[FFmpeg-devel] PATCH: 5.1 AAC SMPTE channel order with libfaad

[Robert Swain]

On 5 April 2010 13:35, Jean-Yves Avenard <jyavenard at gmail.com> wrote:
> On 5 April 2010 19:51, Benjamin Larsson <banan at ludd.ltu.se> wrote:
>> That is the unneded hunk. Does the channel output configuration match
>> the native decoder after your patch ?
>
> yes it does.. It makes 5.1 decoded audio plays just fine in MythTV or
> mplayer; just like when using ffmpeg internal decoder.
>
> ?[ Front Left ] [ Front Right ] [ Center ] [ LFE ] [ Surround Left ] [
> Surround Right ]
>
> It takes whatever libfaad output and put it back in that order (from
> what I've seen, libfaad always output C L R SL SR LFE

What about non-5.1 configurations? I need to check the code for that
but don't have time right now.

> There's a few more features than parametric stereo that makes libfaad
> a requirement (I'm talking from a mythtv perspective here). First the
> LATM/AAC patch available for libfaad (but my understanding is the
> author is working to get something working with the native ffmpeg AAC
> decoder)

Indeed, this needs to be done.

> and H264 HE-AAC used by few european countries for broadcast, and
> likely going to be more and more used as countries are moving to the
> new MPEG-4 AVC standard

This should be covered when Parametric Stereo is completed.

> I am very puzzled by M?ns Rullg?rd message. I don't see where else
> channel re-ordering could be done than within the codec , it's how all
> other codecs do it.

Other codecs do it in-codec because they can set up an array of
pointers to be passed to our float to 16-bit int conversion routines
that will reorder channels and interleave samples on the fly. Generic
channel reordering for codecs that don't output in
waveformatextensible order, could be implemented outside the codec and
be shared between all such codecs. But maybe there's something to be
said for implementing it in a shared manner, but using it in affected
codecs so that they do all output the same channel order. I think such
a thing would simplify matters for anything after the decoder. I think
this is what M?ns was getting at.

Regards,
Rob