[LAU] ambisonics: fons ambdec and muse.demon.co.uk ambidec

Jörn Nettingsmeier nettings at folkwang-hochschule.de
Sat Jan 8 00:06:02 UTC 2011


On 01/07/2011 11:35 PM, Niels Mayer wrote:
> On Fri, Jan 7, 2011 at 1:29 PM, Paul Davis<paul at linuxaudiosystems.com>  wrote:
>> ambisonics assumes a speaker in the midst of an array of speakers that
>> point more or less in the general direction of
>> the listener's head. this phone is just a gimmick.

if you're interested in beamforming with spherical speakers, check out 
franz zotter's and andy schmeder's papers. executive summary: it can be 
done,  but it requires quite some oomph if you want to cover more than 
just one or two octaves. implementing this on a cell phone will always 
remain a marketing gimmick without any real value.

> If that is true, then I guess the link I posted regarding
> CNMAT/MeyerSound "eigenspeaker" is bunk?

no, it's not. that's andy schmeder's project i was talking about. it 
does work, but it's not really about surround sound (unless you happen 
to know your room exactly, in which case you can bouce beams off of 
walls and get some sort of surround effect, but with huge artefacts and 
sidelobes).
these speakers are for acoustic measurements and simulation of radiation 
patterns. you could for instance sample a violin with a sphere of 
microphones, and use one of these spherical speakers to create something 
that looks like a violin and quacks like a violin.

> I'm talking about a different configuration of ambisonics than the
> traditional speakers around the listener and the associated sweet
> spot. IMHO that's a somewhat difficult configuration, especially for
> mobile apps, so I was just wondering whether the effect be gained with
> only four speakers in a "polar" configuration.

you can shoehorn ambisonics theory into this "inside out" application. 
dylan menzies (and maybe dave malham, too) has formulated this idea some 
years ago and called it "o-format". but it's not generally useful.

> And if not four, six, eight, or how many? Can they be arranged in a
> "circle" around the perimeter or is 360 degree configuration of
> drivers required?

you can do circular beamforming, but it will be even less effective than 
spherical. think huge amounts of destructive interference and ugly, ugly 
side lobes. for beamformed bass, you need ridiculously long coil travel 
and loooots of power.




More information about the Linux-audio-user mailing list