[JDEV] [Fwd: Cool idea! (Jabphone ;-)]

[David Waite]

> -----Original Message-----
> From: jdev-admin at jabber.org [mailto:jdev-admin at jabber.org]On Behalf Of
> Max Horn

> Actually, the Frauenhoffer Institute has developed the whole MP3
> encding system. And encoders like Vorbis/Org all build on the ground
> work done by the Frauenhoffer Institute.

They only created parts of it - I thought for the longest time that they
only had a patent on 'their' P.A.M., but apparently they have patents on a
lot more.

>
> I don't want to start a flame war, but IMHO the MP3/Org/Vorbis etc.
> way of encoding data is not well understood by 99.9% of the people
> talking about it (I don't mean any of you, just speaking in general).
> E.g. I often meet disbelieve when I tell ppl that different encoders,
> all producing MP3 at the same bitrate, with same/similiar settings,
> nevertheless do *not* necesarilly produce same results. In fact, the
> quality widely differs, so does the speed of the encoding process.
> That is because the "psycho acustic models" used by the encoders are
> interchangeable; and it's these psycho acustic models that make a
> encoder good or bad. Thes P.A.M. have to be created in tedious and
> expensive tests with hundreds of test listeners etc. This is not
> something a normal small group of open source authors can do, you
> need *money* for that, for the equipment etc.
>

Right - and it is much harder to find a good document on the principles of
P.A.M's than just on encoding - because it was left optional, and because
companies consider their model to be extremely valuable I.P. Different
companies specialize in different areas as well, Fruenhoffer is directed
towards the low bitrate 'internet radio' market.

> I wonder what Vorbis does in this area... I'll investigate that.

If they don't use a simple algorithmic method? It would be hard since 99.99%
of Vorbis users wouldn't care enough about encoding to bother, but it would
be interesting if they could survey users through their own encodings on
which methods sound best.

-David Waite