On 8/19/17, Nicolas George <geo...@nsup.org> wrote: > Le duodi 2 fructidor, an CCXXV, Paul B Mahol a écrit : >> That is not delay between blink and beep. But video timestamps. > > Since the beeps happen every second, the delay is the fractional part, > or the remainder of the frame count divided by the frame rate. But if > you think this arithmetic is too complex for our users, feel free to > propose an option to add it on testsrc2. > >> For lipsync one would need to alter so blinks happens at exact same >> time as beeps. > > They do. > >> Say delay is 10 frames, you modify filter to produce delayed blinks. > > I think it should be a job for setpts, but if you consider that too > complex, again, I am not against a time delay option to testsrc2. > >> > The beep is when the cross is fully contracted. >> That is far for being user friendly and obvious. > > It looks as obvious as the bumping dot on the videos you showed, in my > opinion. > > Anyway, I am not against you wasting your time adding yet another test > pattern, but not as proposed in this patch: use the framework of all the > test patterns and make it bit-exact.
I like idea of having single A/V filter to do this job, not combinations where users might become confused. Also this filter displays audio parameters in video output, which would become pain to do in combination of filters. > > Regards, > > -- > Nicolas George > _______________________________________________ > ffmpeg-devel mailing list > ffmpeg-devel@ffmpeg.org > http://ffmpeg.org/mailman/listinfo/ffmpeg-devel > _______________________________________________ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
