Bug#703732: faad fails to gaplessly decode itunes, nero, fraunhofer aac

Julian Hughes julianhughes at gmail.com
Mon Mar 25 12:31:52 UTC 2013


On Mon, 25 Mar 2013 09:47:36 +0100
Fabian Greffrath <fabian at greffrath.com> wrote:

> This commit refers to the frontend code and does not touch libfaad
> itself.
....
> 
> I am not that sure anymore. It seems that the library is fine to play
> gapless and that its the frontends that need to get patched in order
> to cope with this. But this is only a guess, we are still lacking a
> patch against libfaad to confirm this...
> 
>  - Fabian

Thanks for looking at the Rockbox implementation.

Meanwhile the "official" faad binary itself surely would not be 
one of the frontends which doesn't use libfaad properly?

I just built faad and libfaad from most recent source (now I see 
the bug tracker is at sourceforge, which is why I couldn't find 
it at audiocoding.com) and I get the same result:

I make gapless m4a with neroAacEnc:

to avoid any possible imperfect player's use of libfaad I decode
to wav and play the wavs a) with 'mplayer2 -gapless-audio' and b) on
my player running Rockbox:

decode with neroaacdec: the wavs play gaplessly.
decode with faad: there is a gap inserted.
decode with win32 application foobar2000: the wavs play gaplessly.

I repeat the exercise, this time making the encode with fdkaac instead
of nero:

The result is identical:

wavs produced by faad have a gap inserted.
wavs produced by <another_decoder> play as intended.

I'm still concluding that faad/libfaad can only correctly decode
gapless files if they are produced by faac, which doesn't seem to match
the package description of

"FAAD2 correctly decodes all MPEG-4 and MPEG-2 MAIN, LOW, LTP, LD and
ER object type AAC files."

I hope I haven't completely misunderstood this, and I appreciate that
any fix, if even possible, might not be very likely but thank you for
taking the trouble to look into it

regards

Julian



More information about the pkg-multimedia-maintainers mailing list