On Sun, 19 Feb 2006 21:03:16 -0000, "Serge Auckland"
wrote:


"Roy" roy wrote in message ...

"Serge Auckland" wrote in message
...
This is a long post, so those not interested can ignore it.

This brief history is not exhaustive, and covers mostly surround sound as
applied to records. There were separate developments for broadcasting,
and these have been mentioned in passing, but not covered in detail. I
have also not discussed Ambisonics in any detail.


I would have thought any history of surround sound for recording would
contain more than a passing mention of Ambisonics. It seems to me like the
Betamax of surround sound - technically superior but poorly marketed.

Roy.
Yes, I agree in part, but what I was trying to do was to explain the origins
of surround sound as it applied originally to vynil. I don't think that
Ambisonics was poorly marketed as much as just too late. By the time
ambisonics came about, the public was fed up of "quadrophonic" systems that
didn't work. Anyway, I question the whole premise of surround sound through
four (or five) loudspeakers. It relies on pair-wise phantom images which
just don't work in practice. As we know, frontal phantom images work quite
well, rear phantom images work after a fashion, but don't provide accurate
localisation, and sideways phantom images hardly form at all. For classical
music (which is the only format Nimbus has tried ambisonics, as far as I
know) it will work OK for ambiance, but not for remote soloists. 5.1
surround works for films with the distraction of pictures, but not terribly
well for music. Unless some sound-field synthesis system can be evolved that
doesn't require 200 'speakers (see my earlier posts on the subject) we're
stuck with pair-wise phantom images, and consequently ambisonics or
otherwise, poor surround sound.

S.



One of the problems for Ambisonics was that the patents were assigned
to the NRDC, a famously incompetent bureaucracy that killed quite a
few good ideas in the 70s.

The point of the ambisonic system is that it is a storage/transmission
system that is independent of the reproduction environment. The theory
is that three channels are all that is needed for accurate recording
of directional information in a circle, and that four channels can
completely encode directional information in a sphere. The way it is
reproduced is entirely dependent on the decoder and as many speakers
as needed can be used. The Meridian decoders allow 7 speakers, with
varying layouts, and others have designed decoders using many more
speakers than this.

The problem with the Nimbus and Unicorn recordings is that they had to
use the UHJ system to matrix into 2 channels and hence degraded the
encoding.

It is something of a joke that a system is available that will enable
accurate encoding of directional information into three channels but
the commercially sucessfull but inaccurate Dolby system uses 5
channels and is now starting to use 7! More than twice as much data
for a worse result.

Incidentally the VHS/beta comparison is not apt. There was very little
actual difference between the wo systems. A moe suitable comparison
would be U-matic to Digibeta.

Bill