Documente Academic
Documente Profesional
Documente Cultură
An
attempt
to
connect
another
theatre
to
the
exhibition
failed,
due
to
interference
from
the
footlights.
In
France
from
1890
to
1932
as
the
Thtrophone,
and
in
England
from
1895
to
1925
as
the
Electrophone.
upon
the
scope
of
the
invention
or
the
manner
and
means
whereby
it
may
be
accomplished.
Blumlein
was
a
prolific
inventor,
so
only
some
of
the
ideas
put
forward
to
the
Patent
Office
were
fully
fledged.
Others
were
described
only
in
broad
terms.
What
is
clear
is
that
he
saw
the
primary
use
of
stereo
for
use
with
motion
pictures:
This
invention
relates
to
the
transmission,
recording-
and
reproduction
of
sound
and
is
more
particularly
directed
to
systems
for
recording
and
reproducing
speech,
music
and
other
sound
effects,
especially
when
associated
with
picture
effects
as
in
talking
motion
pictures.
The
fundamental
object
of
the
invention
is
to
provide
a
sound
recording
and
reproducing
system
whereby
a
true
directional
impression
may
be
conveyed
to
a
listener
thus
improving
the
illusion
that
the
sound
is
coming,
and
is
only
coming,
from
the
artist
or
other
sound
source
presented
to
the
eye.
Blumlein
proposed
to
achieve
this
by
using
intensity
difference
between
two
(or
more)
loudspeakers.
Around
1900,
speculative
knowledge
of
how
we
hear
the
direction
of
sound
had
been
distilled
into
Lord
Rayleighs
Duplex
Theory
of
sound
localisation.
By
the
1930s
this
explanation
(which,
while
not
giving
a
complete
explanation,
is
still
correct)
was
well
established,
and
it
is
on
this
understanding
that
Blumlein
bases
his
invention.
In
the
patent
he
makes
clear
reference
to
his
system
differentiating
between
high
(ILD)
and
low
(ITD)
frequencies.
The
system
described
is
more
complicated
than
a
modern
true-stereo
recording
(where
the
mic
technique
used
determines
intensity
/
phase
differentiation),
or
constructed
stereo
(intensity
only).
The
method
proposed
uses
two
(or
more)
binaural
microphone
signals
that
are
matrixed,
using
transformers,
to
derive
sum
and
difference
signals.
Low
frequencies
get
converted
from
phase
difference
to
amplitude
difference,
as
Blumlein
had
discovered
that
even
a
low
frequency
phantom
image
could
be
localized
simply
by
controlling
the
relative
intensities
of
the
two
channels.
The
amplitude
of
all
frequencies
could
therefore
be
controlled
with
resistors.
The
system
he
described
in
detail
uses
two
omni-directional
microphones
separated
by
a
block
of
wood
to
mimic
a
human
head.
These
are
amplified
then
fed
to
a
transformer
matrix
which
derives
sum
and
difference
signals.
This
would
essentially
be
the
recording
part
of
the
signal
chain.
On
the
reproduction
side,
the
two
channels
are
again
processed
using
sum
and
difference
techniques.
The
summed
channel
acts
as
a
reference
and
the
other
uses
a
capacitor3
as
a
phase
comparator
to
convert
any
phase
difference
to
amplitude
difference.
A
third
sum
and
difference
circuit
is
then
employed
to
derive
the
signals
that
feed
the
two
loudspeakers.
This
Mid-Side
technique
was
partially
driven
by
the
need
for
mono
compatibility
in
picture
theatres
(for
backward
compatibility,
or
in
case
one
channel
failed).
Speaker
width
is
stated
as
not
exceeding
the
screen
width,
and
not
closer
together
than
70%
of
the
screen
width
for
low
frequencies.
For
high
frequencies
the
width
should
be
gauged
by
trial
and
error.
Several
microphone
methods
were
described:
a)
Binaural4:
two
pressure
mics
separated
by
a
dummy
head
(block
of
wood).
This
rather
complex
setup
required
decoding
into
stereo
b)
M-S:
using
two
velocity
mics,
one
facing
forwards,
the
other
side-on.
3
The
value
is
not
mentioned,
but
it
would
need
Xc
to
take
effect
around
1500Hz
(giving
a
delay
of
167S).
4
No
distinction
between
binaural
and
stereo
was
made
until
1932
[5]
c)
d)
Blumlein
Pair:
two
velocity
mics
at
45
to
the
centre
of
the
screen.
No
decoding
is
required
XY:
where
the
angle
may
be
varied,
to
avoid
the
necessity
of
combining
and
re-
separating
the
two
channels
Using
additional
mics
in
the
vertical
plane
is
mentioned
to
give
3D
sound
location.
There
are
no
details
on
this,
and
it
appears
that
Blumlein
thought
that
there
would
be
no
perceptual
differences
from
the
horizontal
to
the
vertical
plane.
Blumlein
discussed
how
to
cut
stereo
onto
a
disc,
and
four
recording
head
designs
are
provided.
He
described
combining
hill
and
dale
and
lateral,
and
the
preferred
method
of
45
/
45
cutting.
He
correctly
identified
that
this
stereo
reproduction
would
likely
appeal
to
music
recording,
as
well
as
motion
picture.
In
1916
Bell
Laboratories
produced
the
first
quality
studio
mic
(rather
than
a
Carbon
mic).
This
had
to
wait
for
about
10
years
until
electrical
amplification
was
available,
using
the
triode
thermionic
valve.
These
were
omni-directional
mics.
By
the
time
Blumlein
tendered
his
Complete
Specification
for
approval,
Harry
Olson
had
developed
the
first
practical
ribbon
microphone5.
In
1932
Neumann
produced
the
first
cardioid
mic,
using
a
double
membrane
capsule.
Blumlein
would
no
doubt
have
been
aware
of
these
developments,
and
it
seems
that
some
of
the
open-ended
statements
in
the
application
were
to
allow
for
use
of
new
technological
developments
such
as
unidirectional
microphones.
EMI
was
not
the
only
company
interested
in
multi-channel
sound,
and
in
1933
Steinberg
and
Snow
from
Bell
Laboratories
published
a
paper
[4]
proposing
a
system
with
good
localisation
using
three
channels.
The
experiment
consisted
of
seeing
if
a
panel
of
listeners
could
accurately
locate
the
positions
of
persons
speaking
on
a
stage,
when
reproduced
via
3
loudspeakers.
This
was
found
to
provide
good
results
for
both
fore-aft
and
lateral
localization.
Then
the
experiment
was
repeated
using
two
channels.
This
time
any
positions
on
the
centre-
line
of
the
stage
appeared
to
come
from
the
rear
of
the
stage.
Secondly,
the
virtual
soundstage
was
wider
than
for
3
channels.
Combinations
were
also
tried,
namely:
a)
adding
a
centre
mic
to
the
2
channel
system
(which
pulled
the
centre-line
positions
forward),
b)
bridging
a
third
loudspeaker
across
the
2
channel
system
(which
increased
depth
but
narrowed
the
soundstage),
and
c)
adding
a
third
mic
and
a
third
loudspeaker
across
the
2
channel
system
(this
greatly
reduced
the
soundstage
width).
None
of
these
were
considered
to
provide
realistic
spatial
information,
so
the
3
channel
system
was
chosen.
The
Bell
Laboratories
engineers
performed
further
tests
on
depth
localisation
and
discovered
that
a
reduction
in
playback
loudness,
or
an
increase
of
source
reverberation
ratio,
or
both,
caused
the
sound-source
to
recede.
The
authors
then
add
It
has
not
been
found
possible
to
put
these
relationships
on
a
quantitative
basis.
The
experiment
identified
what
would
happen,
but
the
engineers
were
unable
to
explain
why.
The
next
part
of
their
research
was
an
experiment
to
map
direct
sound
loudness
levels
to
each
ear
as
a
sinewave
source
is
rotated
by
180
from
the
front
to
the
back
of
the
subjects
5
Walter
Schottky
and
Erwin
Gerlach
invented
the
first
ribbon
mic
in
1924
head.
Frequencies
from
300
Hz
to
15kHz
were
used.
Steinberg
and
Snow
identified
three
factors
responsible
for
localization:
phase
difference,
loudness
difference,
and
quality
difference.
They
discounted
phase
difference
(It
is
difficult
to
see
how
phase
differences
in
this
case
can
assist
in
localization
in
the
ordinary
way),
and
chose
loudness,
and
quality.
Quality,
they
surmised
was
due
to
frequency
distortion
of
the
sound
source
wrt
angle.
This
really
is
just
another
way
of
describing
loudness
separation,
so
they
were
barking
up
the
wrong
tree.
And
yet
they
were
very
close;
the
plots
do
indicate
that
the
ears
frequency
response
not
only
is
not
flat,
but
varies
considerably
with
the
direction
of
sound.
Had
they
not
dismissed
phase
so
quickly
they
may
have
headed
in
the
direction
of
discovering
HRTFs.
In
this
era
the
field
of
psycho-acoustics
was
still
in
its
infancy6
(research
on
Pinna
transforms
was
begun
in
1959
by
Dwight
W
Batteau.
His
findings
were
published
in
1964).
The
important
point
here
is
that
the
basis
of
stereo
recording
and
reproduction
was
formed
before
a
complete
theory
of
psychophysical
localization
emerged.
In
the
paper
the
authors
do
make
a
statement
that
predates
discovery
of
the
Haas
(or
Precedence)
effect7
but
do
not
distinguish
between
intensity
difference
and
reverb
to
direct
signal
ratio.
Later
(in
the
summary)
they
do
correctly
note
that
reverb
ratio
plays
no
part
in
localization,
but
does
effect
depth
perception
(with
the
converse
also
being
true,
ie
depth
localisation
was
arbitrary).
Steinberg
and
Snow
noted
that
there
was
good
correlation
between
calculated
and
observed
results
for
apparent
angles.
Two
and
three
speaker
systems
were
compared,
and
the
two
channel
system
gave
a
wider
stereo
image,
but
with
greater
angular
distortion
(ie
there
was
spatial
compression
at
the
extremities).
Thus
they
concluded
that
a
3
channel
system
was
definitely
superior
as
it
prevented
the
hole
in
the
middle,
and
the
centre
channel
would
allow
independent
control
of
the
musical
soloist.
Furthermore,
it
minimized
localization
differences
wrt
listening
position.
They
added
that
for
music
a
spatial
impression
was
more
important
than
accurate
localization
and
therefore
a
2
channel
system
gives
good
satisfaction.
3
channel
reproduction
was
recommended
for
speech
and
motion
picture
sound.
These
early
experiments
in
stereo
share
a
common
goal;
that
the
actual
soundstage
should
be
represented
as
accurately
as
possible
upon
reproduction.
Conclusions
were
based
on
empirical
evidence
and
so
a
considerable
part
of
the
research
time
was
spent
producing
sound
sources
at
known
locations
(these
were
marked
off
at
set
angles)
and
then
finding
what
the
subjective
results
for
matching
these
positions
was
when
reproduced.
Stereo
in
the
Home
Stereo
had
a
very
long
incubation
period.
Although
Blumlein
had
successfully
conceived
of
(and
demonstrated)
a
viable
stereo
recording
and
reproduction
system
in
1931,
the
birth
of
stereo
had
to
wait
until
1958.
Several
reasons
for
this
delay
can
be
proposed.
Firstly,
the
early
1930s
was
the
time
of
the
great
depression.
This
meant
that
sales
of
expensive
luxury
items
went
into
decline.
This
was
also
the
time
of
the
birth
of
public
broadcast
radio.
People
didnt
want
to
pay
for
songs
when
they
could
hear
them
for
free.
As
a
result
of
these
two
factors
Gramophone
record
sales
dropped
substantially.
6
although
not
without
the
odd
breakthrough
eg
Fletcher
&
Munsons
1933
work
on
loudness
curves.
7
Haas (Germany) and Wallach et al (USA) both published their work in 1949.
Another
factor
would
have
been
the
general
reluctance
of
consumers
to
upgrade
equipment.
By
the
1940s
many
people
would
have
owned
a
gramophone
and
a
considerable
collection
of
78s.
The
big
ticket
item
for
the
home
was
a
radio,
rather
than
an
electrical
record
player.
This
would
have
slowed
sales,
as
going
to
a
non-compatible
expensive
new
system
is
always
somewhat
unattractive.
No
doubt
WWII
was
also
part
of
the
reason,
as
research
and
manufacturing
resources
were
redirected
for
military
purposes.
After
the
war
the
USA
was
in
a
good
position
to
accelerate
electronics
development.
In
1948
the
mono
LP
vinyl
record
became
available.
Further
delays
of
the
introduction
of
stereo
well
may
have
been
simply
due
to
getting
a
standard
and
reliable
system
to
the
market.
1958
was
the
first
time
standard
stereo
records
were
available
to
the
general
public.
Acceptance
was
slow,
and
it
was
well
into
the
1960s
before
stereo
LP
sales
outstripped
mono
LPs.
To
promote
the
new
format,
manufacturers
and
record
companies
produced
elaborate
short
films
and
demonstration
discs.
These
invariably
made
exaggerated
claims
about
stereo.
Words
and
phrases
often
used
are
truest,
most
life-like,
perfect
fidelity,
realism,
full-
dimension
and
so
on.
The
demonstration
material
ranges
from
sound
effects
(the
moving
train
being
the
classic
stereo
clich)
to
speech,
and
musical
excerpts.
These
are
interesting,
in
that
the
music
is
mixed
quite
differently
than
it
is
today.
To
ensure
the
listener
would
not
miss
the
advantage
of
stereo
the
recordings
split
sounds
either
to
the
left
or
right
channel.
The
main
reason
for
this
is
obvious:
to
sell
stereo
to
the
public
it
had
to
be
shown
that
it
was
much
better
than
mono,
and
the
way
to
do
that
was
by
ping-ponging
sounds
(musical
or
otherwise)
between
the
two
speakers.
By
1960
many
homes
already
had
a
quality
piece
of
furniture
sitting
in
the
lounge;
the
radiogram.
Very
few
of
these
were
upgradable
so
a
newer
expensive
system
was
required.
Additionally,
a
second
loudspeaker
needed
to
be
place
somewhere.
It
is
exactly
at
this
starting
point
of
stereo
where
the
ideal
and
the
practicalities
are
in
opposition.
Many
homes
simply
did
not
have
the
space
to
set
up
two
loudspeakers
in
an
optimum
layout.
The
manufacturing
companies
were
aware
of
this
and
solved
the
problem
by
adding
the
second
speaker
to
one
end
of
the
radiogram,
creating
the
stereogram.
Thus,
this
slightly
larger
piece
of
furniture
would
fit
into
a
single
position
in
the
room
(in
many
cases
simply
being
put
where
the
old
radiogram
was).
In
terms
of
penetrating
the
market
this
was
a
clever
move.
The
stereo
demonstrations
were
not
just
larger
than
life
due
to
slick
marketing.
They
had
to
have
hard
left
and
right
panning
for
people
to
get
any
impression
of
stereo,
as
the
loudspeakers
in
a
typical
stereogram
were
only
about
1.2m
apart.
This,
of
course,
made
nonsense
of
the
narrators
use
of
words
such
as
realism.
Less
obvious
is
the
fact
that
many
studio
engineers
were
also
part
of
creating
unrealistic
stereo.
It
seems
that
since
30
years
had
passed
since
Blumlein
laid
the
foundation
for
stereo,
recording
engineers
had
lost
touch
with
his
research,
and
were
now
in
the
position
of
providing
stereo
popular
records
to
the
market.
Without
any
set
guidelines,
they
experimented
with
the
new
format
with
a
no
rules
attitude.
Many
60s
records
offer
the
same
stereo
attitude
as
the
demonstration
discs,
for
example
bass
or
lead
vocal
panned
hard
to
one
side.
Examples
of
this
style
of
stereo
mixing
can
be
heard
on
Beatles
recordings,
which
clearly
means
that
lack
of
general
recording
expertise
was
not
a
factor.
The
Beatles
catalogue
is
an
interesting
case
study.
The
first
album
(Please
Please
Me)
was
recorded
in
1963,
some
5
years
after
the
introduction
of
stereo
records
to
the
public.
In
the
UK
the
economy
was
still
pulling
out
of
the
post-war
slump
so
things
were
more
conservative
than
in
the
USA.
This
meant
that
limitations
in
recording
equipment
restricted
the
stereo
to
be
vocal
(channel
1)
+
band
(channel
2).
Arguments
continue
on
Beatles
fan
forums
as
to
which
is
better;
the
mono
or
stereo
version
of
each
album.
Recent
full
collection
releases
have
provided
us
with
both,
and
many
prefer
the
sound
of
the
mono
mixes
as
these
were
the
main
mix
for
release
and
had
the
most
time
and
attention
paid
to
them.
Having
said
that,
from
a
collectors
point
of
view
a
first
pressing
of
Please
Please
Me
is
valued
at
3000
for
stereo
and
750
for
mono.8
In
any
case,
by
todays
stereo
mixing
standards,
early
Beatles
recordings
sound
too
quirky
or
even
just
wrong
to
many
listeners,
indicating
that
we
have
become
conditioned
to
a
certain
style
of
sound
field
for
popular
recordings.
One
statement
from
Steinberg
and
Snows
1931
report
stands
out:
Manipulation
of
circuit
conditions
probably
can
be
used
advantageously
to
heighten
the
illusions
or
to
produce
novel
effects.
The
recordings
of
the
1960s
certainly
applied
such
experimentation
into
production
techniques.
By
the
early
1970s
a
de-facto
convention
for
stereo
placement
from
a
multitrack
source
had
been
established
(for
better
or
worse),
and
has
remained
ever
since.
Attempts
to
go
beyond
stereo
As
we
have
seen,
EMI
experiments
in
the
1930s
were
for
both
two
and
three
channel
sound.
These
numbers
were
decided
not
from
a
perspective
of
moving
up
from
a
single
channel
but,
rather,
moving
down
from
a
large
number
of
channels
(ideally
infinite)
to
a
practical
and
economic
system.
It
is
therefore
not
surprising
that
over
the
years
several
other
ideas
emerged
proposing
adding
extra
channels,
extra
virtual
channels,
or
even
superseding
stereo
altogether.
In
1958
Klipsch
[6,7]
proposed
a
centre
loudspeaker
be
added
to
his
stereo
horn
system.
This
was
due
to
the
horns
being
large
and
therefore
necessitating
a
large
room,
which
led
to
a
hole
in
the
middle
of
the
stereo
image
because
of
the
wide
spacing.
Quadraphonic
sound
can
be
described
as
the
most
successful
failure
of
all
the
multi-channel
systems
for
music.
Although
its
reign
in
home
audio
was
short-lived
it
did
manage
to
gain
enough
momentum
to
provide
the
public
with
a
significant
amount
of
product
both
equipment
and
musical
titles.
The
idea
seemed
simple
enough
two
rear
loudspeaker
channels
were
added
so
that
the
four
loudspeakers
were
spaced
at
90.
The
listener
could
then
be
immersed
in
the
sound
stage.
The
first
systems
used
pre-recorded
4
track
tapes,
making
mapping
to
four
loudspeakers
simple.
How
the
mix
was
recorded
onto
each
track
had
to
be
determined
at
the
recording
studio,
but
this
was
not
needing
deep
consideration
(this
was
only
a
few
years
after
the
availability
of
stereo
to
most
people,
so
no
rules
panning
was
supposedly
still
acceptable)
[8].
Several
incompatible
methods
were
brought
to
the
market.
JVCs
CD-4
pushed
the
limit
of
what
could
be
reliably
delivered
to
a
consumer
market
and
early
systems
were
often
hampered
by
noise
and
distortion.
It
used
a
30kHz
carrier
signal
to
carry
the
two
extra
channels,
and
could
resolve
normal
stereo,
if
played
back
on
a
standard
stereo
system.
To
obtain
the
required
50kHz
bandwidth
from
a
record
considerable
difficulties
had
to
be
overcome.
Among
these
was
record
wear
and
a
special
stylus
that
tracked
at
only
0.5
gm
was
required.
Thus,
CD-4
was
far
from
plug
n
play,
and
some
tinkering
was
required
to
get
good
results.
This
technology
did
improve
and
by
1974
was
a
fully
workable
system.
Another
system
developed
by
Nippon
Columbia9
used
a
combination
of
encoding
matrix
signals
and
8
http://beatlescollectingguide.weebly.com/please-please-me.html
UD-4
modulated
signals
(also
using
a
30kHz
carrier)
onto
the
record
groove.
It
was
very
much
an
also-ran
in
the
quad
race,
but
was
the
only
method
to
be
able
to
accurately
rotate
a
point-
source
through
360
[9].
Commercial
quad
systems
came
at
some
expense
to
the
consumer.
A
fully
discrete
setup
required
addition
of
either
a
CD-4
cartridge
and
demodulator,
or
4-track
tape
recorder,
as
well
as
two
extra
amplifiers
and
loudspeakers.
Several
ideas
were
proposed
(typically
in
DIY
electronics
magazines
which
were
still
popular
at
the
time)
to
do
it
on
the
cheap.
The
simplest
of
these
was
to
wire
a
third
loudspeaker
between
the
+
terminals
of
the
L
and
R
loudspeakers.
A
potentiometer
in
series
controlled
the
level
of
the
third
loudspeaker
which
needed
to
be
set
at
about
-6dB
for
any
sense
of
realism.
The
loudspeaker
was
positioned
at
the
rear
of
the
listening
position,
and
the
pot
adjusted
to
give
suitable
ambience.
A
variation
of
this
was
to
have
two
extra
loudspeakers
wired
in
series
(but
out
of
phase).
In
this
case
they
could
be
place
in
the
rear
corners.
A
commercial
version
was
also
available
(from
Hafler-
Dynaco10).
Another
possibility
was
to
add
the
two
loudspeakers
at
the
front
of
the
listening
area.
The
idea
of
this
is
to
cancel
inter-aural
crosstalk
between
the
listeners
ears
and
the
main
loudspeakers.
If
the
positions
of
the
additional
loudspeakers
are
close
to
the
main
speakers
(and
selected
on
test)
then
the
soundstage
can
be
expanded.
Again,
commercial
versions
of
the
technique
were
released11.
As
well
as
widening
the
stereo
image,
it
gives
some
depth
perception.
The
method
is
explained
in
the
Appendix
1.
A
third
possibility
is
a
Shuffler
circuit.
The
idea
goes
right
back
to
Blumleins
patent
where
he
described
his
main
stereo
method.
In
fact,
that
was
a
quasi-binaural
recording
mic
arrangement
from
which
M-S
was
derived.
This
was
then
shuffled
to
provide
width
control
before
being
converted
back
to
stereo
for
playback.
A
shuffler
can
also
take
a
stereo
signal
and
convert
it
to
M-S
where
the
Side
signal
can
be
manipulated
in
amplitude,
or
with
an
equaliser
before
being
reconverted
into
stereo.
This
works
as
the
L
and
R
signals
are
no
longer
identical,
and
the
technique
was
used
to
create
a
pseudo-stereo
signal
in
the
days
of
monophonic
reproduction.
The
matrixing
is
shown
in
Appendix
2.
Apart
from
discrete
quad,
there
were
two
main
rival
matrix
systems:
Sansuis
QS
and
CBSs
SQ.
These
relied
on
attenuation
and
phase
shifting
to
generate
the
effect,
one
with
+90
the
other
-90,
which
meant
the
two
systems
were
very
similar
but
were
incompatible.
A
perfect
system
could
not
be
realised
and
both
suffered
from
some
inter-channel
leakage
(separation
could
be
as
low
as
3dB).
Both
systems
are
based
on
the
work
of
Scheiber
[10].
Quadraphony
did
interest
the
BBC,
who
between
1969
and
1973
conducted
comprehensive
subjective
tests
of
no
less
than
eight
systems,
including
SQ
and
QS
[11].
A
major
concern
was
that
neither
of
these
commercial
systems
were
mono
compatible
(an
important
consideration
for
broadcasting).
It
was
the
idea
of
quadraphonic
radio
that
the
BBC
(and
other
broadcasters)
were
primarily
interested
in,
and
a
paper12
describing
a
method
for
this
(H-Matrix)
was
published
in
1977,
but
by
then
quadraphony
was
all
but
finished.
The
lessons
learnt
at
the
end
of
the
quad
era
were
several.
The
most
important
was
that
to
avoid
a
marketing
disaster
when
introducing
new
technology
to
the
market
it
is
best
to
have
a
single
standard
format.
Consumers
are
now
very
wary
of
having
rival
formats
battle
it
out
at
their
expense
(eg
VHS
vs
Beta).
It
is
no
surprise
that
hugely
successful
technologies
(such
10
as
MIDI,
CD)
have
provided
a
solid
and
longstanding
standard.
In
terms
of
the
sound
itself,
going
from
two
to
four
channels
is
far
more
than
double
the
trouble.
Stereo
has
difficulties
enough
with
two
actual
signal
sources,
a
phantom
image,
and
acoustic
inter-channel
crosstalk.
With
quad
this
multiplied
to
six
possible
phantom
images
giving
a
total
of
ten
sound
sources
plus
at
least
two
sets
of
crosstalk.
Discrete
systems
could
only
hold
the
channels
in
place
up
to
the
loudspeaker,
and
matrixed
systems
simply
had
no
ability
to
accurately
control
sound
location.
This
point
is
pivotal
to
quads
failure.
A
new
iPhone
costs
about
40
times
more
than
a
landline
telephone,
but
we
don't
see
Apple
going
out
of
business.
Consumers
weigh
up
the
benefit
to
cost
ratio.
Quad
for
all
its
expense
could
only
offer
two
things:
ambience
(if
the
rear
channels
were
subdued),
or
wow
factor
(immersion,
and
artificial
panning
experiences).
By
the
1970s
stereo
recording
techniques
and
listeners
sophistication
had
moved
ahead.
No
longer
were
the
masses
going
to
be
fascinated
with
the
novelty
of
surround
as
they
had
been
with
the
introduction
of
stereo.
Most
quad
demonstration
records
were
trying
to
impress
by
adding
spaciousness
and
realism.
On
this
account,
quad
could
not
survive.
To
sound
natural,
only
a
hint
of
rear
channel
balance
was
required.
By
the
time
the
rear
speakers
were
turned
up
enough
to
make
the
purchase
cost
worthwhile,
it
put
the
listener
inside
the
orchestra.
Great
for
effect,
but
hardly
a
long
term
proposition.
Lastly,
quad
went
beyond
the
WAF13.
For
many
casual
enjoyers
of
music
in
the
home,
having
four
loudspeakers
that
had
to
be
specifically
located
was
simply
a
bridge
too
far,
regardless
of
cost.
What
would
be
next
another
four
loudspeakers
at
ceiling
height?
Quadraphonic
sound
did
generate
a
multi-channel
spin-off
(indirectly).
This
is
Ambisonics
[12],
and
the
method
is
largely
the
work
of
Michael
Gerzon
[13].
It
varies
from
stereo
and
quad
systems
in
that
the
spatial
location
of
each
channel
is
derived
by
matrices.
This
separates
the
number
of
physical
channels
from
the
number
of
spatial
channels.
The
number
of
actual
channels
is
scalable,
with
three
being
the
minimum
for
horizontal
only
reproduction,
and
four
if
vertical
representation
is
also
required.
An
advantage
of
the
system
is
that
it
can
image
over
a
full
360.
A
simple
way
to
explain
it
is
as
a
3D
version
of
M-S.
To
achieve
this
requires
a
special
microphone14
and
several
stages
of
formatting
(A
to
D).
As
the
recording
and
production
of
the
signal
chain
are
separated,
it
is
possible
to
decode
into
other
formats
(eg
stereo,
Dolby
5.1).
Interest
in
the
format
remains
steady
albeit
from
a
tiny
number
of
enthusiasts.
Dolby
Surround
Sound
(particularly
Dolby
Digital)
has
seen
considerable
success
in
the
market
for
home
theatre.
Possibly
this
is
on
the
wane,
as
today
many
consumers
are
putting
their
cash
into
bigger
screens
(with
stereo
bar
loudspeakers),
rather
than
adding
more
loudspeakers.
In
any
case,
Surround
Sound
has
made
almost
no
impact
when
it
comes
to
music
recordings.
Formats
developed
for
this
such
as
DVD-A,
and
SACD
gained
no
acceptance
from
mass
consumers.
13
14
Summary
The
idea
of
stereo
goes
back
a
surprisingly
long
way.
By
the
1930s
all
the
essential
theory
was
known
and
practical
systems
built.
It
was
also
clear
by
then
that
for
the
large
space
of
a
movie
theatre
3
front
channels
were
needed,
particularly
for
speech.
Several
attempts
were
later
tried
to
improve
music
systems
by
adding
more
channels
or
deriving
additional
phantom
channels
from
the
stereo
source.
None
of
these
3,
4
or
5(.1)
channel
systems
have
succeeded
commercially.
The
reasons
seem
twofold;
market
resistance,
and
resistance
from
the
recording
industry
(not,
of
course
from
the
manufacturers).
This
does
not
necessarily
mean
that
multi-channel
systems
sound
inferior,
and
neither
is
stereo
inherently
realistic.
All
these
methods
create
an
audio
illusion
of
some
form,
and
ideas
of
how
stereo
should
be
mixed
range
from
gimmicky
to
life-like.
Originally,
Audio
Engineers
grappled
with
the
problem
of
how
to
maintain
realistic
and
accurate
positioning
of
sounds
in
the
stereo
field.
Today
the
practice
is
quite
different.
Most
signal
sources
in
a
mix
are
mono
and
are
positioned
with
a
panpot.
This
supposedly
sets
the
sounds
in
place
in
the
stereo
field.
It
is
very
easy
to
do,
and
most
Engineers
I
know
spend
a
lot
longer
on
dynamic
control
and
equalization
than
on
ensuring
that
the
stereo
image
of
a
mix
is
transferable.
Delay
is
never
used
to
try
to
position
a
sound
in
the
stereo
field.
In
fact,
FX
such
as
delay
are
often
added
to
sounds
in
the
mix
and
this
can
often
completely
undermine
the
listeners
ability
to
precisely
locate
a
sound.
An
accurate
soundstage
has
been
sacrificed
for
more
additives
in
the
mix.
Stereo
continues
to
be
the
main
format
for
music
reproduction,
having
survived
while
other
more
complex
systems
have
not.
It
is
a
relatively
simple
system
and
gives
a
good
impression
of
spaciousness,
suitable
for
enjoyable
music
listening.
In
short;
overall,
it
is
the
best
solution.
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
References
Stereo,
a
Misunderstanding,
The
Anstendig
Institute,
1982
http://www.anstendig.org/Stereo.html
The
Telephone
at
The
Paris
Opera
Scientific
American,
December
31,
1881,
422-423
Blumlein
A.
D.,
Electric
and
Musical
Industries
Ltd.,
assignee.
Improvements
in
and
relating
to
Sound-transmission,
Sound
recording
and
Sound-reproducing
Systems
(patent
394,325.
June,
1933)
https://worldwide.espacenet.com/publicationDetails/originalDocument?CC=GB&NR
=394325A&KC=A&FT=D&ND=3&date=19330614&DB=EPODOC&locale=en_EP
Steinberg
J.
C.,
Snow
W.
B.,
Audio
Perspectives
-
Physical
Factors,
1933
http://www.aes.org/aeshc/docs/bell.labs/auditoryperspective.pdf
K.
Hammer,
W.
Snow:
Binaural
Transmission
System
at
Academy
of
Music
in
Philadelphia.
Memorandum
MM-
3950,
Bell
Laboratories,
Nov.
1932
P.
W.
Klipsch,
Circuits
for
Three-Channel
Stereophonic
Playback
Derived
from
Two
Sound
Tracks,
IRE
Transactions,
Nov-Dec
1959
P.
W.
Klipsch,
Three-Channel
Stereo
Playback
of
Two
Tracks
Derived
from
Three
Microphones,
IRE
Transactions,
Mar-Apr
1959
Everest
A.,
(1975).
Handbook
of
Multichannel
Recording,
TAB
books.
Gayford
M.
L.,
(1975).
Hi-Fi
for
the
Enthusiast,
Pitman
Press,
2nd
Ed.
Scheiber
P.,
Quadrasonic
Sound
System
(patent
3,632,886,
1969)
BBC
(1974),
The
Subjective
Performance
of
Various
Quadraphonic
Matrix
Systems,
RD
1974/29.
Brice,
R.,
(2014)
Ambisonics,
http://www.pspatialaudio.com/Ambisonics%202014%20article.pdf
Gerzon
M.
(1991),
Optimal
Reproduction
Matrices
for
Multispeaker
Stereo,
91st
AES
conf.,
Appendix
1
A
stereo
reproduction
system
produces
signals
thus:
Ls
to
Le,
Ls
to
Re,
Rs
to
Le,
and
Rs
to
Re
where
subscripts
s
and
e
are
speaker
and
ear,
and
Ls
to
Le
=
L,
and
Rs
to
Re
=
R,
and
Ls
to
Re
=
L,
and
Rs
to
Le
=
R
The
two
additional
speakers
will
produce
a
difference
signal
and
are
wired
out
of
phase
so
one
will
give
the
L-R
and
the
other
the
R-L
signal.
Therefore,
at
the
listener:
L
=
L
+
R
-
R
=
L
R
=
R
+
L
-
L
=
R
Appendix
2
M
=
L
+
R
and
S
=
L
-
R
M
+
S
=
(L
+
R)
+
(L
R)
=
2L
M
S
=
(L
+
R)
-
(L
R)
=
2R
and
(M
+
S)
+
(M
-
S)
=
M
+
M
+
S
-
S
=
2M
(MS
sums
to
mono)