On 4/20/2018 9:22 PM, James Almer wrote:
Wouldn't this hint that the file in question is broken?
I don't follow. If the inversion was a mistake, then frames currently
not marked by demuxer are in fact keyframes and a stream decoded from
that frame onwards should show correct output. I see precisely the
opposite. Output generated by streamcopying from currently marked KF
decode correctly and those generated by streamcopying from currently
marked non-KF decode incorrectly*.
This should rule out errors in the muxer as well.
* ffmpeg -i logo-capcom.cpk -ss 5.4 -c copy -copyinkf -an copy-test.cpk
&& ffplay copy-test.cpk
It's very much possible I'm missing something but I don't see what.
Regards,
Gyan
_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
