Hi On Tue, Oct 14, 2014 at 12:05:19AM +0000, Carl Eugen Hoyos wrote: > Michael Niedermayer <michaelni <at> gmx.at> writes: > > > > Attached patch fixes ticket #1614 for me (and > > > passes fate-aac). > > > > > > Please review, Carl Eugen > > > > i can still hear some rather loud artifacts at > > around 10sec in the file > > Yes, I have no idea how to fix this. > The artefacts do not appear with faad.
if it works in another decoder like faad then its probably possible to instrument the code with lots of "printf" and then just do a diff between our decoders output and faads to see where the differences start thats not fun work but as long as decoders are reasonably similar structured its quite monotone work. That is start adding printfs to some function like random pick, before the windowing&imdct to both decoders and maybe to the packet decode functions to print the input bits. If one kind matches and the other differs the issue must be in code between them if both match it must be in later stages also we have #define TRACE that dumps the bitstream nicely formated as it is being parsed. [...] -- Michael GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB Opposition brings concord. Out of discord comes the fairest harmony. -- Heraclitus
signature.asc
Description: Digital signature
_______________________________________________ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
