U.S. patent number 3,745,252 [Application Number 05/112,168] was granted by the patent office on 1973-07-10 for matrixes and decoders for quadruphonic records.
This patent grant is currently assigned to Columbia Broadcasting System, Inc.. Invention is credited to Benjamin B. Bauer.
United States Patent |
3,745,252 |
Bauer |
July 10, 1973 |
MATRIXES AND DECODERS FOR QUADRUPHONIC RECORDS
Abstract
Method and apparatus for recording four separate channels of
information on a medium having only two independent tracks, such as
a phonograph record, and method and apparatus for reproducing such
information and presenting it on four loudspeakers to give the
illusion of sound coming from four separate sources. The signals
recorded in the manner described in this application are also
reproducible on conventional stereophonic playback systems,
distributing the four separate channels to the two loudspeakers in
a manner to give a balanced and symmetrical reproduction. Two
embodiments of encoding apparatus for combining the four input
signals preparatory to recording on the two-track medium, and two
decoders, one for each of the encoding systems, are described.
Inventors: |
Bauer; Benjamin B. (Stamford,
CT) |
Assignee: |
Columbia Broadcasting System,
Inc. (New York, NY)
|
Family
ID: |
22342450 |
Appl.
No.: |
05/112,168 |
Filed: |
February 3, 1971 |
Current U.S.
Class: |
381/21 |
Current CPC
Class: |
H04S
3/02 (20130101) |
Current International
Class: |
H04S
3/00 (20060101); H04S 3/02 (20060101); H04r
005/00 () |
Field of
Search: |
;179/15BT,1G,1GP,1.4ST,1.1TD |
References Cited
[Referenced By]
U.S. Patent Documents
Other References
"Four Channels and Compatibility" by Scheiber Audio Engineering
Society Preprint, Oct. 12, 1970. .
"A New Quadraphonic System" by Hafler Audio Magazine, July
1970..
|
Primary Examiner: Claffy; Kathleen H.
Assistant Examiner: D'Amico; Thomas
Claims
I claim:
1. Encoding apparatus for matrixing four signals designated
L.sub.f, L.sub.b, R.sub.b and R.sub.f into two composite signals
designated L.sub.T and R.sub.T for recording or transmission on a
two-track medium, said apparatus comprising:
first, second, third and fourth input terminals to which said
L.sub.f , L.sub.b, R.sub.b and R.sub.f signals are respectively
applied,
first and second output terminals, and
signal transfer means connected in circuit between said input
terminals and said output terminals to transfer signals from the
input terminals to the output terminals, said signal-transfer means
including
a first summing circuit connected to add a predetermined first
portion of the L.sub.f signal from the first input terminal to a
like first portion of the L.sub.b signal from the second input
terminal to produce a first sum signal,
a second summing circuit connected to add said predetermined first
portion of the R.sub.f signal from the fourth input terminal to a
like portion of the R.sub.b signal from the third input terminal to
produce a second sum signal,
a third summing circuit connected to combine in phase opposition a
second predetermined portion smaller than said first predetermined
portion of the L.sub.f signal from the first input terminal and a
like smaller portion of the L.sub.b signal from the second input
terminal to produce a third sum signal,
a fourth summing circuit connected to combine in phase opposition
said second predetermined smaller portion than said first
predetermined portion of the R.sub.f signal from the fourth input
terminal and a like smaller portion of the R.sub.b signal from the
third input terminal to produce a fourth sum signal,
a fifth summing circuit connected to add said first sum signal to
said fourth sum signal and first all-pass phase-shifting means
operative to cause the first sum signal received by said fifth
summing means to be substantially in phase quadrature with the
fourth sum signal received by said fifth summing circuit,
a sixth summing circuit connected to add said second sum signal to
said third sum signal and second all-pass phase-shifting means
operative to cause the second sum signal received by said sixth
summing circuit to be substantially in phase quadrature with the
third sum signal received by said sixth summing circuit, and
means connecting the output terminals of said fifth and sixth
summing circuits to said first and second output terminals,
respectively, thereby to produce at said first output terminal a
composite signal L.sub.T consisting of said predetermined first
portion of said L.sub.f and L.sub.b signals and said second
predetermined smaller portion of said R.sub.f and R.sub.b signals
and wherein the R.sub.f signal is substantially in quadrature with
one of the L.sub.f and L.sub.b signals and the R.sub.b signal is
substantially in quadrature with the other of the L.sub.f and
L.sub.b signals, and to produce at said second output terminal a
composite signal R.sub.T consisting of said predetermined first
portion of said R.sub.f and R.sub.b signals and said second
predetermined smaller portion of the L.sub.f and L.sub.b signals
and wherein the L.sub.f signal is substantially in quadrature with
one of the R.sub.b and R.sub.f signals and the L.sub.b signal is
substantially in quadrature with the other of the R.sub.f and
R.sub.b signals.
2. Apparatus according to claim 1, wherein said predetermined first
portion is substantially the decimal fraction 0.924 and said
predetermined second smaller portion is substantially the decimal
fraction 0.383.
3. Encoding apparatus according to claim 1, wherein for in-phase
input signals L.sub.f, L.sub.b, R.sub.b and R.sub.f said summing
circuits and said phase-shifting means are operative to cause the
L.sub.f and L.sub.b signals in the composite signal L.sub.T and the
R.sub.f and R.sub.b signals in the composite signal R.sub.T to all
be substantially in phase, to cause the R.sub.f and R.sub.b signals
in the composite signal L.sub.T to be substantially in phase
opposition with one another, and to cause the L.sub.f and L.sub.b
signals in the composite signal R.sub.T to be substantially in
phase opposition with one another.
4. Encoding apparatus according to claim 1, wherein the said
all-pass phase-shifting means are operative to shift the phase of
said first, second, third and fourth sum signals to cause the
relative phase angles of said first, second, third and fourth sum
signals to be 90.degree., 135.degree., 45.degree. and 0.degree.,
respectively.
5. Decoding apparatus for reproducing four different signals from
two composite signals L.sub.T and R.sub.T wherein the L.sub.T
signal contains a predetermined first portion of signals designated
L.sub.f and L.sub.b and a second predetermined smaller portion than
than said first predetermined portion of signals designated R.sub.f
and R.sub.b and wherein the R.sub.f signal is substantially in
quadrature with one of the L.sub.f and L.sub.b signals and the
R.sub.b signal is substantially in quadrature with the other of
said L.sub.f and L.sub.b signals and wherein the R.sub.T signal
contains said predetermined first portion of the said R.sub.f and
R.sub.b signals and said second predetermined smaller portion of
the said L.sub.f and L.sub.b signals and wherein the L.sub.f signal
is substantially in quadrature with one of said R.sub.f and R.sub.b
signals and the L.sub.b signal is substantially in quadrature with
the other of said R.sub.f and R.sub.b signals, said apparatus
comprising:
first and second input circuits to which the L.sub.T and R.sub.T
signals are respectively applied, said input circuits including
all-pass phase-shifting means operative to shift the relative phase
of said L.sub.T and R.sub.T signals by a predetermined angle to
position corresponding components thereof either in phase or in
phase opposition to permit later addition or subtraction of said
corresponding components, and
first, second, third and fourth signal-combining networks, each
having first and second input terminals and an output terminal,
each connected to receive at its first and second input terminals a
signal from said first and second input circuits, respectively,
said first signal-combining network being operative to combine a
portion of the signal from said first input circuit equal to said
predetermined first portion with the inverse of a portion of the
signal from said second input circuit equal to said predetermined
second portion and to produce at its output terminal a first output
signal containing one of said L.sub.f and L.sub.b signals as its
predominant component,
said second signal-combining network being operative to combine a
portion of the signal from said first input circuit equal to said
predetermined first portion with a portion of the signal from said
second input circuit equal to said predetermined second portion and
to produce at its output terminal a second output signal containing
the other of said L.sub.f and L.sub.b signals as its predominant
component,
said third signal-combining network being operative to combine a
portion of the signal from said first input circuit equal to said
predetermined second portion with a portion of the signal from said
second input circuit equal to said predetermined first portion and
to produce at its output terminal a third output signal containing
one of said R.sub.f and R.sub.b signals as its predominant
component, and
said fourth signal-combining network being operative to combine the
inverse of a portion of the signal from said first input circuit
equal to said predetermined second portion with a portion of the
signal from said second input circuit equal to said predetermined
first portion and to produce at its output terminal a fourth output
signal containing the other of said R.sub.f and R.sub.b signals as
its predominant component.
6. Decoding apparatus according to claim 19 wherein said
predetermined first portion is substantially the decimal fraction
0.924 and said predetermined second smaller portion is
substantially the decimal fraction 0.383.
7. Decoding apparatus according to claim 5, wherein said input
circuits are operative to introduce a relative phase shift of
substantially 90.degree. between said L.sub.T and R.sub.T composite
signals, and wherein said first, second, third and fourth output
signals respectively contain the L.sub.f, L.sub.b, R.sub.b and
R.sub.f signals as its predominant component.
8. Decoding apparatus according to claim 5, wherein said input
circuits are operative to introduce a relative phase shift of
substantially 45.degree. between the L.sub.T and R.sub.T composite
signals, and wherein said first, second, third and fourth output
signals respectively contain the L.sub.b, L.sub.f, R.sub.f and
R.sub.b signals as its predominant component.
9. Encoding apparatus for matrixing four input signals designated
L.sub.f, L.sub.b, R.sub.b and R.sub.f into two composite signals,
said apparatus comprising:
first, second, third and fourth input terminals to which said
L.sub.f, L.sub.b, R.sub.b and R.sub.f are respectively applied,
first and second output terminals,
circuit means including means for shifting the phase of said
R.sub.f input signal at said fourth input terminal by a reference
phase-shift angle and for shifting the phase of said L.sub.b input
signal at said second input terminal by said reference phase-shift
angle plus 90.degree. and operative to couple first and second
predetermined portions, respectively, of said phase shifted R.sub.f
and L.sub.b signals to said first output terminal,
circuit means including means for shifting the phase of said
R.sub.b input signal at said third input terminal by said reference
phase-shift angle and for shifting the phase of said L.sub.f input
signal at said first input terminal by said reference phase-shift
angle plus 90.degree. and operative to couple said first and second
predetermined portions, respectively, of said phase-shifted R.sub.b
and L.sub.f signals to said first input terminal,
circuit means including means for shifting the phase of said
L.sub.b input signal at said second input terminal by said
reference phase-shift angle and for shifting the phase of said
R.sub.f input signal at said fourth input terminal by said
reference phase-shift angle plus 90.degree. and operative to couple
said first and second predetermined portions, respectively, of said
phase-shifted L.sub.b and R.sub.f signals to said second output
terminal, and
circuit means including means for shifting the phase of said
L.sub.f input signal at said first input terminal by said reference
phase-shift angle and for shifting the phase of said R.sub.b input
signal at said third input terminal by said reference phase-shift
angle plus 90.degree. and operative to couple said first and second
predetermined portions, respectively, of said phase-shifted L.sub.f
and R.sub.b signals to said second output terminal.
10. Encoding apparatus in accordance with claim 9, wherein said
first and second predetermined portions are respectively the
decimal fractions 0.383 and 0.924.
Description
BACKGROUND OF THE INVENTION
This invention relates to apparatus for recording and reproducing
four separate channels of information on a medium having only two
independent tracks, and more particularly to improved apparatus for
recording such information and for reproducing the information and
presenting it on four loudspeakers to give the listener the
illusion of sound coming from a corresponding number of separate
sources. In a system of this kind described in applicant's
co-pending U.S. application Ser. No. 44,224 filed June 8, 1970,
entitled "Quadruphonic Recording and Reproducing System" and
assigned to the assignee of the present invention, in which a
stereophonic record, which may be in the form of disc or tape,
etc., is used as the two track medium, there are recorded on the
left and right channels signals to be presented on the "left front"
and "right front" loudspeakers, respectively, together with the
signals on both channels identified with "left back" and "right
back" loudspeakers at 90.degree. out of phase relative to each
other, with the "left back" signals leading the "left front"
channel signal on the left track and the "right back" signals
leading in the "right front" channel signal on the right track.
Also described in aforementioned application is a decoding system
which accepts the two outputs from the disc record, one from each
track, and by appropriate electronic manipulation separates them
into a simulation of four independent channels, for presentation on
four separate loudspeakers, each carrying predominantly the
information contained in the orginally recorded sound channels with
attenuated information from other channels.
There being considerable current interest in multiple-channel
recording and reproduction systems of this general type, which are
termed "stereo quadruphonic" systems, other techniques have been
suggested for matrixing the four channels of information
preparatory to recording on the two-track medium, and for decoding
the information to give the desired illusion of four channels of
information upon reproduction. In one such system of which
applicant is aware, described in an Audio Engineering Society
Preprint entitled "Four Channels and Compatibility" by Peter
Scheiber, Oct. 12, 1970, the encoding matrix is as diagrammatically
illustrated in FIG. 1 of the drawings, and phasor diagrams useful
in explaining the operation of the decoder are presented in FIGS. 2
and 3. The encoding matrix of FIG. 1 is operative to encode four
separate channels of program information designated L.sub.f (left
front), L.sub.b (left back), R.sub.b (right back) and R.sub.f
(right front) into two new channels L.sub.T and R.sub.T. The four
signals are respectively applied to input terminals 2, 4, 6, 8,
each of which is connected to both of a pair of summing devices 10
and 12 as shown. The summing device 10, which may be a matrix of
operational amplifiers and resistors, is operative to add 0.924 of
signal L.sub.f, 0.924 of signal L.sub.b, -0.383 of signal R.sub.b
and 0.383 of signal R.sub.f, these factors being indicated by the
numbers adjacent the point of connection of the respective
terminals to summing circuit 10. The summing device 12, of similar
construction, sums 0.383 of signal L.sub.f, -0.383 of signal
L.sub.b, 0.924 of signal R.sub.b and 0.924 of R.sub.f. It is to be
noted that 0.383 is the sine of 22 1/2.degree. while 0.924 is the
cosine of 22 1/2.degree..
The resulting composite signals L.sub.T and R.sub.T from summing
circuits 10 and 12, respectively, may be applied to the transducers
of a recorder, which may be either a two-track magnetic tape
recorder or a stererophonic disc cutter, etc. The form of the
composite signals is shown in phasor form in FIG. 2, FIG. 2A
representing the signal L.sub.T applied to the "left" recording
track, and FIG. 2B representing the total signal R.sub.T applied to
the "right" recording track. It will be noticed that the phasor
corresponding to signal L.sub.T is made up of two equal signals
L.sub.f and L.sub.b, each 0.924 long and of the same phase, and two
signals 0.383 R.sub.b and 0.383R.sub.f, also equal in length but
directed oppositely. Likewise, the signal R.sub.T is made up of two
larger signals, 0.924R.sub.f and 0.924R.sub.b, of the same phase,
and two smaller signals 0.383L.sub.b and 0.38L.sub.f of opposite
phase. It should be noticed further that while, in general, signals
L.sub.f, L.sub.b, R.sub.b and R.sub.f are complex program signals
which cannot be represented by phasors, if we consider their
relationship on a signal frequency basis then the use of the phasor
representation is justified and it serves to better visualize the
function of the apparatus.
A recorded program comprising the signals L.sub.T and R.sub.T can
be reproduced on a conventional stereophonic player, or it can be
de-matrixed in a special resistive de-matrixing network, the
details of which are not of concern to the discussion to follow.
Suffice it to say that the de-matrixing network produces four
signals each of which is composed of a predominant signal L.sub.f,
L.sub.b, R.sub.b, and R.sub.f, respectively, together with two
additional signals of the same series at a 3dB lower level,
diluting the predominant signal.
From the discussion to follow it will be seen that the
above-described matrixing technique produces results which
seriously detract from the realism of four channel operation.
Consider, for example, the operation of the network when equal
signals, L.sub.f and R.sub.f, representing a "center-front" signal
are applied to terminals 2 and 4. It will be observed that in this
situation the signal L.sub.T becomes elongated, forming a new
signal, designated A, since 0.924 + 0.383 = 1.307. The signal
R.sub.T also becomes elongated, for the same reason, forming a new
signal, designated B, which is also 1.307 long. However, if one
wishes to record a "center-back" signal by applying equal signals
L.sub.b and R.sub.b to terminals 4 and 6, respectively, then the
signal L.sub.T becomes foreshortened by a factor 0.924 - 0.383 =
0.541, forming a new signal C, and the composite signal R.sub.T is
likewise foreshortened, forming a new signal D, which is also 0.541
long. It is seen, therefore, that the efficiency of recording a
signal applied to the "back" terminals differs from the efficiency
of recording a signal applied to the "front" terminals; this has
been considered to be a serious deficiency of this matrix.
In an attempt to improve some of the characteristics of the
just-described matrix, it is suggested in the aforementioned Audio
Engineering Society Preprint that a relative phase shift of
90.degree. be introduced between the signals L.sub.T and R.sub.T as
by inserting all-pass shift networks 14 and 16 between the output
terminals to produce a new set of signals L.sub.T ' and R.sub.T '
which are then applied to recorder. While the inclusion of this
relative 90.degree. phase shift improves the performance from some
points of view, it causes significant deterioration in other
respects. For example, upon application to the matrix of equal
inputs L.sub.f and R.sub.f (as one would do to obtain a
"center-front" signal), the two new phasors which are obtained,
namely, A' and B' in FIGS. 3A and 3B, respectively, are at
90.degree. with respect to each other, thus resulting in a fuzzy
and indefinite virtual image. Furthermore, the basic problem of
partial cancellation of the "center-back" signal is not overcome.
In order to obtain a sharp virtual image upon stereophonic replay
it is important that the two phasors applied to the loudspeakers be
as nearly in phase as possible, and furthermore, it is desirable
that the matrix permit relatively uniform transmission of all
signals applied individually or jointly to its input terminals.
SUMMARY OF THE INVENTION
It is the principal object of the present invention to provide an
improved method and apparatus for matrixing four signals which
results in a vastly superior arrangement of the matrixed signals
and which is free of the aforementioned defects of the prior art
matrix illustrated in FIG. 1.
Briefly, this object is obtained with a matrix which places the
four original signals in appropriate phase relationship by the use
of suitable all-pass networks prior to summation into the two final
signals L.sub.T and R.sub.T. More specifically, to achieve this
object with a minimum number of all-pass networks, the L.sub.f and
L.sub.b signals are twice summed, 0.924 of each being added
together in a first summing network, and 0.383 of L.sub.f and
-0.383 of L.sub.b being added in a second summing network.
Similarly, 0.924 of each of signals R.sub.b and R.sub.f are added
in a third summing network and 0.383 of signal R.sub.b and -0.383
of signal R.sub.f are summed in a fourth summing network. The
outputs of the first and third summing circuits are summed in a
fifth summing circuit after introducing a 90.degree. phase-shift in
the output of the first sum signal relative to that of the third
sum signal, and the outputs of the two remaining summing circuits
are similarly combined in a sixth summing network, the outputs of
the fourth summing network being subjected to a 90.degree.
phase-shift with respect to the output from the second summing
network. The outputs of the fifth and sixth summing networks are
the composite signals L.sub.T and R.sub.T, respectively. These
composite signals may be applied to a stereophonic loudspeaker
system directly, or recorded on a tape recorder or a stereophonic
disc recorder for subsequent replay on a two-channel stereophonic
system. By combining the signals in this manner, the "front" and
"back" signals have symmetry and the phase relationship of the
"center" signal is improved.
In another embodiment of the invention, in which essentially the
same matrix is utilized, the front-to-back symmetry is altered
(which in some applications is desirable by introducing a different
phase-shift between the ultimately summed signals than was used in
the just-described embodiment).
In both embodiments there is an enhanced measure of performance
during reproduction on a stereophonic system, and at the same time
provides improved performance on a four-channel de-matrixed
quadruphonic system.
BRIEF DESCRIPTION OF THE DRAWINGS
An understanding of the foregoing and additioal objects and aspects
of the invention may be gained from consideration of the following
detailed description, taken in conjunction with the accompanying
drawings, in which:
FIG. 1 is a schematic diagram of the prior art system to which
reference has already been made;
FIGS. 2 (A and B) and 3 (A and B) are phasor diagrams to which
reference has been made in the discussion of the system of FIG.
1;
FIG. 4 is a schematic diagram of an encoding matrix embodying the
present invention;
FIG. 5 (A and B) are phasor diagrams useful in explaining the
operation of the circuit of FIG. 4;
FIG. 6 is a schematic diagram of a system for decoding signals
recorded with the encoder of FIG. 4;
FIG. 7 is a schematic diagram of a variation of the encoder of FIG.
4;
FIG. 8 (A and B) are phasor diagrams useful in explaining the
operation of the encoder of FIG. 7; and
FIG. 9 is a schematic diagram of a portion of a reproducing system
for decoding signals recorded with the encoder of FIG. 7.
DESCRIPTION OF THE PREFERRED EMBODIMENT
A preferred form of encoder embodying the present invention is
illustrated in FIG. 4 and is similar in some respects to the
encoder described in applicant's aforementioned co-pending
application Ser. No. 44,224, differing therefrom in the manner in
which the signals are summed and in the phase-shift constant used.
The four separate channels of program information, again designated
L.sub.f, L.sub.b, R.sub.f and R.sub.b for left front, left back,
right front, and right back, respectively, are applied to input
terminals 18, 20, 22 and 24 of the encoder. Terminals 18 and 20 are
both connected to a pair of summing devices 26 and 28, and input
terminals 22 and 24 are both connected to another pair of summing
devices 30 and 32. Summing devices 26 and 32 are of similar
construction, both being operative to add the decimal fraction
0.924 of each of the two signals applied thereto. Summing devices
28 and 30 are alike but differ from summing devices 26 and 32 in
that 0.383 of one of the signals applied to each is added to -0.383
of the other applied signal. More specifically, in the case of
summing device 28, 0.383 of the signal L.sub.f is added to -0.383
of the signal L.sub.b, and in device 30, 0.383 of the signal
R.sub.b is added to -0.383 of the signal R.sub.f.
The outputs of the four summing devices 26-32 are applied to
all-pass phase-shift networks 34, 36, 38 and 40, respectively. The
networks 34 and 40 each introduce a phase-shift of (.psi. +
90.degree.) and the networks 36 and 38 each introduce a phase-shift
(.psi. + 0.degree.) into their respective circuits, the reference
angle .psi. being arbitrarily chosen, the only requirement being
that this reference angle is substantially the same in any one
encoder or decoder. It should be noted that although a convention
that these phase-shift angles are lagging has been observed,
leading phase-shifts could also be employed, as long as the same
convention is observed in any one encoder or decoder. The
respective phase-shift networks introduce the differential phase
angles shown; that is, networks 34 and 40 introduce a phase-shift
of 90.degree. in the signals from summing devices 26 and 32
relative to the signals from summing devices 28 and 30. The outputs
of phase-shift networks 34 and 38 are added in equal proportions in
an additional summing device 42, and the output signals from
phase-shift networks 36 and 40 are similarly added in summing
device 44. Summing devices 42 and 44 produce at output terminals 46
and 48, respectively, the final composite output signals L.sub.T
and R.sub.T which may be applied to a stereophonic loudspeaker
system directly, or first recorded on a tape recorder or a
stereophonic disc recorder for subsequent replay on a two-channel
stereophonic system.
Referring now to FIG. 5, the resulting voltages L.sub.T and R.sub.T
are displayed in phasor form in FIGS. 5A and 5B, respectively. It
will be noted that the voltage L.sub.T is comprised of two equal
relatively large signals, 0.924L.sub.f and 0.924L.sub.b, which are
directed in the same phase, and two equal smaller signals,
0.383R.sub.b and 0.383R.sub.f which are oppositely directed, but
which also are at 90.degree. with respect to the signals L.sub.f
and L.sub.b. Likewise, the signal R.sub.T is composed of two
relatively large signals, 0.924R.sub.f and 0.924R.sub.b, which are
in phase, and two smaller signals, 0.383L.sub.f and 0.383L.sub.b,
which are directed in opposite phase, and at the same time are at
90.degree. to the larger signals R.sub.f and R.sub.b. This
90.degree. relationship between the larger and smaller signals
gives the present decoder a significant advantage in that if now a
front-directed signal consisting of equal measures of signals
R.sub.f and L.sub.f is applied to the matrix, the resulting signals
will be as denoted by the phasors A and B, shown in dashed lines in
FIGS. 5A and 5B, respectively, which, it will be noted, are at
45.degree. relative to each other. This does not meet fully the
earlier stated desirable condition that these phasors be in phase,
but the results are much improved from that attainable in the prior
art device in which the corresponding phasors are at 90.degree. to
each other (FIG. 3). At the same time, if one applies a
back-directed signal, by applying equal measures of L.sub.b and
R.sub.b to terminals 20 and 22, respectively, the resulting total
signals are shown by the dashed-line arrows C and D in FIGS. 5A and
5B, respectively. It will be observed that phasors C and D are of
the same magnitude as the A and B phasors obtained when only a
"front-center" signal is applied. Thus, the encoder of FIG. 4 has
symmetry front-to-back and provides an improvement with respect to
the in-phase tendency of the center signal.
One important consideration in connection with the form of the
invention illustrated in FIG. 4 is that the phasor pairs A and B
are of equal magnitude, but of opposite relative phase positions
compared to the phasors C and D which result from the application
of "back-center" signals. This allows the advantageous possibility
during decoding to specifically distinguish between "front-center"
and "back-center" signals which other decoding schemes do not
allow. It should be noted that the signs of the summation network
30 can be reversed without significant deterioration of the
operation of the encoder, and such reversal would result in the
signals L.sub.T and R.sub.T being fully in phase for the
application of either the "front-center" or the "back-center"
signals. However, such an encoder would not allow distinguishment
in the decoding process between the "front-center" and the
"back-center" signals.
It is significant to note that the step of summing the signals
L.sub.f, L.sub.b, R.sub.b and R.sub.f with the summing networks 26,
28, 30 and 32 prior to phase-shifting with the networks 34, 36, 38
and 40 results in an economy of circuitry. Instead of employing
these first-mentioned summing operations, one could provide, for
example, eight separate phase-shifting networks followed by
summation networks 42 and 44 at which the four appropriate phasors
are summed in the required proportion. By using the four initial
summing operations, only four phase-shift networks are required in
the encoder in place of eight.
Referring now to FIG. 6, there is illustrated one form of apparatus
for decoding the stereophonic signal encoded with the encoder of
FIG. 4 into four signals which carry predominantly the original
information (albeit diluted with the information from two adjacent
channels) for presentation over a four loud-speaker system. The two
signals L.sub.T and R.sub.T are derived from the record medium by a
suitable transducer (e.g., a conventional stereophonic pickup in
the case of a disc record) and are applied to terminals 50 and 52,
respectively, of the decoder apparatus. In order to position these
signals properly for de-matrixing, the signals L.sub.T and R.sub.T
are applied to differential phase-shift networks 54 and 56,
respectively, which differentially alter the position of the
phasors L.sub.T and R.sub.T by 90.degree., in accordance with the
teachings of applicant's aforementioned co-pending application Ser.
No. 44,224. The relative orientation of the two signals after
phase-shift are as depicted in the phasor diagrams associated with
the output terminals of phase-shift networks 54 and 56. These
signals are applied to each of four summing devices 58, 60, 62 and
64 in the proportion indicated in the blocks representing the
summing devices. That is, in summing device 58, 0.924 of the
L.sub.T signal is added to -0.383 of the R.sub.T signal, in summing
device 60, 0.924 of L.sub.T is added to 0.383 of R.sub.T, in
summing device 62, 0.383 of L.sub.T is added to 0.924 of R.sub.T,
and in summing device 64, -0.383 of L.sub.T is added to 0.924 of
R.sub.T. Without going into the geometry, which is believed will be
self-evident, summation of the signals in these proportions
produces four new signals L.sub.f ', L.sub.b ', R.sub.b ' and
R.sub.f ' at the output terminals 66, 68, 70 and 72 of summing
devices 58, 60, 62 and 64, respectively, the phasor diagrams of
which are depicted adjacent their respective terminals. It will be
noticed that these phasors have respective predominant signals
L.sub.f, L.sub.b, R.sub.b and R.sub.f in combination with flanking
signals from adjacent terminals. This is considered by some to be a
preferred way of forming a quadruphonic signal. By contrast, if one
were to reverse the signs of summation, say in the summing network
30, as herein before described, this would invert the small phasors
in one of the signals portrayed in FIGS. 5A or 5B -- for example,
causing the position of small phasors 0.383R.sub.b and 0.383R.sub.f
to be reversed. It will be found that under this circumstance upon
de-matrixing, each of the diluted signals would have a predominant
signal in combination with two opposite-end signals, which
according to some listeners produces less desirable results.
It is to be noted that while the phasors in FIG. 5 are tagged with
signs and phase angle positions, this practice has not been
followed in FIG. 6 in the interest of avoiding confusion and
unnecessary repetition. In FIG. 6, rather, sign and phase are are
shown simply by the relative position of a phasor in the phasor
diagram.
After de-matrixing, it may be desirable to place the principal
phasors L.sub.f, L.sub.b, R.sub.b and R.sub.f appearing at
terminals 66, 68, 70 and 72 in some other phase orientation
relative to each other; for example, it may be desired to have them
all in phase. This may be accomplished by applying the signals to
respective all-pass phase-shift networks 82, 84, 86 and 88, the
latter two of which introduce a relative phase-shift of 90.degree.
relative to the other two. The signals delivered by the phase-shift
networks may then be amplified by respective gain control
amplifiers 90, 92, 94 and 96, the gain of each of which may be
controlled by applying a control signal to its control electrode
and applied to respective loudspeakers 98, 99, 100 and 101.
Should it be desired to "enhance" the quadruphonic illusion of the
reproduced signals, the gains of amplifiers 90, 92, 94 and 96 may
be controlled by a control and switching logic network 102 in
response to signals derived from the output terminals of
phase-shifters 54 and 56 in the manner described in co-pending
application Ser. No. 44,196 filed June 8, 1970 by applicant and
Daniel W. Gravereaux, and assigned to the assignee of the present
application. Briefly, the control and switching logic 102 processes
the signals appearing at the output terminals of phase-shifters 54
and 56 through an automatic gain control circuit which maintains
them at a constant predetermined level. They are then de-matrixed
in a circuit similar to that formed by summing circuits 58, 60, 62
and 64, the resulting four signals are summed with an appropriate
time-constant network, and the sum signal fed back to the automatic
gain control circuit so as to maintain the sum substantially
constant. The individual signals are combined through linear
additions and subtractions to produce control signals at the output
lines 104, 106, 108 and 110 which are respectively connected to the
gain control terminals of amplifiers 90, 92, 94 and 96. The system
is designed to apply control signals to the gain control amplifiers
to increase the gain of the channel containing the instantaneously
dominant signal and to reduce the gain of the other channels to
give a substantially perfect illustion of four separate independent
sources of sound. As the sound diminishes in the channel first
identified and another sound appears on a different channel, the
logic circuit functions to rapidly attenuate the gain in the first
channel and to increase the gain in a different channel. In this
manner, it is possible to produce a high level of simulation of
four discrete channels.
Not only are these four channels produced in a highly realistic
manner, but because of the symmetry of the phasors, it is possible
to 37 pan" a signal between any two adjacent pairs of input
terminals, i.e., L.sub.f - L.sub.b ; L.sub.b - R.sub. b ; and
R.sub.b - R.sub.f ; and R.sub.f - L.sub.f, without significant
change of level during the panning operation. Furthermore, if the
record is played on a conventional two-track stereophonic system,
the signals appear to be relatively crisp and well defined.
Referring now to FIG. 7, there is shown another embodiment of the
invention by which the front-back symmetry of the signals is
changed, which for some applications might be preferable. The
general configuration of the encoder resembles the system of FIG.
4, but as will be seen from the discussion to follow the signals
are combined in a different way. The four original input signals
L.sub.f, L.sub.b, R.sub.b and R.sub.f at the terminals 112, 114,
116 and 118, respectively, are applied to four summing devices 120,
122, 124 and 126 in the manner and in the proportions indicated by
the decimal fractions appearing on the summing devices. More
specifically, 0.383 of each of signals L.sub.f and L.sub.b are
added in summing device 122 and 0.924 of each of them are added in
summing device 124. Similarly, 0.383 of each of the R.sub.b and
R.sub.f signals are summed in summing device 120, and 0.924 of each
of them are summed in summing device 126. The outputs of summing
devices 120, 122, 124 and 126 are connected to respective all-pass
phase-shift networks 128, 130, 132 and 134 which produce relative
phase-shifts to the signals transmitted thereby of 0.degree.,
45.degree., 90.degree. and 135.degree.. The outputs of phase-shift
networks 128 and 132 are added in equal proportions in summing
circuit 136, and the outputs of phase-shift networks 130 and 134
are similarly added in equal proportions in summing device 138. The
composite signals L.sub.T and R.sub.T appearing at the output
terminals 140 and 142 of summing devices 136 and 138, respectively,
may be recorded on two-track magnetic tape or on a stereophonic
disc in conventional manner for subsequent replay over a
stereophonic system, or they may be de-matrixed into four signals
as hereinafter described.
Phasor diagrams of the composite signals L.sub.T and R.sub.T are
displayed in FIGS. 8A and 8B, respectively. For convenience, these
phasor diagrams are presented in reference to a geometrical system
of coordinates the axes of which are marked 0.degree., 45.degree.,
90.degree. and 180.degree., etc., albeit the actual electrical
phase will be drawn with respect to an imaginary electrical axis
which is labeled "O' axis." Thus, in FIG. 8A, the signal
0.383R.sub.b, which undergoes the minimum phase-shift .psi., is
shown as being coincident with the O' axis, while the 0.924L.sub.b
signal, which undergoes a relative phase lag of +90.degree., is
shown lagging behind the phasor 0.383R.sub.b by 90.degree., which
places it at 22 1/2.degree. to the geometrical 90.degree. axis.
Similarly, in FIG. 8B, the signal 0.383L.sub.f, which undergoes the
phase-shift .psi. +45.degree., is displaced 45.degree. from the O'
axis, while the 0.924R.sub.f signal, which undergoes a relative
phase lag of +90.degree., is shown lagging behind the phasor
0.383L.sub.f by 90.degree., which places it at 22 1/2.degree. to
the geometrical 90.degree. axis.
One important advantage of the matrixor on FIG. 7 will now be
demonstrated, Let it be assumed that a "center-front" signal is
applied to the matrix by applying equal L.sub.f and R.sub.f signals
to terminals 112 and 118, respectively. It will be seen from
reference to FIG. 8A that in this case the signal L.sub.T is
composed of 0.924 parts of L.sub.f and 0.383 parts of R.sub.f which
resolve into the dashed-line phasor A, which it will be noted,
lines up precisely with the geometrical 90.degree. axis. At the
same time, the total signal R.sub.T shown in FIG. 8B is comprised
of 0.924 parts of R.sub.f and 0.383 parts of L.sub.f which form the
dashed-line phasor B, which has a length of unity and which also
lines up precisely with the geometrical 90.degree. axis. Therefore,
the two phasors produced at terminals 140 and 142 are equal and
in-phase and will, therefore, produce a precise and sharp center
signal.
The situation is different, however, when a back-center signal is
applied to the matrix by applying equal in-phase signals L.sub.b
and R.sub.b to terminals 114 and 116. In this case, the output
L.sub.T signal, shown in FIG. 8A, is comprised of 0.924L.sub.b and
0.383R.sub.b which is aligned with the 45.degree. geometrical axis.
The R.sub.T signal, on the other hand, shown in FIG. 8B, consists
of 0.924 portions of R.sub.b and 0.383 parts of L.sub.b, which
resolve to form the phasor D, also of unity length but aligned with
the 135.degree. geometrical axis. It will be observed that the
phasors C and D are at 90.degree. with respect to each other, and
as was noted earlier, it is known that this will produce a fuzzy
image upon replay over a stereophonic system. Thus, with the
encoder of FIG. 7, application of a "center-front" signals produces
a sharp image and application of a center back signal produces a
fuzzy image. This turns out to be a highly desirable method of
differentiating front from back when the signals are reproduced
over a two loudspeaker stereophonic system. In any case, since all
the phasors A, B, C and D have the same unity amplitude, the
strength of the center front and center back signals is the same,
which also is as it should be.
Referring now to FIG. 9, there is illustrated a de-matrixor for
decoding the signals encoded by the system of FIG. 7. The composite
signals L.sub.T and R.sub.T are applied to terminals 144 and 146,
respectively, and thence through respective all-pass phase shift
networks 148 and 150 which introduce a relative 45.degree. phase
shift between the signals. This phase-shift action causes the two
signal phasors to be aligned with respect to each other in such a
way that they can be de-matrixed. The phasor diagrams of signals
L.sub.T and R.sub.T, which correspond to FIGS. 8A and 8B,
respectively, are reproduced for convenient reference adjacent
terminals 144 and 146, respectively, except that the markings
referring to the lag phase angle and sign have been eliminated for
the sake of clarity. After passing through the phase-shift
networks, the phasor diagrams of the resulting signals, designated
L.sub.T ' and R.sub.T ', are as shown associated with the output
terminals of the respective phase-shift networks. These latter
signals are de-matrixed by applying them to four summing networks
152, 154, 156 and 158 in the proportions indicated in the circles
representing the summing devices. The operation of the summing
devices, which is generally similar to that described in connection
with FIG. 6, produces signals at the respective output terminals
160, 162, 164 and 166 of the summing devices 152-158 which are
predominantly L.sub.f, L.sub.b, R.sub.b and R.sub.f, respectively.
The phasor diagrams of the four signals are depicted in association
with the respective terminals, it being noted that each predominant
phasor is combined with the two flanking phasors from adjacent
channels, thereby providing the same quadruphonic image action as
in the embodiment of FIG. 5.
It will be understood that the output terminals of the de-matrixor
network of FIG. 9 may be connected to additional phase-shift
networks to place the principal phasors in any desired phase
orientation relative to each other in the manner described in the
system of FIG. 6. It will be further understood that the
just-described phase-shift networks (which should be regarded as
optional) may be connected to corresponding gain control amplifiers
for application to respective loudspeakers, and arranged to be
controlled by a logic and switching network, again as described in
connection with FIG. 6 and in the aforementioned co-pending
application.
It will be evident that the herein described modification of the
encoding or matrixing circuit disclosed in applicant's
aforementioned co-pending application produces an enhanced measure
of performance when the record medium is reproduced on a
stereophonic system, and at the same time provides improved
performance on a four-channel de-matrix quadruphonic system. The
latter improvement is achieved because the encoding matrixes
described herein do not possess directional ambiguity of the kind
inherent in the matrixor of FIG. 1; in the prior art circuit of
FIG. 1 there is a directional ambiguity which makes it impossible
to be sure whether the encoded signals had originated between the
front or back pairs of loudspeakers. Examination of the topology of
the FIG. 1 matrix suggests that at least two points of directional
ambiguity are bound to exist, and as a matter of fact, exhibit a
continuous and broad region of ambiguity, extending along the two
paths from the left back to the right back loudspeaker. In other
words, a sound panned between the input terminals corresponding to
these loudspeakers will appear to travel from the left back to the
left front loudspeakers, then to the right front, and next to the
right back loudspeakers. The present matrixes do not possess such
an ambiguity.
* * * * *