On Thu, Apr 19, 2018 at 12:35 AM, Michael Niedermayer <mich...@niedermayer.cc> wrote: > > This breaks fate: (have not tested anything beyond this) > > TEST adtstoasc_ticket3715 > --- ./tests/ref/fate/adtstoasc_ticket3715 2018-04-17 14:20:30.500366780 > +0200 > +++ tests/data/fate/adtstoasc_ticket3715 2018-04-18 23:34:00.751854625 > +0200 > @@ -92,4 +92,4 @@ > 0, 83968, 83968, 1024, 465, 0xeb3ce0af > 0, 84992, 84992, 1024, 326, 0x7be4a667 > 0, 86016, 86016, 1024, 339, 0x2cf4a71f > -0, 87040, 87040, 1028, 258, 0xd4007ad4 > +0, 87040, 87040, 1024, 258, 0xd4007ad4 > Test adtstoasc_ticket3715 failed. Look at > tests/data/fate/adtstoasc_ticket3715.err for details. > make: *** [fate-adtstoasc_ticket3715] Error 1 >
Just an off-hand comment that given that audio frames in AAC tend to have 1024 samples usually (or in more general, "1028 samples is quite unlikely for AAC"), I'd say 1024 is correct for a duration of an AAC frame. Looking at the hash, the actual data doesn't seem to have gotten changed. Best regards, Jan _______________________________________________ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
