Uploaded by Hanson Tsai

History of Spatial Coding

HISTORY OF SPATIAL CODING
Mark F. Davis, AES Member
Dolby Laboratories, Inc., San Francisco, CA, USA
INTRODUCTION
The goal of spatial audio coding and
reproduction is to convey some measure of the dimensional aspects of a
sound event to the end listener, preferably in a practical and cost-effective
manner. The importance of the spatial
aspects of sound has been clear since
before the dawn of audio reproduction, as evidenced, for example, in the
design of concert halls. So it is not
surprising that concern for the accurate capture, transmission, and reproduction of the spatial components of a
sound event goes back almost to the
birth of the audio industry as we know
it today. Yet even now, completely accurate spatial reproduction remains
elusive, and is perhaps the largest single imperfection in modern audio
technology.
This article is a brief overview of
some of the notable efforts that have
been made in the name of spatial coding. Although some may tend to conceive of spatial coding as a recent technology—part of the digital audio
age—many of the concerns around spatial coding, and even some of the fundamental techniques employed, have
existed for many decades, and in some
cases have had to be recycled multiple
times before reaching viability.
The basic approaches to spatial audio
systems tend to fall into two groups:
the brute-force solution of trying to reproduce an approximation of the original soundfield and the psychoacoustic
approach of conveying the essential
spatial cues without actually recreating
the soundfield itself. Not all spatial audio systems fall cleanly into one of
these categories, and some are crude
enough to leave the question of categorization open to debate.
Certain elements do tend to be common to almost any spatial audio system.
554
These include means for capturing the
sound (for example, some sort of microphone array); means for combining
and/or encoding the raw signals into a
net composite signal; means for decoding at the reproduction site; and presentation means, such as a loudspeaker
array or headphones. A problem underlying most spatial audio systems is that
of trying to capture and reproduce a
three-dimensional wavefield using essentially zero-dimensional transducers
because microphones and loudspeakers
are basically point-in-space devices.
That is in contrast to, for example,
video processing, in which cameras and
display screens can capture and display
two-dimensional images. Happily, the
audio problem is made somewhat more
tractable by the ability of the human auditory system to create a three-dimensional sonic perception from just the
two ear signals.
Of course, any viable sound system
must be practical and cost effective,
with realistic hardware requirements
(for example, low MIPS for digital systems) and low data rate. Ideal content
preparation requirements should be reasonable without placing undue time or
equipment constraints on the producer.
To the extent possible, the system
should be tolerant of suboptimal playback configurations or listener positions.
A number of sometimes interdependent threads run through the history of
spatial audio. Much of the design effort to date has been devoted to
squeezing as many channels as possible into the available storage and transmission media. A lot of work has also
gone into exploring the most effective
uses, orientations, and positions of the
available microphones and loudspeakers. Another issue has been differing,
market-dependent focus. In the consumer domain, spatial sound has large-
ly meant stereo until recent years,
while for the most part the cinema industry went straight from mono to full
surround systems.
One final thread is the need to develop infrastructures to support a particular spatial audio configuration as a prerequisite for the viability of that
technology. In addition to content producers and end users, such infrastructures may include a multiplicity of
wired and wireless transmission media
and storage formats that have included
mechanical, magnetic and/or optical
disks, tape, and film.
THE DAWN OF AUDIO
To provide some initial perspective, it
is worth noting a few of the signatory
events that marked the dawn of audio
technology. The desire to be able to
capture and reproduce sound at will is
not a recent development. References
to audio systems of one sort or another
can be found in literature long before
such systems were possible. However,
the theoretical foundations for practical audio systems were not established
until the 1800s when pioneers such as
Faraday, Henry, and Ohm in the electrical sciences and Helmholtz, Tyndall, Lissajous, and others in acoustics
demystified and quantified these phenomena, providing a basis for physical
instrumentation.
Work done by these pioneers in turn
led to the first great audio invention in
1867, the string and tin can “telephone.”
The identity of the inventor appears to
be lost to history. Although this invention was primitive, it clearly demonstrated a strong intuitive understanding
of the nature and workings of sound.
A much more practical and farreaching invention followed nine years
later when Alexander Graham Bell inaugurated the telephone on March 10,
J. Audio Eng. Soc., Vol. 51, No. 6, 2003 June
History of Spatial Coding
J. Audio Eng. Soc., Vol. 51, No. 6, 2003 June
A. G. Bell
First telephone
transmitter
T. Watson
Photo courtesy of Oberlin Archives
1876 by speaking the immortal words
to his assistant, “Mr. Watson, come
here. I want you.”
Bell actually filed his patent application on February 14, 1876, just two
hours before electrician and inventor
Elisha Gray filed a caveat (intent to
patent) for a similar device.
The invention of the telephone was
significant not only for its intrinsic value, but also for establishing the principles of electroacoustic transduction in
both directions. A long succession of
refinements and improvements in microphones, headphones, loudspeakers,
and other transducers would follow.
With the transmission of sound established, it was only a little more than
a year later that Thomas Edison
demonstrated the ability to record and
reproduce sound with his recording on
December 8, 1877, of “Mary Had a
Little Lamb.”
Edison’s invention is elegant and almost starkly simple, employing only
mechanical transduction; in hindsight
it is surprising that it took mankind so
long to devise it.
Despite its simplicity, Edison’s initial phonograph employed separate
record and reproduce heads, so “offthe-tinfoil” monitoring has presumably
been a feature of recording devices
since day one.
In 1889 Edison followed up with the
invention of motion pictures, using
some new 35-mm film stock prepared
by George Eastman. The first film
showed Edison’s assistant, Fred Ott,
sneezing. Although initially not strictly
an audio invention, Edison always intended motion pictures to be accompanied by sound. It would be some years
before this became commercially viable. Edison was but one of several people pursuing the idea of motion pictures
during this period, and it appears that
the first motion picture on a strip of film
was produced a year earlier, in 1888, by
Auguste Louis Le Prince, who mysteriously disappeared shortly before his invention could be commercialized.
A number of other soon-to-be important formats got their start in this
period. In 1894 Guglielmo Marconi, at
the ripe old age of 20, invented wireless transmission. It would be some
years before full continuous-wave
transmission of audio was possible, ➥
Elisha Gray
Frame from Le Prince
1888 Film
Three photos of Thomas Edison
Edison’s first phonograph
Fred Ott sneezes in
Edison’s first film.
555
History of Spatial Coding
Marconi
but Marconi established the basic principles of wireless transmission.
On December 1, 1898, Valdemar
Poulsen became the first to patent a
magnetic recording device, using steel
wire as the recording medium. Poulson
was partly motivated by a desire to
avoid the wear of grooved media
played with a stylus.
deForest
Audion
Leon Gaumont, in 1901, began experimenting with optical sound on
film. This, too, was intended in part to
avoid the wear of grooved media, as
well as to provide reliably synchronized sound with video.
Rather amazingly, all the inventions
of this time period were realized without the aid of electronic amplification,
as it was not until 1906 that Lee deForest invented the “Audion” triode vacuum tube. Indeed, it would be another
10 years before the vacuum tube became a commercially viable mass-produced item and another decade or so
before it permeated various designs. It
was not until the late 1920s or early
556
Drawing of Clement Ader’s 1881 demonstration in Paris
1930s that basic monaural audio technology was refined to the point that it
made sense to start considering multichannel systems capable of conveying
spatial audio information.
THE DAWN OF SPATIAL AUDIO
Regardless of the time it would take
for single-channel audio to achieve initial maturity, early experiments
into spatial audio began soon after the development of the telephone. In 1880 Bell did some tentative binaural experiments using
a pair of telephone transmitters
connected to a pair of receivers.
The first widely noted public
demonstration of spatial audio
followed a year later in 1881
when Clement Ader set up a series of microphones across the
stage of the Paris Opera and fed
their outputs via wires to headphones in nearby hotel rooms.
Listeners perceived a crude but
effective binaural rendering of the performance,
noted for its compellingly
natural quality. Unfortunately, there appears to
have been little immediate follow-up interest.
The first crude system
for presenting spatial
sound to an audience was
patented on January 12,
1915 (Patent Number
1,124,580) by Edward H.
Amet. This system employed a monaural record
synchronized to a movie
projector. A mechanical
commutator attached to the turntable
allowed the sound to be panned to any
of a number of loudspeakers arrayed
across the screen, presumably allowing
spoken dialog to follow the position of
actors as they moved about. The system could also direct sounds to loudspeakers in the audience. This was a
remarkably far-sighted invention, as it
would be another dozen years before
even mono-synchronized sound was
commercially employed in the cinema.
At this point some additional formative events bear mentioning, although
none formally employed spatial sound
reproduction. On April 28, 1916, Edison conducted a convincing live-versus-recorded demonstration of his Diamond Disk acoustical phonograph in
Carnegie Hall, New York. Although
the phonograph was monophonic, the
hall imparted its exceptional spatial
acoustics to both the live performer
and the reproduced version. Perhaps if
the average living room had similar ➥
J. Audio Eng. Soc., Vol. 51, No. 6, 2003 June
Stereo
Mono
SRS Labs is a recognized leader in developing audio solutions for any application. Its diverse portfolio
of proprietary technologies includes mono and stereo enhancement, voice processing, multichannel
audio, headphones, and speaker design. • With over seventy patents, established platform partnerships
with analog and digital implementations, and hardware or software solutions, SRS Labs is the perfect
partner for companies reliant upon audio performance.
Multichannel
Product Applications
Technologies.
• Circle Surround II • TruBass
• Home Theater/Entertainment
• FOCUS
• TruSurround XT
• Wireless + Portable
• SRS 3D
• VIP
• Telecom + Voice
• SRS Headphone
• WOW
• Gaming
• Internet + Broadcast
World Wide Partners
Aiwa, AKM, Analog Devices, Broadcom, Cirrus Logic, ESS, Fujitsu, Funai,
Hitachi, Hughes Network Systems, Kenwood, Marantz, Microsoft,
Mitsubishi, Motorola, NJRC, Olympus, Philips, Pioneer, RCA, Samsung,
Sanyo, Sherwood, Sony, STMicroelectronics, Texas Instruments, Toshiba
The Future of Audio.
C 2002 SRS Labs, Inc. All rights reserved. The SRS logo is a registered trademark of SRS Labs, Inc.
Technical information and online demos
at www.srslabs.com
History of Spatial Coding
quality acoustics, the goal of viable
spatial sound reproduction would have
been reached much sooner.
On November 2, 1920, the first commercial radio station (KDKA in Pittsburgh) began broadcasting.
In the 1920s, Fox Newsreels used
optical sound on film to provide narration, and on October 6, 1927, synchronized dialog with film was commercially employed for the first time with
the release of The Jazz Singer by
Warner Bros. Although the soundtrack
was monophonic (a record), the stage
was set for the eventual advance to
surround sound with film.
With the arrival of the 1930s the foundation and infrastructure of single-channel audio was now established enough
to support more aggressive experiments
in spatial audio. During this period Bell
Labs took a pioneering role in advancing the state of the art, something they
would eventually be deterred from by
the U.S. Securities and Exchange Commission, which was afraid the company
would extend its monopoly on telephony to cover all of commercial audio. In
many of its experiments during this
time, Bell Labs was materially aided by
noted conductor Leopold Stokowski,
who became something of an evangelist
for spatial sound reproduction.
In 1931 Stokowski and Harvey
Fletcher of Bell Labs took up where
Clement Ader had left off, using im-
proved recording equipment to convey
monaural and binaural sound from the
Academy of Music in Philadelphia.
The following year Stokowski would
cut a crude stereo record using two
groove bands.
Before that occurred, however,
something extraordinary happened. On
December 14, 1931, British engineer
and inventor Alan Blumlein patented
stereo.
Blumlein was a prolific inventor
with 128 patents to his credit. His 1931
stereo patent didn’t just focus on one
aspect of stereo sound but examined
the entire audio infrastructure of the
time and determined in detail what it
would take to convert it to stereo.
Blumlein’s patent, number 394325,
contains 70 claims addressing such refinements as:
• Microphone design
• Microphone techniques, such as the
use of crossed figure-8 microphones
• Disk cutting, particularly the 45/45
stereo disk-cutting technique, in
which the left channel is recorded on
one groove wall and the right channel is recorded on the other groove
wall. This system would be “invented” at least two more times before
becoming a commercial reality.
• Stereo broadcast techniques including the use of AM and FM on a
common carrier to convey the two
channels
• Sum/difference matrixing, in part to
support backward compatibility with
mono equipment
• A stereo “shuffler” circuit. There is
some question about the purpose of
this circuit. At one point it is described as a preprocessor of stereo
signals, intended to make the acoustic waves reaching the listener more
nearly like a live situation, perhaps
foreshadowing Michael Gerzon’s
Ambisonics system. But at another
point, a shuffler is described as separately using interchannel amplitude
and phase to independently pan a
movie soundtrack left/right and
up/down behind a movie screen. If
this configuration had just been
flipped horizontally, so that its outputs were distributed around the
room instead of around the screen, it
might have comprised the first matrix
surround system.
One of the few areas not significantly addressed in Blumlein’s 1931 patent
was stereo sound on film, an omission
corrected four years later, in 1935,
when Blumlein produced the first films
with a stereo optical soundtrack, most
notably short films of trains at Hayes
Station. Such was the nature of Blumlein’s foresight that these films can still
be played on current 35-mm projectors
with stereo sound heads.
Unfortunately, Blumlein’s work was
at least 20 years ahead of its time. His
Used by permission of Robert Alexander
Illustration of 45/45 recording heads from Blumlein’s 1931 patent.
Alan Blumlein, as shown on
cover of Robert Alexander’s
biography (recommended
reading)
558
Detail from Patent No. 394,325
(1931) demonstrating recorder
assembly whereby both channels
may be cut by single tool on same
groove. This results in recording at
45o to the wax (or other) surface
giving sum and difference.
Detail from Patent No. 394,325
(1931) similar to figure on left except
here the driving force is generated
from electromagnet
J. Audio Eng. Soc., Vol. 51, No. 6, 2003 June
History of Spatial Coding
colleagues apparently regarded his
work as fascinating but with uncertain
commercial potential. EMI, his employer, rewarded him by transferring
him out of the audio group to pursue
more pressing projects like television
and radar. He died shortly before his
39th birthday in an airplane crash during a test of an experimental radar system. Sadly, during an earlier attempt to
produce a biography, Blumlein’s notebooks were loaned out and apparently
lost. EMI never attempted to capitalize
on Blumlein’s stereo patent before it
expired.
With Blumlein’s work remaining
largely unknown, work by other
groups continued as before. In 1933
Bell Labs’ researchers Fletcher,
Snow, and Steinberg attempted to
widen the sweet spot of stereo by
adding a third channel in the center.
They conducted a well-received
demonstration with a three-channel
wire transmission from the National
Academy of Sciences to Constitution
Hall in Washington, D.C.
That same year binaural recordings
made using a dummy head were
demonstrated at the Chicago Century
of Progress Exhibition.
In 1936 Arthur Keller of Bell Labs
reinvented the 45/45 stereo disk recording system, patent number 2,114,471.
The filing was delayed because Bell
Labs felt there was no commercial ap-
Philips-Miller Recording
2-Channel PhilipsMiller Recording
J. Audio Eng. Soc., Vol. 51, No. 6, 2003 June
Leopold Stokowski
plication. Keller was unaware of Blumlein’s work until the 1950s, when
Westrex Corporation invented 45/45
recording for the third time.
1936 was also the year the PhilipsMiller optical film sound system was
introduced. This system used an
opaque strip on film that was selectively sliced away by a vibrating stylus,
leaving a variable-area soundtrack that
could be immediately played with a
conventional photocell arrangement.
The system was capable of recording
two-channel sound and had very high
quality for its day.
1940 was a watershed year for spatial
audio. On April 9 and 10, Fletcher and
Stokowski demonstrated a novel threechannel recording system at Carnegie
Hall. This system used optical sound on
a film track that was running at high
enough speed to support a frequency
response of 30 to 15000 Hz. A fourth
optical track contained a gain control
signal, effectively implementing a
wideband compander, resulting in an
impressive system dynamic range of
120 dB.
Without question, however, the big
news in the world of spatial audio in
1940 was the release of the film Fantasia, with music conducted by Leopold
Stokowski. Road-show presentations of
this film featured a full surround sound
system called “Fantasound,” developed
by Stokowski, Disney, and RCA.
The Fantasound system also used a
four-channel optical soundtrack on a
separate strip of film, synchronized
with the projected film. There were
three audio channels plus a control
track. The control system allowed the
audio tracks to be panned to any of 10
loudspeakers: Front Left/Center/Right;
Front Left/Right corners; two side
loudspeakers; two rear loudspeakers;
and one on the ceiling. The side and
rear loudspeakers were used sparingly,
except in the “Ave Maria” segment,
where they provided an enhanced sense
of envelopment.
Unfortunately there was little subsequent interest in Fantasound, and the
equipment was donated to a Russian
company. Tragically, it was lost at sea
during shipping.
Audio progress in the next few years
was understandably slowed, as the
world took time out to fight a war. German researchers made significant refinements to magnetic recording during
this period, and in 1942 the first stereo
tape recordings were produced by Helmut Kruger at German Radio in Berlin.
As the war wound down, British
record company Decca began experimenting with lateral/vertical stereo
records, but the heavy-tracking pickups of the era made the records difficult to play, and it would be another
dozen years before stereo records became a commercial reality.
In 1948 three milestones occurred
that, while not directly related to spatial audio, helped provide the foundation for further progress. One of these
was the development of the 33-1/3
RPM LP record by Peter Goldmark of
CBS Laboratories. Similar formats had
been used in broadcast applications
(and Muzak) for some time, but its use
as a consumer format was revolutionary and led to the introduction of the
stereo LP nine years later.
The second event of note to occur in
1948 was the formation of the Audio
Engineering Society in New York.
On June 30, 1948, the transistor was
invented by William Shockley at Bell
Labs. In the following years transistors
would almost entirely replace vacuum
tubes in analog circuits, perhaps to the
consternation of listeners enamored of
“tube sound,” and some 20 years later
latter-day descendents of the transistor
would be a cornerstone of the digital
revolution in audio technology.
THE 1950S: STEREO COMES OF
AGE
Probably the biggest audio news of the
1950s was that stereo finally came of
age as a consumer medium. On the ➥
559
History of Spatial Coding
cinema side, there were some
notable experimental forays into
multichannel sound systems, although mainstream cinema surround sound was still a couple
of decades away.
The arrival of stereo disk
recordings may have seemed imminent in 1951, when Emory
Cook made a series of stereo
recordings of railroad trains, realized as a dual-band LP titled Rail
Dynamics, which was demonstrated at the Audio Fair in New
York. However, the dual-band
technique required a pair of precisely mounted mono pickups
and had limited playing time,
thus it did not achieve commercial acceptance.
In the meantime the cinema
industry experimented with
some truly memorable surround
sound systems, although none
would enjoy an extended life
span. On September 30, 1952,
the Cinerama process was introduced with the release of the
film, This Is Cinerama. This amazing
system, developed by Fred Waller, employed three synchronized projectors,
each covering one-third of the screen.
The sound system, developed by Hazard E. Reeves, was based around a
synchronized dubber running a sevenchannel magnetic soundtrack (six audio tracks, one control track). The
loudspeaker system included five
screen loudspeakers (predating SDDS)
and an array of surround loudspeakers
that could be fed a mixture of the
source channels. The effect produced
by this system was almost visceral,
particularly during the roller coaster
sequence of This Is Cinerama. It is
worth the effort to seek one of the few
remaining theaters capable of showing
this film in its original format. Unfortunately, the expense of the Cinerama
process was prohibitive, and the system was abandoned in 1963.
Cinemascope, a system somewhat
less complicated than Cinerama, was
introduced on September 16, 1953,
with the release of the film The Robe.
The system used an anamorphic (horizontally magnifying) lens to project a
wide image with a single 35-mm projector. The sound system, developed
560
Diagram of Cinerama presentation
by Ampex, used four magnetic tracks
striped onto the print, intended for L,
C, and R front, plus a mono effects
track. The Robe is credited as the first
film to use off-screen voices. The high
quality of the Cinemascope sound system allowed films to be presented
without the dreaded Academy Mono
equalization, a high-frequency rolloff
used in theaters to minimize the hiss
associated with optical soundtracks of
the period. Although several dozen
films were made with the full Cinemascope system, it also proved too expensive in practice, and its use of magnetic surround tracks was eventually
discontinued. Magnetic tracks were
also used on 70-mm prints, usually
comprising five screen channels and a
mono effects channel, beginning in
1955 with the film Oklahoma and 1956
with Around the World in Eighty Days.
One more cinema surround system
notable for its practicality was the Perspecta Sound system, developed in
1954 by recording engineer C. Robert
Fine, also remembered for his “PingPong” stereo recordings on Command
Records some years later. Perspecta
Sound used the existing mono optical
sound track, to which was added three
low-frequency control tones used
by an external controller to pan
the mono sound to three screen
loudspeakers. Somewhat reminiscent of Amet’s 1915 switched
system, Perspecta Sound allowed
dialog tracks to follow the actors
on the screen using a single inventory print, and improved on
Amet’s switched system by allowing different levels to be fed
simultaneously to different loudspeakers. The slow time constants of the system resulted in a
type of pretransient artifact, as
the gain of a channel had to be
ramped up before the transient
arrived. Although the prenoise
and the lack of true multichannel
sound led to the system’s eventual retirement, it was used on
many Merry Melodies cartoons.
For Todd-AO 35-mm magneticstripe versions of Around the
World in Eighty Days, Perspecta
Sound was used to expand the
mono effects track to three surround loudspeakers.
Most of the multichannel cinema
systems of this era were used primarily
for spot sound effects, like airplane flyovers, rather than for true surround
sound. Ultimately the film industry
would require a high-quality, low-cost,
optical surround soundtrack to replace
mono as the standard sound format,
and that would take at least another
couple of decades.
While film sound engineers
scratched their collective heads over
the lessons learned from these pioneering systems, stereo sound was gearing
up to make its grand entrance as a consumer format. The next volley was delivered by Murray Crosby, who
demonstrated a compatible stereo FM
system in 1954. The performance of
this system so impressed the engineers
at RCA that, although they did not
commit to using it, they did initiate the
release of stereo material on 30-ips
two-track tapes. The first of these was
a recording of Also Sprach Zarathustra
recorded by Fritz Reiner and the
Chicago Symphony, which sold for
$18.95.
1957 was to be the breakthrough
year for consumer stereo. By this time
Arthur Haddy and associates at British
J. Audio Eng. Soc., Vol. 51, No. 6, 2003 June
History of Spatial Coding
Decca Records had significantly refined their vertical/lateral stereo disk
system and demonstrated it to RCA.
But on September 5, 1957, Westrex
started giving demonstrations of its
newly reinvented 45/45 stereo disk,
some 26 years A.B. (after Blumlein).
On October 11, 1957, Westrex demonstrated this system to some acclaim at
the New York AES Convention. And
in November 1957 Audio Fidelity
Records copped the honor of releasing
the first disk in this format, a curious
pairing of The Dukes of Dixieland
with Railroad Sounds, produced by
Sidney Frey. The story, perhaps apocryphal, is that Frey sent the master
tapes to Westrex to create a pair of
evaluation disks, and to discourage
their release Westrex cut one side of
each, but Frey went ahead and released
the record anyway. In short order the
RIAA gave its official stamp of approval to the system, including the use
of a 0.7 to 1.0 mil stylus and 6 grams
vertical tracking force, and the rest was
history.
THE 1960S: STEREO
INFRASTRUCTURE
AND RESEARCH
With the rapid acceptance of the stereo
LP disk, the decade of the 1960s saw
the development of complementary
technology that greatly expanded the
use of stereo in consumer audio.
In 1961 stereo radio was born (at
least in the U.S.) when the FCC selected the GE/Zenith FM multiplex system, largely designed by Carl Eilers, to
support national stereo broadcasts.
Stereo broadcasts began on June 1,
1961, by WEFM in Chicago and
WGFM in Schenectady. The system
chosen employs sum/difference matrixing for compatibility with mono receivers and conveys the difference information on an amplitude-modulated
38-kHz subcarrier. The FCC’s decision
was somewhat controversial, as the
system chosen exhibits 20 dB less dynamic range than mono transmission,
or that of the competing Crosby system, but allows spectral space for
background music services. The FCC
also withheld stereo certification of
AM broadcast at the time in order to
promote the struggling FM medium.
The fate of AM stereo was ultimately
J. Audio Eng. Soc., Vol. 51, No. 6, 2003 June
Arthur Haddy
left to the marketplace, where confusion and apathy have largely relegated
it to limbo.
The stereo infrastructure was extended with the introduction of two new
consumer tape formats principally configured for stereo. In 1963 Philips introduced the compact cassette, followed in 1966 by the introduction of
the eight-track cartridge format. Acceptance of the cassette as a high-fidelity medium was aided in 1969 with
the advent of the Dolby B-type noise
reduction system.
The establishment of these formats
motivated numerous investigations of
preferred recording and reproduction
arrangements, with the sometimes-conflicting goals of preserving both image
position and ambience.
In 1959 Goldmark and Hollywood of
CBS Labs found that using two small
loudspeakers with a subwoofer produced imaging little different from that
produced by a pair of full-range loudspeakers. Thus was born the satellite
loudspeaker system, although it would
be a while before it became popular.
The following year Ben Bauer of
CBS Labs demonstrated that crossfiring the loudspeakers somewhat could
widen the sweet spot in stereo audition. This principle has since been used
in a number of commercial designs.
That same year Beaubien and Moore
investigated stereo virtual imaging as a
function of frequency and concluded
that the fusion of such images could be
improved by panning low-frequency
and high-frequency information along
separate trajectories. This notion has
since been investigated for surround
systems as well. However, the conventional wisdom regarding widening the
sweet spot while preserving image fu-
sion is that the most robust solution is
to use additional loudspeakers, especially a center loudspeaker for stereo.
In a somewhat different vein, in
1963, Schroeder and Atal developed the
acoustic crossfeed cancellation circuit.
Although very sweet-spot dependent,
this system eliminates the infamous 2kHz notch encountered with centered
virtual images and allows the rendering
of images outside the span of the loudspeakers. Although initially intended as
a tool for studying concert hall acoustics, it has become the basis of many
modern “virtual surround” systems.
QUAD: THE BIG NEWS OF THE
1970S
The 1970s saw the start of two significant trends in audio. Unquestionably,
the higher public profile of these at the
time was the rise, fall, and eventual rebirth of four-channel audio, otherwise
known as quadraphonic sound, or
“quad.”
Quad began innocently enough as a
series of test recordings produced and
demonstrated by Robert Berkovitz at
Acoustic Research Corporation. The
stated intent was to simply show what
happens when a pair of rear channels
was added to a front stereo pair, but the
idea rapidly spiraled into a massive
commercial enterprise. To some extent
this was technically justified, because
quad avoided the compromise in stereo
between precise imaging and enveloping ambience, with the front channels
providing the imaging and the
surround loudspeakers providing the
ambience. But part of the underlying
motivation was almost certainly commercial one-upmanship, arguably resulting in products and systems being
rushed into the marketplace prematurely. In any case, audio engineers who
had worked so long to push two channels through a variety of broadcast and
storage media were suddenly faced
with trying to squeeze another two
channels through the same media.
The proposed systems included discrete tape systems, a discrete multiplexed LP, and a discrete quad FM
system, Quadcast, proposed by Lou
Dorren. But the primary entries were
matrix systems designed to extract four
channels from two. The godfather of
quad matrix systems was Peter ➥
561
History of Spatial Coding
Scheiber, who established most of the
basic principles of matrix systems then
and since. Other matrix systems introduced on the heels of Scheiber’s included those by David Hafler, ElectroVoice, the CBS SQ system, and the
Sansui QS system. These systems all
used a square loudspeaker arrangement. One notable variant was a system described in 1971 by Shiga,
Okamoto, and Cooper that used a diamond-pattern matrix: center front, left,
right, and a mono surround channel,
presaging the later Dolby MP matrix
system.
Each of these matrix systems employed an encoder to reduce four
source channels to two. Usually the encoder was a passive downmixer employing signed weighted sums of the
source channels, sometimes including
phase shift networks.
The decoders, most of which were
originally passive, then employed
complementary equations to upmix the
two encoded channels back to four.
Unfortunately, matrix systems involved some serious compromises that
limited their effectiveness, most notably minimal adjacent channel separation, which was usually only about 3
dB. In an attempt to overcome this limitation, designers added gain-riding
logic subsystems to better isolate dominant signals, but these were only partially effective and sometimes introduced objectionable gain-pumping
artifacts. The matrix systems also tended to be unduly sensitive to interchannel phase and amplitude differences,
such as might be encountered in a
slightly misaligned tape deck. With
some tweaking some of the surviving
quad recordings of the 1970s can be
played to good effect with current-generation matrix decoders.
One matrix system that stood apart
from the rest was the Ambisonics systems originally invented by Michael
Gerzon in 1969. Rather than trying to
recover discrete loudspeaker signals,
the goal of Ambisonics is to recreate a
valid approximation of the original
soundfield in the neighborhood of the
listener’s ears. Within limits, the system is compatible with both upmixing
and downmixing, and normally does
not use any active components, like
gain riding. Properly set up, Ambison562
1970s CBS SQ Matrix Decoder
ics is one of the most compelling surround systems ever devised, but as it
uses wavefront synthesis with a small
number of loudspeakers, it tends to be
sweet-spot dependent.
Recognizing the limitations of matrix
systems, several companies chose to
develop more discrete systems. The
highest profile of these was the
RCA/JVC CD-4 Quadradisc system,
which used sum/difference matrixing to
carry composite stereo baseband signals plus stereo difference signals conveyed by a subcarrier centered at 30
kHz. Unfortunately, tracking such
high-frequency signals without significant groove wear was largely beyond
the capabilities of phono pickups—
even those equipped with the Shibata
stylus developed for this application—
and the system was not widely adopted.
From a technical perspective, probably the most effective of the quad formats were the discrete tape systems.
Quad versions of eight-track cartridges
and reel-to-reel tape decks were commercialized, and there was at least one
quad cassette format proposed. Given
the weaknesses in the rest of the quad
infrastructure and the general unpopu-
larity of the eight-track format for
high-fidelity applications, it is not surprising that these formats failed to
achieve lasting consumer acceptance.
Just as the wide deployment of
stereo had motivated research into preferred modes of recording and reproduction, so did the quad era boost audio exploration. During the 1970s and
1980s the role of technical evangelist
that had been occupied earlier by
Leopold Stokowski was effectively assumed by musician Quincy Jones.
One of the more compelling demonstrations of surround sound ever
mounted was implemented in 1970 by
Tom Horrall of Bolt, Beranek, and
Newman (now Acentech). Horrall constructed a model of Boston Symphony
Hall and used it to calculate an approximation of the spatial impulse response
of the hall. He then recreated that response using an elaborate tape delay
system with multiple playback heads,
standard stereo material, and a dozen
loudspeakers strategically arranged
around the listener. The amazing thing
about this system was not just that it
created a highly satisfactory sense of
envelopment, but that it actually
sounded identifiably like Boston Symphony Hall. Few commercial spatial
sound systems of the era were able to
make such a claim. Horrall’s system
was subsequently commercialized in
1988 as the Pioneer DSP-3000.
Another interesting surround sound
investigation was published in October
1971 by Nakayama et al., in which the
effectiveness of a number of horizontally arrayed loudspeaker arrangements
was examined. Their preferred ar-
Discrete stereo groove with frequency distribution
J. Audio Eng. Soc., Vol. 51, No. 6, 2003 June
History of Spatial Coding
max VCR. A year later in 1976 JVC
introduced the VHS format VCR, originally with a monophonic linear analog
soundtrack, then stereo linear analog
tracks in 1980, and in due course
(1983), with a separate companded FM
stereo track, VHS Hi-Fi. These stereo
formats provided two channels of sufficient quality to support Pro Logic
surround in the home. The Laserdisc
format was developed in 1978 and formally introduced in 1980.
As had been the case with audioonly content a couple of decades earlier, the availability of recorded video
content with stereo audio tracks resulted in marketplace pressure for broadcast to follow suit, and in 1984 the
Horrall Surround Test
BTSC stereo sound system was introduced for commercial U.S. broadcast,
using sum/difference matrixing for
backward compatibility and a subcarrier modulation scheme developed by
Carl Eilers at Zenith. Taking a cue
from the impaired-SNR lessons
learned with FM stereo, a dual- ➥
Photo used by permission of Acentech Inc.
with a diamond pattern of
output channels, to supply the needed center
loudspeaker, feeding the
mono surround signal to
pairs of loudspeaker arrays along either wall.
Nonetheless, the indusQuad 8-track deck
try’s initial response was
somewhat tentative until
the following year, 1977,
Quad 8-track cart
when the release of Star
Wars, followed by the rerangement consisted of four front loudlease some months later of Close Enspeakers plus two at the sides and two
counters of the Third Kind, demonstratin the rear. A similar arrangement was
ed the system’s potential. In 1978 the
espoused by Theile in 1990. Within a
basic matrix was refined by Craig Todd
four-channel constraint, Nakayama’s
of Dolby to improve separation (the
preferred arrangement was two medi“motion picture,” or MP
um-spaced front loudspeakers and two
matrix, which included
wide-spaced side loudspeakers. These
phase shifters), and in 1982
and other studies consistently support
and 1983 the gain riding in
the current standard of 5.1 channels as
consumer decoders was upan effective surround compromise.
dated with the Pro Logic alWhile the consumer domain was
gorithm developed by Doustruggling with the difficulties of
glas Mandell, which is still
adapting stereo media to quad, more
in use. This system, along
substantive progress was being made
with the gain-adjusting cirin cinema sound systems. In 1974 Dolcuitry used in Dolby cineby Laboratories at last made stereo opma matrix systems, uses a
tical prints viable by applying its Aconstant-power dynamic
type noise reduction, including a
matrix, instead of variable
derived center channel with gain riding
gain applied to a fixed
system, to the film process. Lacking
matrix, to reduce pumping
surround channels, the system met
artifacts.
with muted response. Two years later
The sustained establishment at last
in 1976 Dolby corrected this problem
of surround sound in cinema would
with the addition of a four-channel maprove to be the gateway through which
trix system, used initially on the film
consumer audio evolved from stereo to
A Star Is Born.
surround. In 1975 foundation of the
This system initially used what was
consumer video revolution was begun
basically a modified Sansui QS matrix,
with the introduction of the Sony Beta-
Teac Quad deck
J. Audio Eng. Soc., Vol. 51, No. 6, 2003 June
563
History of Spatial Coding
band noise-reduction system developed at dbx, Inc. by the author was included in the system. Other countries
have used similar analog modulation
schemes for stereo TV, or gone directly from analog mono to digital stereo,
as Britain and other European countries did in 1986 with the Nicam-728
(Near Instantaneous Companded Audio Multiplex) system, a 14-bit to 10bit block compander.
In 1978 Emil Torick attempted to
apply the use of noise reduction to
U.S. FM stereo in a backward-compatible manner, using a novel companded
quadrature subcarrier, but the system
was mired in controversy about its susceptibility to multipath conditions. Although it enjoyed some limited use, it
was abandoned a few years later.
THE DAWN OF DIGITAL AUDIO
While quadraphonic sound was getting
the lion’s share of attention in audio
during the 1970s, another revolution
was quietly getting started: digital audio. As usual, the establishment of this
technology was predicated on the development of certain other technologies. One of these was the transistor, as
the use of large numbers of vacuum
tubes for logic circuits would have
been impractical on a large scale. But
even using large numbers of discrete
transistors would not have been sufficient; it took the development of digital integrated circuits, affordable highprecision
analog/digital/analog
converters, and digital storage media to
provide the necessary basis for digital
audio. Once these were in place, their
effects on the audio engineering community would be far reaching, leading
many audio design engineers to learn
to do in DSP (digital signal processing)
what they already knew how to do
with analog circuits. Eventually they
would begin to exploit capabilities of
DSP that were beyond the practical
abilities of analog electronics.
The availability of digital audio
components began to reach critical
mass around 1969, when Thomas
Stockham began experimenting with
digital recording, using a modified
Hewlett-Packard data recorder.
Stockham’s demonstration of digital
recording at the November 1976
New York AES Convention was one
564
of the signatory moments of the organization. A few years later he produced the first commercial digital
recording at the Santa Fe Opera.
In 1970 one of the first commercial
digital products was released, the Lexicon Delta-T101 delay line, designed
by Francis Lee. It used companded 12bit PCM.
In 1975 EMT released the first digital
reverb unit, designed by Barry Blesser,
which dramatically advanced the state
of the art of DSP. Its debut at the New
York AES Convention was another of
the standout moments of the Society.
Extension of digital recordings to the
consumer domain followed with
demonstrations of the audio compact
disc in 1981, released commercially
the following year. The CD had a profound influence on the audio engineering community, as it effectively eliminated in one stroke a whole raft of
traditional problems such as flutter,
wow, speed error, clicks, pops, noise,
and distortion.
5.1-CHANNEL SURROUND
Despite the acceptance of Dolby’s
four-channel matrix surround system,
there was clearly pressure to advance
to a more discrete system with two
surround channels and a subwoofer
channel, an arrangement first characterized by Tom Holman of Lucasfilm
as a 5.1-channel system, and standardized in June 1991 in Ottawa as ITU-R
Rec. BS.775-1.
In 1977 close on the heels of the Star
Wars matrix surround release, John
Mosely proposed a hybrid “quintaphonic” sound system, which employed a combination of discrete and
matrix-derived channels, qualifying it
as another system to anticipate the current Dolby Digital Surround EX system in operation of a matrix-derived
channel.
In 1978 Superman became the first
movie with a 5.1-channel soundtrack,
with 70-mm prints using a Dolby sixtrack magnetic format developed by
Ioan Allen that combined two surround
channels with two front LFE channels.
In 1980 Terry Beard and Nuoptix
Corporation developed a discrete fourchannel analog optical soundtrack.
Since each track was only 11.5 mils
wide, about one-third that of a track in a
stereo optical print, noise was a major
concern, and a wideband compander for
each channel was included. The system
was used for the movie Popeye.
In 1982 the film Return of the Jedi
was released and Tom Holman designed his THX surround systems. Separate systems were designed for home
and theater use, with common goals.
The THX theater system employed
constant-directivity horns, a novel
fourth-order Linkwitz-Riley crossover,
and a new theater alignment program to
assure consistency of sound quality
from theater to theater. For the home
system, a modified X-curve EQ was
used with dipole surround loudspeakers
and a surround decorrelator to provide
surround envelopment with direct
frontal imaging.
The use of Dolby Pro Logic matrixing with loudspeaker systems consisting of Left/Center/Right/Surround Array/and optical bass extension
continued as the standard surround arrangement until the early 1990s, when
digital soundtracks were developed for
35-mm theater film systems.
The first public presentation of a
film with a digital soundtrack was performed by Boston audio engineer John
Allen, who in 1989 fittingly ran a print
of Fantasia synchronized to a stereo
digital soundtrack on videotape, played
with a Sony PCM-F1 processor, and
expanded to four channels with an MPmatrix decoder.
The honor of producing the first optical digital soundtrack system went to
the combined forces of Kodak and Optical Radiation Corporation, who in
1990 introduced the Cinema Digital
Soundtrack (CDS) system with the release of the film Dick Tracy. This system used 5.1 channels of 12-bit ADPCM, with the digital data recorded
optically on the film in place of the
normal analog stereo optical track. Unfortunately this substitution required
double inventory prints with separate
analog or digital soundtracks, and
meant that there was no backup analog
track if there was a problem with the
digital presentation, and the system
was eventually withdrawn.
In 1992 Dolby introduced its SR-D
digital soundtrack system with the
films Star Trek VI and Batman
Returns. This system retained the anaJ. Audio Eng. Soc., Vol. 51, No. 6, 2003 June
History of Spatial Coding
log soundtrack, with Dolby SR noise
reduction and placed the digital information in the unused space between
the sprocket holes on one edge of the
film, thereby avoiding the double-inventory restrictions of the CDS system.
In order to stay within the data-rate restrictions imposed by the sprocket-hole
placement, this system employed a
more aggressive multichannel coder
that processed the channels as an ensemble, rather than coding each channel individually. Later that year DTS
introduced its digital soundtrack system with the films Dr. Giggles and
Jurassic Park (1993), which featured
audio data recorded on external CDROMs that were played in sync with
the film. In 1993 Sony introduced the
SDDS 7.1-channel soundtrack system
with the release of the film The Last
Action Hero. The extra two channels
were used to feed a total of five screen
loudspeakers. This system recorded the
data on the two outer edges of the film,
outside the sprocket holes, and used
the Sony ATRAC coder. As of this
writing, all three of these systems are
still in use, and in fact all three can coexist with each other and with the analog soundtrack on a single 35-mm
movie print.
With the desire for more audio channels came the need to improve the storage and transmission bandwidth of
available media. The 1990s saw the development of a number of multichannel
coders, including Dolby Digital (AC3), Musicam surround, AAC, WMA,
MLP, and MPAC. Some of the techniques employed to exploit the presence of multiple channels include use
of a common bit pool for all channels,
interchannel masking, interchannel prediction, and coupling or intensity
stereo. One previously tried-and-true
technique that has been problematic in
multichannel coders is sum/difference
matrixing. The problem with this approach in the context of low bit-rate
perceptual audio coders is that the dematrixing operation tends to direct the
signal and the quantization noise to different output channels, leading to directional unmasking of the noise.
In 1993 the 5.1-channel infrastructure was extended to the domain of
television broadcasting with the selection by the ATSC of the Dolby Digital
J. Audio Eng. Soc., Vol. 51, No. 6, 2003 June
Experimental 65-loudspeaker canopy array at University of Gottingen
(AC-3) coder for the audio component
of the U.S. HDTV system. This coder
was also chosen in 1997 to supply audio on DVD-Video, with the DTS
coder as an alternate. In 1999 DVDAudio was introduced with very high
quality 24-bit multichannel audio and
MLP lossless coding.
The past few years have seen the rise
of Internet audio, the unprecedented establishment by consumers of MP3 as the
preferred format for compressed stereo
content, and the introduction of various
streaming audio formats by Real Networks, Microsoft, Apple, and others.
TAKING STOCK: THE PRESENT
At this point in the quest for ultimate
spatial audio fidelity, the 5.1-channel
format is established as the standard
surround sound format, supported by
DVD, film, Internet audio, and
HDTV. Yet there are already alternatives deployed or proposed, including
the SDDS 7.1-channel format, various
systems with three or four surround
loudspeakers, the IMAX channel configuration, and Tom Holman’s musicoriented 10.2-channel system. Tentative forays into better matching of
side and screen loudspeakers, top-ofscreen loudspeakers, and the use of a
vertical (“voice of God”) ceiling
channel have been made. Yet people
still have but two ears, and stereo continues to enjoy a high degree of consumer acceptance, supported by such
formats as audio CDs, MP3, much Internet audio, cassettes, and broadcast.
(For that matter, AM broadcast is still
largely monophonic.)
In fact, the continued strength of
the stereo format has in recent years
given rise to the 2:N channel upmixer
or surround synthesizer, which attempts to extract an aestheticallypleasing surround presentation from
conventional stereo content. Representative systems include Dolby Pro
Logic II, DTS Neo6, Lexicon Logic 7,
and the TC6000 Unwrap. Within limits, these systems help bridge the gap
between stereo and 5.1-channel discrete surround.
To date an astonishing array of systems have been proposed or implemented to approximate spatial sound
reproduction, including:
• Binaural
• 45/45 and vertical/lateral stereo disks
• FM/AM and FM multiplex broadcast
• Two-band stereo disks
• Stereo optical analog soundtracks on
film, usually with noise reduction,
and sometimes accompanied by auxiliary control tracks
• Stereo and multichannel tape recording and magnetic film soundtracks
• 4:2:4 and 5:2:5 matrix surround systems
• Helican scan FM stereo on videotape
• Digital PCM on magnetic tape and
optical disk (CD)
• Digital low bit-rate coders, including
multichannel coders.
There is every likelihood that this
list will continue to expand.
FUTURE DIRECTIONS?
In contemplating future advances in
spatial sound reproduction, it is im- ➥
565
History of Spatial Coding
portant to be realistic about the
strengths and weaknesses of current
technology. Existing 5.1-channel surround systems represent the cumulative contribution of an enormous number of innovative engineers and
inventors and, properly used, are capable of rendering a highly satisfactory
spatial audio experience. To be viable,
any candidate for a next-generation
system will likely have to provide a
significantly enhanced experience in a
cost-effective manner.
What might be desired in such a system? For one thing, a larger sweet spot
might be welcome. 5.1-channel systems can support fairly effective virtual
imaging for centered listeners but tend
to be perceived as individual discrete
channels by off-center listeners. Increasing the number of channels can
reduce this problem, but how many are
necessary? In this regard, a singular
point in the history of spatial audio has
now been reached, in that there exist
media capable of supporting more
channels than we know how to employ. For example, if 60-minute playing time is required and low bit-rate
coding at 32 kb/s per channel is used, a
standard DVD can hold in excess of
400 channels. There is considerable divergence of opinion on whether this is
still hopelessly inadequate or in fact
overkill. On the one hand, strict application of the spatial sampling theorem
would imply the need for a loudspeaker every 3/16th of an inch, conveying
potentially millions of channels. On
the other hand is the assumption that,
even in a large auditorium or theater,
virtual imaging becomes effective once
the channel density is great enough, so
far fewer channels would be adequate.
Jens Blauert, in his renowned book
Spatial Hearing, suggests that roughly
two dozen channels might suffice.
Before ramping up the number of
channels, however, some practical considerations may have to be addressed.
For example, content producers may
be less than enthusiastic about having
to mix several hundred discrete channels for a single program. Further, as
the number of channels increases, the
incremental contribution of any one
channel diminishes, to the point where,
from an engineering standpoint, it represents an inefficient use of available
566
bandwidth. This raises the question of
how side-chain information might potentially be used with a reduced number of conveyed channels to render a
larger number of output channels.
Another desirable element often
mentioned for future spatial audio systems is configuration neutrality: the
system automatically makes optimal
use of an arbitrary number of output
channels from a common program
source, chosen by the end user. This
would avoid the need to revamp the
entire system infrastructure every time
an additional channel is added. Perhaps
some sort of holographic system conveying spatial transform information
would be the preferred basis of a configuration-neutral audio format.
Beyond the difficulties of conveying
full spherical spatial information, there
are practical difficulties in presenting
it. As difficult as it may be to place
loudspeakers in a canopy above the listeners, it is probably even more challenging to arrange for sounds to come
from beneath listeners without a suspended floor. Indeed, in a theater it is
generally not even possible to have
loudspeakers at ear level, since they
will be too loud for near listeners and
all but inaudible on the far side of the
room. Instead they are placed above
the audience, resulting in much more
even sound distribution, at the cost of
limiting the ability to present sounds
that should be truly horizontal.
Even a seamless spherical presentation by itself won’t necessarily comprise an ultimate spatial audio system,
as a means to project sound into the
room would also be desired. Several
technologies are currently being explored to accomplish this, including ultrasonic drivers, phased loudspeaker arrays, or multichannel echo-cancellation.
Even if a fully immersing presentation
scheme can be developed, there will
still be the question of how to extend
current discrete channel spatial coding
techniques to the realm of full 3-D.
There may even be philosophical
questions to resolve. In cinema presentation, for example, each viewer
should receive an equivalent presentation. If, say, a parrot flies out from the
screen, should everyone perceive the
parrot as landing on one patron’s
shoulder, or should each viewer per-
ceive the parrot landing on his or her
shoulder? Is three-diminsional audio
even compatible with a two-dimensional movie screen?
One of the ultimate benchmarks of a
spatial audio system may be the simple
experience of standing out in a rainstorm. A million raindrops fall every
second, in every direction, and at every
distance from a fraction of an inch to a
mile. A proper spatial audio system
will reproduce the perception of each
splash at its proper distance and orientation for a single listener or an audience of hundreds. The realization of
such systems may not be that far off.
IN CONCLUSION
It seems appropriate to conclude with
two quotes about spatial audio.
Stereophonic systems do not
consist of two, three, or any other
fixed number of channels. There
must be sufficient of these to give a
good illusion of an infinite number.
And I know there will be
compromises. There wouldn’t be
any need for engineers if we didn’t
need compromises. It’s the one
that does the wise compromising
that succeeds best.
–HARVEY FLETCHER, 1953
[Resolving the spatial audio quest
is] the last great problem in audio.
–FRANCIS RUMSEY, 2001
ACKNOWLEDGMENTS
This review was made possible by the
heroic efforts of Dolby Laboratories’
librarians Tamara Horacek and Elizabeth Azinheira, who dug up a mountain of reference material and to whom
I extend my deepest thanks.
Thanks are also extended to Ioan
Allen, Gilbert Soulodre, Craig Todd,
and David Gray for constructive comments and contributions; to Tom Holman for his related talk that inspired
this work; to Kate Barrett for polishing
the prose; Rick Weldon for clearing
the photos; and to Eric Benjamin, Dolby’s ad hoc Ambisonics expert.
BIBLIOGRAPHY
http://www.acmi.net.au/AIC/
LE_PRINCE_BIO.html.
Alexander, Robert Charles, The
Life and Works of Alan Dower
J. Audio Eng. Soc., Vol. 51, No. 6, 2003 June
History of Spatial Coding
Blumlein (1999 Focal Press).
Allen, I., “The Production of Wide
Range, Low Distortion Optical Soundtracks Utilizing the Dolby Noise Reduction System,” SMPTE J., vol. 83,
no. 9, pp 729 (1974 Sept.).
Bauer, Ben and Stoles, George W.,
“Stereophonic Patterns,” J. Audio Eng.
Soc., vol. 8, pp. 126-129 (1960 Apr.).
Bauer, Ben, “Broadening the Area of
Stereophonic Perception,” J. Audio
Eng. Soc., vol. 8, pp. 91-94 (1960 Apr.).
Bauer, Ben, “Some Techniques Toward Better Stereophonic Perspective,”
J. Audio Eng. Soc., vol. 17, pp. 410415 (1969 Aug.).
Beaubien, William H. and Moore,
Harwood B., “Perception of the
Stereophonic Effect as a Function of
Frequency,” J. Audio Eng. Soc., vol. 8,
pp. 76-86 (1960 Apr.).
Bell, A. G., “Experiments Relating
to Binaural Audition,” Am. J. Otol.
(1880 July).
Bennett, J. C., Barker, K., and
Edeko, F. O., “A New Approach to the
Assessment of Stereophonic Sound
System Performance,” J. Audio Eng.
Soc., vol. 33, pp. 314-321 (1985 May).
Bernfeld, Benjamin, “Simple Equations for Multichannel Stereophonic
Sound Localization,” J. Audio Eng.
Soc. vol. 23, pp. 553-557 (1975 Sept.).
Blake, L., “Mixing Dolby Stereo
Film Sound,” Recording Engineer/Producer, vol. 12, p. 68 (1981 Feb.).
Blauert, J., Spatial Hearing (1983
MIT Press).
Blumlein, A. D., “Improvements in
and Relating to Sound-Transmission,
Sound-Recording and Sound-Reproducing Systems,” British Patent
394,325 (June 14, 1933).
Bore, Gerhart, “Principles and Problems of Stereophonic Transmission,”
presented at the 9th Convention of the
Audio Engingeering Society (1957
Oct.), preprint 30.
Burden, R. W., and Frohock, S. E.,
“Audio Considerations for Stereophonic Broadcasting,” J. Audio Eng. Soc.,
vol. 10, pp. 36-38 (1962 Jan.).
Camras, Marvin, “A Stereophonic
Magnetic Recorder,” Proc. Inst. Radio
Engrs., vol. 37, pp. 440-447 (1949).
Camras, M., “Approach to Recording a Sound Field,” J. Acoust. Soc.
Am., vol. 43, pp 1425-1431 (1968).
Ceoen, Carl, “Comparative StereoJ. Audio Eng. Soc., Vol. 51, No. 6, 2003 June
phonic Listening Tests,” J. Audio Eng.
Soc., vol. 20, pp 19-27 (1972 Jan./Feb.).
Cooper, Duane H. and Shiga, Takeo.
“Discrete-Matrix
Multichannel
Stereo,” J. Audio Eng. Soc., vol. 20,
pp. 346-360 (1972 June).
Corcoran, James and Williams, Douglas, “The Recording and Re-recording
of Stereophonic Sound for WideScreen Motion Pictures,” SMPTE J.,
vol. 77, pp. 1292-1294 (1968 Dec.).
Crowhurst, Norman, “Advantages,
Scope and Limitations of the Perspecta
Stereophonic System,” SMPTE
Journal, vol. 64, pp. 184-189 (1955
Apr.).
Crowhurst, Norman, “Basic Requirements for a Stereophonic System,” J. Audio Eng. Soc., vol. 5, pp.
129-135 (1957 July).
Davisson, L. D., “Rate Distortion
Theory and Application,” Proc. IEEE,
vol. 60, pp. 800-808 (1972 July).
Dutton, Gilbert F., “The Assessment
of Two-Channel Stereophonic Reproduction Performance in Studio Monitor
Rooms, Living Rooms and Small Theatres,” J. Audio Eng. Soc., vol. 10, pp.
98-105 (1962 Apr.).
DVD News, vol. 1., no. 3 (Feb. 9,
1998).
DVD News, vol. 1, no. 13 (1998
July).
Eargle, John, “Multichannel Stereo
Matrix Systems: An Overview,” J. Audio Eng. Soc., vol.19, pp. 552-559
(1971 July/Aug.).
Eargle, John and Streicher, Ron,
“Acoustical Perspectives in Commercial Two-Channel Stereophonic
Recording,” presented at the AES 8th
Int. Conference on The Sound of Audio (1990 May), paper 8-020.
Feldman, L., “Four Channel Sound,”
Sams, #20966 (1973 Indianapolis).
Fels, Peter, “20 Years of the Delta
Stereophony System High Quality
Sound Design,” presented at the 100th
Convention of the Audio Engineering
Society (1996 May), preprint 4188.
Finney, H. R., “Spacial Stereo—The
Subjective System Concept to Stereophonic Listening in the Home,” presented at the 11th Convention of the
Audio Engineering Society (1959
Oct.), preprint 122.
Forman, Albert J., “A New
Multiplex Stereo for Three-Dimensional Sound,” Tele-Tech,
vol. 12, p. 92 (1953 Apr.).
Fouque, M. and Redlich, H., “Space
Information in Stereophony,” presented at the 14th Convention of the Audio
Engineering Society (1962 Oct.),
preprint 267.
Frederick, H. A., “The Development
of the Microphone,” J. Acoust. Soc.
Am., vol. 3, 8 (1931)
Geluk, J. J., “Compatible Stereophonic Broadcasting Systems for Spatial Reproduction,” J. Audio Eng. Soc.,
vol. 28, pp 136-139 (1980 Mar.).
Gerrity, W. E. and Hawkins, J. N.
A., “Fantasound,” J. Soc. Motion Picture Eng. (1941 Aug.).
Gerzon, Michael, “Three Channels:
The Future of Stereo?” Studio Sound
(1990 June).
Gerzon, Michael, “Directional
Masking Coders for Multichannel Subband Audio Data Compression Systems,” presented at the 92nd Convention of the Audio Engineering Society
(1992 Mar.), preprint 3261.
Gilbert, Mark and Schulein, Robert
B., “Stereosurround™—A Compatible
Multichannel Encoding/Decoding Process for Audio and Audio/Video Applications,” presented at the 87th Convention of the Audio Engineering
Society (1989 Oct.), preprint 2855.
Goldmark, Peter C. and Hollywood,
John M., “Psychoacoustics Applied to
Stereophonic Reproduction Systems,”
J. Audio Eng. Soc., vol. 7, pp. 72-74
(1959 Apr.).
Goldmark, Peter C., “Maverick Inventor: My Turbulent Years at CBS,”
(1973 E. P. Dutton & Co.).
Gotoh, T., Kimura, Y., and Yamada,
A., “A New Sound Localization Control System for Stereophonic Recording,” presented at the 67th Convention
of the Audio Engineering Society
(1980 Oct.), preprint 1700.
Grignon, Lorin D., “Experiment in
Stereophonic Sound,” J. Soc. Motion
Picture Engrs., vol. 52, pp. 280-292
(1949).
Halstead, William S. and Burden,
Richard W., “A Compatible FM Multiplex System for Stereophonic Television Service,” J. Audio Eng. Soc., vol.
10, pp. 16-22 (1962 Jan.).
Harvey, F. K. and Schroeder, M.
R., “Subjective Evaluation of Factors Affecting Two-Channel
Stereophony,” J. Audio Eng. ➥
567
History of Spatial Coding
Soc., vol. 9, pp. 19-28 (1961 Jan.).
Hilliard, John K., “A Brief History
of Early Motion Picture Sound
Recording and Reproducing Practices,” J. Audio Eng. Soc., vol. 33, pp.
271-278 (1985 Apr.).
Hirsch, Charles J., “Progress Report
of Panel 1 of the National Stereophonic Radio Committee,” J. Audio Eng.
Soc., vol. 8, pp. 2-6 (1960 Jan.).
Hoffner, Randy, “Multichannel Television Sound Broadcasting in the United States,” J. Audio Eng. Soc., vol. 35,
No. 9, pp. 660-665 (1987 Sept.).
Holman, Tomlinson, “New Factors
in Sound for Cinema and Television,”
J. Audio Eng. Soc., vol. 39, pp. 529539 (1991 July/Aug.).
Horrall, Thomas R., “Auditorium
Acoustics Simulator: Form and Uses,”
presented at the 39th Convention of the
Audio Engineering Society (1970
Oct.), preprint 761.
Hugonnet, C. and Jouhaneau, J.,
“Comparative Spatial Transfer Function of Six Different Stereophonic Systems,” presented at the 82nd Convention of the Audio Engineering Society
(1987 Mar.), preprint 2465.
Jenrick, Paul W., “Integrating Multichannel Sound into Home Video Systems,” presented at the AES 4th Int.
Conf. on Stereo Audio Technology for
Television and Video (May 1986), paper 4-020.
Joel, Irv, “Multi-Channel Audio for
Television Broadcasting,” presented at
the 76th Convention of the Audio Engineering Society (1984 Oct.), preprint
2162.
Jonquet, A. and Pignon, J. P.,
“Holophony and Tetrahedrophony:
Some New Views about 3-D
Stereophony,” presented at the 56th
Convention of the Audio Engineering
Society (1977 March), preprint 1209.
Kellogg, Edward W., “History of
Sound Motion Pictures Part I,”
SMPTE J., vol. 64, pp. 291-302 (1955
June).
Kirby, D. G., “Experiences with
Multichannel Sound for HDTV,” presented at the 91st Convention of the
Audio Eningeering Society (1991
Oct.), preprint 3198.
Klipsch, Paul W., “Stereophonic
Sound With Two Tracks, Three Channels By Means of a Phantom Circuit
(2PH3),” J. Audio Eng. Soc., vol. 6,
568
pp. 118-123 (1958 Apr.).
Kohsaka, Osamu, Satoh, Eiji, and
Nakayama, Takeshi, “Sound-Image
Localization in Multichannel Matrix
Reproduction,” J. Audio Eng. Soc., vol.
20, pp. 542-548 (1972 Sept.).
Maxfield, J. P., Colledge, A. W., and
Friebus, R. T., “Pick-Up for Sound
Motion Pictures (Including Stereophonic),” J. Soc. Motion Picture Engrs. (1938 June).
McKnight, John G., “Why Stereo?
The Philosophy of Multichannel
Recording of Music,” J. Audio Eng.
Soc., vol. 8, pp. 87-90 (1960 Apr.).
Mosely, John, “Quintaphonic
Sound,” SMPTE J., vol. 86, pp. 20-29
1977 Jan.).
Mullin, John T., “Monogroove
Stereophonic Disk Recording,” J. Audio Eng. Soc., vol. 2, pp. 249-251
(1954 Oct.).
Nakayama, Takeshi, Miura, Tanetoshi, Kosaka, Osamu, Okamoto, Michio, and Shiga, Takeo, “Subjective
Assessment of Multichannel Reproduction,” J. Audio Eng. Soc., vol. 19,
pp. 744-751 (1971 Oct.).
Nguyen, Chieu, “Implementation of
Digital Audio Systems for Television
Multi-Channel Sound,” presented at
the 78th Convention of the Audio Engineering Society (1985 May), preprint
2226.
Nigro, John, “A Stereodynamic
Multichannel Amplifier for Single or
Binaural Input,” J. Audio Eng. Soc.,
vol. 1, pp. 287-291 (1953 Oct.).
Offenhauser Jr., W. H. and Israel,
J. J., “Some Production Aspects of
Binaural Recording for Sound Motion Pictures,” J. Soc. Motion Picture Engrs., vol. 32, pp. 139-155
(1939 Feb.).
Olson, Harry F., “Stereophonic
Sound Reproduction in the Home,” J.
Audio Eng. Soc., vol. 6, pp. 50-60
(1958 Apr.).
Potter, R. K., “System for Binaural
Transmission of Sound,” U.S. Patent
1,608,556 (Nov. 30, 1926).
Rawlence, Christopher, The Missing
Reel (1990 Collins, UK).
Scheiber, Peter. “Four Channels and
Compatibility,” J. Audio Eng. Soc.,
vol. 19, pp. 267-279 (1971 Apr.).
Schoenerr, Steve, http://history.
acusd.edu/gen/recording/stereo.html.
Schroeder, Manfred R., “An Artifi-
cial Stereophonic Effect Obtained from
a Single Audio Signal,” J. Audio Eng.
Soc., vol. 6, pp. 74-79 (1958 Apr.).
Schroeder, M. R. and Atal, B. S.,
“Computer Simulation of Sound
Transmission in Room,” IEEE Int.
Conv. Rec. P. 150 (1963).
Selsted, Walter T., “Multichannel
Sound Reproduction,” J. Audio Eng.
Soc., vol. 2, pp. 20-24 (1954 Jan.).
Shiga, Takeo, Okamoto, Michio, and
Cooper, Duane H., “Dual-Triphonic
Matrix Stereo System,” presented at
the 40th Convention of the Audio Engineering Society (1971 Apr.), preprint
783.
Snow, W. B., “Effect of Arrival
Time on Stereophonic Localization,” J.
Acous. Soc. Am., vol. 26 (1965 Nov.).
Snyder, Ross H., “History and Development of Stereophonic Sound
Recording,” J. Audio Eng. Soc., vol. 1,
pp 176-179 (1953 Apr.).
Snyder, Ross H., “Stereophonic
Sound for the Videotape Recorder,” J.
Audio Eng. Soc., vol. 7, pp. 213-216
(1959 Oct.).
Sobol, Norbert, “The DSP 610—A
Computer Controlled Processor for a
Truly Directional Sound Reinforcement System (The Delta Stereophony
System),”presented at the AES 6th Int.
Conf. on Sound Reinforcement (1988
May), paper 6-015.
Steinberg, J. C. and Snow, W. B.,
“Symposium on wire transmission of
symphonic music and its reproduction
in auditory perspective: physical factors,” Bell System Technical J., vol. 13
(1934 Apr.).
Snyder, R. H., “History and Development of Stereophonic Sound
Recording,” J. Audio Eng. Soc., vol. 1,
pp. 176-179 (1953 Apr.).
Stockham, Thomas G., “Restoration
of Old Acoustic Recordings by Means
of Digital Signal Processing,” presented at the 41st Convention of the Audio
Engineering Society (1971 Oct.),
preprint 831.
Stockham, Thomas G., “Records of
the Future,” J. Audio Eng. Soc., vol.
25, pp. 892-895 (1977 Oct./Nov.).
Stockham, Thomas G., “The Impact
of Digital Audio,” presented at the
AES 2nd Regional Convention (1987
June), preprint 2647.
Tanaka, E., Kusunoki, Y., Sugiyama, Y., Nakajima, O, Furukawa T.,
J. Audio Eng. Soc., Vol. 51, No. 6, 2003 June
History of Spatial Coding
Runii, S., “On PCM Multi-Channel
Tape Recorder Using Powerful Code
Format,” presented at the 67th Convention of the Audio Engineering Society (1980 Oct.), preprint 1690.
Tager, Pavel G., “Some Features of
Physical Structure of Acoustics Fields
of Stereophonic Systems,” SMPTE J.,
pp. 105-110 (1967 Feb.).
Templin, E. W., “Recent Developments in Multichannel Stereophonic
Recording Systems,” SMPTE J., vol.
676, pp. 53-58 (1957 Feb.).
Theile, Gunther, “On the Performance of Two-Channel and MultiChannel Stereophony,” presented at
the 88th Convention of the Audio Engineering Society (1990 Mar.), preprint
2887.
Theile, Gunther, “Further Developments of Loudspeaker Stereophony,”
presented at the 89th Convention of the
Audio Engineering Society (1990
Sept.), preprint 2947.
Theile, Gunther, Stoll, Gerhard, and
Silzie, Andreas, “MUSICAM-Surround: A Multi-Channel Stereo Coding
Method,” presented at the 92nd Convention of the Audio Engineering Society (1992 Mar.), preprint 3337.
Theile, Gunther and Stolle, Gerhard,
“MUSICAM-Surround: A Universal
Multi-Channel Coding System Compatible with ISO 11172-3,” presented
at the 93rd Convention of the Audio
Engineering Society (1992 Oct.),
preprint 3403.
Toole, F. E., “In-Head Localization
of Acoustic Image,” J. Acous. Soc.
Am., vol. 48, p. 943 (1970).
Toole, Floyd E., “The Acoustics
and Psychoacoustics of Loudspeakers
and Rooms—The Stereo Past and the
Multichannel Future,” presented at
the 109th Convention of the Audio
Engineering Society (2000 Sept.),
preprint 5201.
Torick, Emil L., “AM Stereophonic
Broadcasting—An Historical Review,”
J. Audio Eng. Soc., vol. 23, pp. 802805 (1975 Dec.).
Torick, Emil L., “Improving the Signal-to-Noise Ratio and Coverage of
FM Stereophonic Broadcasts,” J. Audio Eng. Soc., vol. 33, pp. 938-943
(1985 Dec.).
Torick, Emil L., “Highlights in the
History of Multichannel Sound,” J.
Audio Eng. Soc., vol. 46, pp. 27-31
(1998 Jan./Feb.).
Uhlig, Ronald E., “Stereophonic
Photographic Soundtracks,” SMPTE J.,
pp. 292-295 (1973 Apr.).
Uhlig, Ronald E., “Two- and ThreeChannel Stereophonic Photographic
Soundtracks for Theaters and Television,” SMPTE J., vol. 83, pp. 729-732
(1974 Sept.).
Voelker, E. J., Mueller, M, and Teuber, W., “Multi-Channel Digitally Delayed Sound System with Perfect Direction Impression,” presented at the
80th Convention of the Audio Engineering Society (1986 Mar.), preprint
2353.
http://web.inter.nl.net/users/anima/pr
e-cinema/leprince/index.htm.
Welch, Walter L. and Burt, Leah B.
S., From Tinfoil to Stereo (1994 University Press of Florida).
Wood, Irving W., and Fichman, Joel
S., “The Sound System at the New
York State Theatre,” J. Audio Eng.
Soc., vol. 13, pp. 104-110 (1965 Apr.).
http://www.xs4all.nl/~rabruil/
phmil.html.
THE AUTHOR
After receiving his Ph.D. in psychoacoustics and
electrical engineering from MIT in 1980, Mark Davis
worked at dbx, Inc., where he designed the
dbx/MTS stereo television noise-reduction system
and the Soundfield One phased array loudspeaker.
Since 1985 he has been a senior engineer in the
R&D Department of Dolby Laboratories, where he
helped design AC-3, Dolby Virtual Surround, and
DSP ports of Dolby Surround Pro Logic and Dolby
SR noise reduction.
J. Audio Eng. Soc., Vol. 51, No. 6, 2003 June
THE PROCEEDINGS
OF THE AES 18TH
INTERNATIONAL
CONFERENCE
AUDIO FOR
INFORMATION
APPLIANCES
Challenges, Solutions, and
Opportunities
n
le o
ilab
a
v
A
M
CD
-RO
$40.00 Members
$60.00 Nonmembers
This conference looked at the
new breed of devices, called information appliances, created
by the convergence of consumer electronics, computing,
and communications that are
changing the way audio is created, distributed, and rendered.
FOR ORDERING INFORMATION
You can purchase the book and
CD-ROM online at
www.aes.org.
For more information email
Andy Veloz at
aav@aes.org or
tel: +1 212 661 8528 ext. 39.
569