C[omp]ute

From: "andrew cooke" <andrew@...>

Date: Mon, 19 Jun 2006 18:29:28 -0400 (CLT)

Just noticed these.  http://www.elac.com/en/products/imago/dml.html has a
description and images.

My first reaction was - that must sound terrible!  How can you get good
stereo imaging if the phase information at high frequencies is lost?  For
years people have been trying to build extremely rigid speaker cones to
avoid the kind of behaviour that is a "feature" here; the only advantage
seems to be wide dispersion.

But then I thought - maybe there's something really cool happening here. 
Some kind of mix between chaotic motion and synthetic apertures.  It
seemed crazy, but then some highish-end audio companies seemed to be using
the technology.  So I poked around some more.

The technologoy is patented (how can you patent a bad speaker?) by NXT -
http://www.nxtsound.com/index.php?id=476&no_cache=1

It was difficult to find neutral opinions.  Since the technology enables
thin flat and curved panels it is very "marketable".   It's difficult to
separate marketng blurb from "this really sounds good".

Eventually I found this review -
http://www.findarticles.com/p/articles/mi_zdext/is_200308/ai_ziff46255/print
- for computer speakers.  You wouldn't expect that to be a very demanding
application, but the review was surprisingly negative.

Next, a HIFI Choice review of Mission speakers.  Two reviews, in fact. 
Their site is dead, but Google has caches if you click through the first
two hits for http://www.google.cl/search?q=hifi.choice+mission+NXT  Both
reviews are negative.

Finally, here's a detailed argument on ribbon speakers -
http://www.svconline.com/mag/avinstall_turning_pro/index.html - which
says:

Electrostatics and planar magnetics (ribbons) are driven uniformly over
their entire diaphragms. This is the exact opposite to the distributed
mode ideal of flat NXT panels, which operate in random phase across their
surface. Cones, on the other hand, are uniform (pistonlike) up to the
midrange. It is generally conceded that ribbons and electrostatic panels
tend to sound better than cones, while the distributed mode panels sound
worse. However, in certain pro applications in which a uniform phase line
array's excellent pattern control is not needed, the nature of NXT panel
radiation (the “antiribbon”) is an advantage. The wide coverage and random
phase of a well-executed NXT design, such as the Armstrong i-Ceilings
panel, is an advantage in the struggle to achieve wide uniform coverage
with low ceilings as well as high acoustic levels before feedback. The
random phase nature of the panel helps prevent feedback, effectively
randomizing the phase of the signal. The Armstrong panels actually sound
quite decent, to boot.

So, no cool chaotic synthetic aperture...

Andrew