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.





Brief History of Quadrophonics Surround Sound



Some time in the sixties, David Hafler proposed a way of extracting ambience
information from two channel stereo, by connecting a third loudspeaker
between the L and R live connections of an amplifier, and placing this third
loudspeaker behind the listener. This loudspeaker would therefore respond to
the L-R information and which allowed a degree of ambience to be heard. Out
of phase information, which would normally result in a reduction in stereo
localisation in the front pair is brought out to the back. Clearly, if a
signal is all left or all right, then there is also a considerable amount of
L-R information, so the Hafler surround "channel" carries all sorts of
information as well as the out-of-phase ambience. Nevertheless, on
orchestral music with a reverberant acoustic, some quite pleasant effects
can be heard.

By 1971, we began to see papers at Conferences, in the AES journals and
popular press describing ways of providing four channels of audio which
could be amplified discretely with four channels of amplification, and heard
over four loudspeakers arrange round the listener. There were many competing
systems propsed, but all accepted the commercial need to provide these four
channels from conventional vinyl LPs which could be compatible with
non-surround systems and be broadcastable on mono and stereo radio. At that
time, Reel-to-Reel was not considered acceptable for mass-market programme
distribution - EMI had tried to release stereo recordings on 3 ¾ ips tape,
but with little commercial success in spite of apparently excellent
technical quality. Cassettes were not at the stage of development that could
be called hi-fi, Dolby noise reduction was still uncommon, and the need for
compatibility (and Philips patents) stopped discrete 4 channel cassettes
being made. So, that left the LP. As mentioned, the imperative of
compatibility required that the four channel information be "coded" onto an
LP such that it could be played in stereo and mono by normal equipment.
There were many coding systems proposed, but 3-4 saw commercial release.
These were divided into two types, the matrix systems and the "discrete"
system(s)

4 channels could be matrixed into a two-channel signal using mathematical
matrixing, with or without phase shifts between channels. The main two
commercial systems were the CBS SQ system invented by Ben Bauer et. al. and
adopted by CBS, EMI and others, and the Sansui QS system adopted by Pye
records and others. It will come as no surprise that these two systems were
incompatible with each other.

The main attributes of the two systems we

SQ

Crosstalk

Lf-Rf - None

Lr-Rr - None

Lf-Lr - 3dB

Rf-Rr - 3dB

Diagonals - 3dB

There were two further alternative decoding matrixes that provided better
f-b and diagonal crosstalk at the expense of L-R crosstalk

Stereo compatibility: No L-R crosstalk, Lr and Rr collapse to front L and R

Mono compatibility: Centre rear sounds are suppressed.

If you look at what the playing stylus is doing when replaying an SQ record,
Lf and Rf are the conventional +-45 degree modulation, Lr and Rr are
clockwise and anticlockwise helical modulations respectively.

QS

Crosstalk

Lf-Rf - 3dB

Lr-Rr - 3dB

Lf-Lr - 3dB

Rf-Rr - 3dB

Diagonals no crosstalk

Stereo compatibility : Front crosstalk 7.7dB

Mono compatibility: Centre back signals reduced.

These basic attributes don't include the phase shifts and time delays
between channels, so the situation with crosstalk and compatibility is more
complicated than I have indicated here.





Pickup cartridges of the time were capable of 20-25dB of separation at mid
frequencies, so particularly the QS system and to some extent the SQ system
provided much less crosstalk than stereo. In an attempt to overcome the
problems inherent with matrixed systems, JVC launched the CD-4 system
(Compatible Discrete - 4) This system provided 4 discrete channels on an LP
using a modulation technique similar in principle to FM stereo. The Left
channel on the record carried Lf + Lr in the normal way. There was also a
supersonic carrier at 30kHz (+15khz-10kHz) which carried the Lf-Lr
information. The Right channel was similarly modulated for right signals.
Simple mixing would recreate the Lf,Lr,Rf,Rr channels with no more crosstalk
than normal for stereo. The penalty for this is that the cartridge had to
produce a flat output up to 45kHz to recover the f-b carrier. A special
stylus profile (Shibata) was developed to allow this. Whilst CD-4 worked
fine under laboratory conditions, once in the field, taking into account
record wear and normal domestic dirt, decoding the 4 channels was a very hit
and miss affair. Nevertheless, it showed that it was possible to get fully
compatible 4 channels off an LP.

In response to CD-4, matrix decoder manufacturers started to produce
decoders with logic steering which would adjust levels in the channels to
provide more perceived separation, but at the expense of unstable images.
One UK manufacturer (TATE) claimed 35dB separation between channels for
their SQ decoder. As can be expected, manufacturers stressed the separation
but failed to mention the wandering images.

A second carrier-based system (UD-4) was also launched in the USA, but never
made it to Europe.

Amplifier and decoder manufacturers started to produce multi-standard
decoders, with and without steering logic, and with different matrix
coefficients, phase shift and delays, in an attempt to make their products
"better". At the time there did not seem to be any great insistence from the
patent holders that decoder manufacturers keep to a fixed specification.



Whilst all this was going on, the BBC were being very cautious about 4
channel sound, stereo and mono compatibility was for them the most important
aspect. They decided that no commercial matrix system was sufficiently
compatible with stereo and mono, and that creating another sub-carrier on
the FM signal wasn't an option. To do so would have reduced the modulation
available for the mono signal still further (FM stereo had already reduced
the mono modulation by around 11%, and the stereo service area was already
much smaller than the mono area.) A further reduction in mono power would be
difficult and an even smaller service area for quadrophonic radio over
stereo wasn't considered acceptable.

In November 1974's Wireless World, Carey and Sager published proposals for
broadcasting three channel surround, with a fourth being used for height
information, but I have not come across any evidence that their proposals
were acted upon. By then, the public's perception of surround sound was four
loudspeakers arranged in a square or rectangle round the listener.

So, in the time-honoured way of the BBC, they decided to research their own
matrix system, and after several iterations, came up with Matrix H. This was
allied to the separate work done on ambisonics, which carried on into the CD
era in Nimbus records releases.



By the time the BBC had carried out service testing, the whole issue of
quadrophonics had pretty much gone away. The first appearance of
Quadrophonics (a dreadful Greek-Latin combination) in the Hi-Fi yearbook was
in 1973 with 41 products from 18 manufacturers. By 1979 Quadrophonics had
disappeared, and the entry for "Surround Sound" consisted of 4
manufacturers, all of which were offering "Surround Sound" synthesisers, not
proper decoders

So, the Quadrophonic era was just 6 years, from 1972/3 to 1978/9. What went
wrong? This is a personal view albeit from someone who lived through it.
Neither SQ nor QS provided the technical performance necessary for it to be
seen as an improvement over stereo. CD-4 could never have worked using the
LP as a carrier, and discrete 4 channel was never a cassette option, as
Philips's prime concern was to maintain compatibility. Add to that the
incompatibility of recordings, the practical difficulties and additional
costs of four loudspeakers and four channels of amplification, and
quadrophonics never stood a chance. In spite of extensive service testing of
Matrix H, the BBC never implemented a regular quadrophonic service.

The legacy of quadrophonics did live on though. Dolby Labs realised that the
problems of matrix quadrophonics were mostly to do with localisation and
incompatible standards. For music, the crosstalk issue and logic steering
moving the phantom images about were serious, but for films, where the
distraction caused by the pictures would allow less critical appraisals, a
matrix system could work. Dolby Surround was, as I understand it, loosely
based on the QS system. This was followed by Dolby Pro-Logic decoding which
provided the dynamic steering necessary to reduce the perceived crosstalk.
As long as the system was used for films, with the rear channels being used
for effects, it worked satisfactorily, much better than such a system would
have worked with music without pictures. However, relying on a phantom
centre-front image for dialogue was found to be very wearing, as the action
of the pro-logic resulted in dialogue wandering about the stage. With the
advent of digital audio, and data reduction techniques Dolby Labs brought
out Dolby Digital. DD provides a centre channel that would lock dialogue to
the screen, with a further four channels of surround, and a low frequency
effects channels for, well, low frequency effects. The DD system is also
usable for music, as Producers can decide whether to use the centre channel
or not, and the use of the LFE channel is also optional. DD however is not a
linear digital channel as it uses lossy data reduction techniques.
Nevertheless, the Public seems to have accepted DD as used in DVDs as a
valid carrier for films and music programme.



References and further reading:

B. Bauer, D Gavereaux, A. Gust J. Audio Eng. Soc. Vol 19 1971

P. Scheiber. Four Channels and Compatibility. J. Audio Eng. Soc. Vol 19 1971

G. Shorter, Wireless World Jan & Feb 1972

D. Aldous. Brief Survey of Four Channel Stereo Techniques. Hi-Fi Year Book
1973

B. Bauer et. al. Quadrophonic Matrix Perspective - Advances in SQ Encoding
and Decoding Technology. Paper read at 44th AES meeting Rotterdam 22nd Feb
1973.

D. Aldous. Wither Quadrophony. Hi-Fi Year Book 1974

Carey & Sager, Wireless World Nov 1974

G. Shorter. The Surround Sound Panorama. Hi-Fi Year Book 1975

D. Meares. Quadrophonic Broadcasting: Why Matrix H? Hi-Fi Year Book 1978





S.