Bug#419890: include aoTuV patch

Rogério Brito rbrito at ime.usp.br
Sat Jul 7 10:00:56 UTC 2007


Hi, Fabio.

I'm including your detailed description of these improvements to
libvorbis so that people from the Debian Mentors mailing list may know
what this patch is about.

On Apr 18 2007, Fabio wrote:
> Would be nice to include aoTuV [1] patch to libvorbis. aoTuV is an
> improved libvorbis encoder, that, while keeping ABI compatibility,
> gives many improvement:
> 
> * better quality at all bitrates versus libvorbis-1.1.2 [2], giving
> aoTuV better quality than other codecs (AAC, MP3, MPC, WMA, ...) [3];
> * support of quality down to -2 (32kb/s at 44kHz stereo), versus -1
> (45kb/s at 44kHz stereo) of libvorbis-1.1.x;
> * encoding speed 10% better thanks to the included Vorbis-OptSort [4]
> patch (improve ordering loop).
(...)

I have already packaged a private version of libvorbis with the aotuv
patch (which applies cleanly) and I also fixed some lintian warnings.

If some mentors are interested in what I did so far, I put the sources
on my homepage:

http://www.ime.usp.br/~rbrito/libvorbis/libvorbis_1.1.2.dfsg-2.0+aotuv5.dsc

Please, note that they are almost in a point to be uploaded to, say,
experimental, but there are some cosmetic documentation facts that need
some more attention (any comments are *QUITE* welcome).

As Fabio already said, Aoyumi's patch to the libvorbis reference
implementation brings Vorbis to the state-of-the-art and with the
proliferation of Vorbis based players, it really shows how much this
*Free* format is capable of giving in terms of audio, even if the
xiph.org foundation doesn't seem to be updating it anymore (which is a
shame).

The updated libvorbis is all that is needed to have oggenc use the new
Vorbis engine beautifully.

I'm using libvorbis + aotuv at quality 8 to save 

> aoTuV versions have been tested and found to have a better quality by
> the audio community (especially on hydrogenaudio forum [5]) and this
> is the reccomended vorbis encoder [6].

Let me add here that the tests were ABX/HR tests.

The original references that Fabio cites are on his first e-mail at
<http://bugs.debian.org/419890>.


Regards, Rogério Brito.

-- 
Rogério Brito : rbrito at ime.usp.br : http://www.ime.usp.br/~rbrito
Homepage of the algorithms package : http://algorithms.berlios.de
Homepage on freshmeat:  http://freshmeat.net/projects/algorithms/




More information about the pkg-xiph-maint mailing list