[LAU] A new approach

Hermann Meyer brummer- at web.de
Fri Apr 1 19:00:35 UTC 2016



Am 01.04.2016 um 19:08 schrieb Fons Adriaensen:
> On Fri, Apr 01, 2016 at 10:06:00AM -0400, Paul Davis wrote:
>
>> Since our patent has now been granted, I'm able to disclose part of the
>> reason why Ardour sounds much better than any other DAW (and uses less
>> CPU). MOD deliberately attempts undecided bits, but we realized that we
>> could harness Heisenbergian dynamics to exploit these undecided bits to
>> provide unparallelled fidelity at the same time as parallel processing at
>> the quantum level.
>> ...
> As has happened many times in the history of science and technology,
> independent groups of researchers have again developed the same or
> similar ideas at roughly the same time. But it doesn't happen often
> that they are able to announce their results on exactly the same day!
>
> I can now to disclose a significant result obtained by the group of
> researchers at ERC of which I am a member. While others were looking
> at 'decidedness' our group has for the last year investigated the
> very much related concept of bit spin and its effects. Bit spin
> results in bits being not being exactly spherical, and this has a
> detrimental effect on the quality of Ambisonics audio reproduction.
>
> We found that loss of quality can be avoided by using a modified
> set of anti-oblate spherical harmonics. This are known in other
> branches of science as 'slurpian functions' as they can be used
> as well to model the deformation of liquid droplets in a gradient
> field. Our main result is that we derived closed expressions for
> the parmaters required to exactly compensate for (variable) bit
> spin.
>
> New versions of Ambdec, Tetraproc and the AMB panner plugins,
> using these results, will be made available later today.
>
> Ciao,
>

Bit spinning is what we use in the guitarix project as well, but, as we 
are "rockers" we use high speed spinners and cutting the bit corners 
with bladed wheels.
The resulting bits fits perfectly in any simple non-linear function 
without ever producing aliasing effects.

regards
hermann


More information about the Linux-audio-user mailing list