utgg wrote: > Thanks - I hadn't spotted that there. > > Well... It is a genuine bug in the LMS patched faad (for windows)! > > As part of the machinations dealing with stdin input and looking for a > header, the code does a SEEK_END on the input file - which is expected > to fail. But it doesn't actually fail on windows - it seeks forward 4096 > bytes in the stream. This puts it into the middle of the aac data, > missing the first header. The code then searches for a header and for > this stream there happens to be something that looks like a header in > that data before the next real header, and decoding fails. > > This means that on the windows platform the faad transcoder will always > be missing the start of the stream - I'm surprised this hasn't been > noticed before. > > I've attached a new faad.exe which works. We really need Michael to > update the distributions 'officially' I guess. Great work utgg, thanks!
Are you saying there is something in the aac data of the "Desmond" programme that looks (to the parser at least) like a genuine header?! If so, that is bad luck indeed! ------------------------------------------------------------------------ Davidg1's Profile: http://forums.slimdevices.com/member.php?userid=35004 View this thread: http://forums.slimdevices.com/showthread.php?t=53229 _______________________________________________ plugins mailing list [email protected] http://lists.slimdevices.com/mailman/listinfo/plugins
