppose implementing the conversion and a > smart time base selection would be enough. With just the conversion, you > might get lucky if the time base of the first segment is accurate enough. >
> I do not object to the patch implementing the conversion, it will be needed > at some point anyway to support segments with different settings, but a > comment that this is incomplete would probably be a useful remainder. I agree. At the moment nor the code nor the documentation mentions this. Ideally the library should complain in case different settings are detected. As for the timebase issue, probably a workaround can be found using a settb filter. -- FFmpeg = Friendly & Fanciful Miracolous Picky Educated Gargoyle _______________________________________________ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
