IL37872A - Quadruphonic reproducing system - Google Patents

Description

QUADRUPHONIC REPRODUCING SYSTEM Columbia Broadcasting System, Inc 37871/2 This invention relates to apparatus for recording and reproducing four separate channels of information on a medium having only tvo independent tracks, and more particularly to apparatus for reproducing such information and presenting it on four loudspeakers to give the listener the illusion of sound coming from a corresponding number of separate sources* The tvo tracks may be provided by any one of several available two-track systems, such as two-track tape, or a sterophonic disc record. In a system of this kind described in copending Israel Application Ser. No. 37418 in which a stereophoni 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 channel8 signals to be presented on the "left front" and "right front" loudspeakers, respectively, together with signals on both channels identified with "left back" and "right backQ loudspeakers at 90° out-of-phase relative to each other, with the "left back" signal leading in the "left" channel and the "right back" signal leading in the "right" channel. Also described in the aforementioned co-pending 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 our independent channels, for presentation on four separate loudspeakers, each carrying predominantly the C-1061 information contained in the originally recorded sound channels with attenuated information from other channels. The present invention is directed to an improved decoding apparatus of the general type described in the aforementioned co-pending application.

To better understand the nature of the problem to be overcome by the improved decoding apparatus of the present invention, the technique and apparatus for recording the four channels of information on a two-track : stereophonic record, described in detail in the afore- ! mentioned application, will be briefly described with reference to FIG. 1 of the accompanying drawings. In this figure, the encoding apparatus is diagrammatically illustrated as a block 8 which includes a number of all-pass networks (characterized as ψ-networks in the aforementioned application) which provide uniform transmission of all the important frequencies of the audio spectrum while at the same time providing a fixed differential phaseshift between the various networks so that a fixed phase difference relationship exists between the output voltages. An understanding of the details of these all-pass circuits being unnecessary for an understanding of the present invention, and, moreover, since they are described in detail in the aforementioned application, only the function of the encoding apparatus , as it applies to the present invention, will be described. The encoding apparatus includes four input terminals 10, 12, 14 and 16 for receiving four separate signals which C- 061 designations signify the locations in a listening area of the four loudspeakers on which the signals are intended for ultimate presentation. The characteristics of the all-pass networks are' such that if input voltages L . L„ , R and R_ of the same amplitude and F B B F frequency are applied in succession to their respective input terminals, the two output voltages and R^, respectively appearing at output terminals 18 and 20 display the amplitude and phase relationship, referred to the input signals (phase-shifted by frequency dependent reference phase angle ψ) shown by the phasor diagrams adjacent the respective output terminals. It is demonstrated in the aforementioned application that if the resulting voltages and are recorded on a two-track medium, such as a stereophonic disc record, the resulting record is fully compatible with conventional stereophonic, as well as monophonic, players, with all the sounds appearing in full quality and amplitude over the two-channel stereophonic system.

. " It has been found that if the sounds are represented predominantly by a set of two musically related and similar, albeit incoherent, signals L and F R f the decoder described in the aforementioned F application not only turns "on" the "front" loudspeakers, as it should, but also tends to some degree to turn "on"-the "back" loudspeakers, obviously an undesirable result. It is the primary object of the present invention, therefore, to provide a decoder which gives better separation " " C- SUMMARY OF THE INVENTION Briefly, the foregoing object is attained by providing means in the decoder of the reproducing apparatus for controlling the gains of the amplifiers associated with the four separate loudspeakers with signals derived by appropriate auto correlation of signals closely resembling the four independent .channels being presented on the loudspeakers. More specifically, the two signals predominantly carrying left front (L ) F and right front (R ) sound information are multiplied F and the product integrated, and the two signals predominantly containing left back (L ) and right back B (R ) information are multiplied together and their B product integrated. Because of the relative phase relationships of the voltages representing the four signals, the two signals delivered by the integrators correlate differently, particularly when the integration is carried out over a time period corresponding to approximately 150 Hz. The signals delivered by the integrators are rectified and applied in parallel, but with opposite polarity, to a pair of subtracting circuits that are operative to product a pair of gain control signals. The signal delivered by one of the subtracting circuits controls in unison the gains of the gain control amplifiers associated with the "left front" and "right front" loudspeakers , and the signals delivered by the other subtracting circuit is applied in parallel to the gain control amplifiers associated with "left back" and - 'X.. channels together with the use of autocorrelation techniques improves the separation between the "front" and "back" sounds to enhance the realism of four-channel simulation. The improved decoder is adaptable for operation in any encoder-decoder system in which four signals are encoded for recording on a two-track medium and the decoder decodes them for display over four loudspeakers in a manner to cause the predominant information in the four signals, together with attenuated information from others of the si over corresponding loudspeakers.

BRIEF DESCRIPTION OF THE DRAWING An understanding of the foregoing and additional aspects of this invention may be gained from consideration of the following detailed description, taken in conjunction with the accompanying drawing, in which : FIG. 1 is a schematic diagram of encoding apparatus -to which reference has already been made; FIGS. 2A and 2B together constitute a schematic diagram of decoder and playback apparatus embodying the invention; FIG. 3 is a curve illustrating the transmission characteristics as a function of frequency of wave-shaping networks embodied in the system of FIG. 2; and FIG. 4 is a curve illustrating the charactereisti of other wave-shaping networks embodied in the system of FIG. 2. 37872/2 applicant's a orementioned Israel Application Ser. No. 37418 certain details of vhich, in turn, are described in U.S.

Patent No. 3,708,631. The disclosures of these applications are incorporated herein by reference to the extent necessary for understanding the operation of certain aspects of the herein described decoder; however, the description to follow; is believed to be sufficiently complete to enable ones skilled in the art to understand this operation without recourse to the co-pending applications* Referring now to FIG. 2, the decoder of the present invention is shown in the context of reproducing information recorded on a disc record 22 mounted for rotation on a turntable 24 in the usual manner. The record is played with a stereophonic phongraph pickup 26 of conventional design having two output circuits 28 and 30 which respectively carry the output signals transduced from the "left" and "right" channels of the record* The "left" and "right" channel voltages are portrayed adjacent their respective output circuits* and4t will be seen from comparison with FIG. 1 that these voltages correspond to the voltages and produced by the encoding apparatus of FIG. 1. Should it be necessary in the replication of the voltages and to amplify the signals, they may be applied to respective amplifiers 36 and 38, which preferably include the necessary amplitude-frequency modifying networks C-1C61 for restoring the frequency balance that may have been used in recording the record. It will be evident that the apparatus described thus far is completely conventional and well understood by ones skilled in the art.

Essential to the decoding of the outputs L • T and R to provide a simulation of the original sounds T is the shifting of the signal from the "left" channel by 90°· relative to the signal on the "right" channel.

To this end, the signals from amplifiers 36 and 38 appearing at their respective output terminals 37 and 39 are respectively applied to corresponding all-pass phase-shift networks 40 and 42, the latter providing a reference (lagging) phase shift of , and network 40 providing a lagging phase shift of + 90°. The value of (which generally is a function of. frequency) may be the same as the phase shift used in the encoding apparatus 8 (FIG. 1) , or it may be different, the principal requirement being that the reference phase angle "be the same in both of networks 40 and 42.

The phase-shift networks 40 and 42 transform their respective input signals and into two new signals and R^ which are very nearly equal to the signals and R^ except that they are displaced at 90° with respect to each other. This is illustrated by the phasor diagrams 44 and 46 in which the phasors of the signal are in the same relative position as in the R signal, but the vectors of the L' signal T ° relative effects have been disregarded in the phasor diagrams. it will be observed that these two voltages contain predominant and voltages, with the former also including .707 of each of Ι and R^, and the latter further including .707 of each of R1 and L'.

B B The signals delivered by the all-pass networks 40 and 42 are further processed to derive two additional voltages containing predominant and information.

The former is derived by. summing the two signals from networks 40 and 42, each after multiplication by the factor .707, in a summing circuit 48 of known configuration. A signal containing predominant "right back" information is derived in a similar way by subtracting one of the signals from the other again after multiplication of each by the factor .707. As illustrated by the phasor diagrams 52 and 54, the voltages produced by the operation of circuits 48 and 50 now contain predominant I#B and information.

The four voltages represented by phasor diagrams 44, 46 52 and 54 are applied over conductors 56, 58, 60 and 62, respectively, to corresponding gain control amplifiers 64, 66, 68 and 70 and thence to corresponding loudspeakers 72, 74, 76 and 78. If necessary in a given system application, additional amplifiers {not shown) may be provided between the gain control amplifiers and their respective loudspeakers. The loudspeakers are positioned in a listening area (represented by the dotted line enclosure) with the loudspeaker carrying predominantly of the listening area, the loudspeaker predominantly carrying the signal L' at the left back corner , and ' lithe loudspeaker predominantly carrying the signal R' at the right back corner of the listening area.

The arrangement thus far described is similar to the decoding apparatus described in the aforementioned co-pending applications and as mentioned therein produces a good similitude of the original four channel sound even if the gain control amplifiers are maintained in a quiescent, or normal "gain state, or even if they are replaced with ordinary linear amplifiers. This result obtains because the predominant voltage in each channel is- 3dB higher in level than the other two voltages, and corresponds to the signal originally intended for reproduction at the particular location. However, the realism is enhanced in the systems described in the previous applications by providing a logic and switching circuit which senses whether or not a particular channel should be increased or decreased in gain so as to emphasize those sounds which momentarily predominate in particular channels. While the action of the logic employed in the systems of the earlier applications greatly enhances the simulation of four independent channels, under* certain conditions of operation there tends to be a deterioration in the simulation. The present invention is concerned with an improved form of logic which overcomes the shortcomings of the previous circuits and consistently achieves the enhanced realism of four channel simulation. c-ioei To better understand the problem to which the present invention is directed, i't is convenient to assume that the two signals encoded on the disc record consist of only two signal voltages , L and R , which are shown by F F heavier lines than the 3dB-down components in the phasor diagrams presented in FIG. 2. Further, in the analysis to follow the two signals are considered to be musically similar and in compass, but mathematically they are incoherent. Examples of such signals would result from two persons preforming a two-part musical selection in unison; while the melody or melody and accompaniment would be in proper tempo and compass, the mathematical correlation, consisting of multiplication and integration over a period of time equal to the reciprocal of the lowest frequency, would tend to be reduced to zero. Only in the case of the two signals being identical would the correlation take on a finite value. Now, it is noted that when the signals have passed through networks 40 and 42, respectivel they appear as signals L' and R' at 90° with respect to F F their previous positions, so that eveh if they were originally indentical, after phase shifting they would no longer show coherency. The reason for this is that the product of a sine and cosine function over one cycle is precisely zero.

After the voltages L' and R* in phasor diagrams F F 44 and 46 are acted upon by summing networks 48 and 50 they appear as .707. R1 and .707 L' in both of phasor F F diagrams 52 and 54. Now, it should be noted, if the C-1061 ' diagrams 44 and 46 would be 3dB greater than the magnitude of the corresponding signals in the phasor diagrams 52 and 54 and thus readily identifiable as to magnitude by logic circuitry of the aforementioned applications so as to cause the appropriate gain control amplifier to turn "on" or "off" as the case may be.

However, should the voltages L' and R' be in compass and F F appear almost simultaneously, as they tend to do during a performance by two musicians playing different parts of the same musical selection, then the R.M.S. value of j i the sum of the phasors 52 and 54 is the same as the s value of the corresponding phasors in diagrams 44 and 46.

In this circumstance, the logic of the system described Israel ' 37418 in/application Ser. No. Aje-r9-6- becomes confused and tends to turn all loudspeakers "on" whereas what is desired is that only the "front" loudspeakers 72 and 78 be turned "on" .

It will have been observed from the foregoing example, in which only the voltages L' and R' are present, F F that while the voltages in the phasor diagrams 44 and 46 are uncorrelated , the corresponding voltages in the phasors 52 and 54, namely, .707R* and .707L' are individually F F either positively or negatively correlated. This can be seen from the fact that the arrow labeled 707R1 in F diagram 52 is parallel to, but in the opposite direction, to the corresponding arrow in diagram 54; consequently, upon multiplication and integration the output of these two phasors would be a finite negative number . Although C-1061 points in the same direction and, accordingly, multiplication and integration of these two signals would produce a finite positive number. It will be appreciated that if both the signals L' and R* had F F equal long-term average magnitudes, a long-term integration would reduce to zero, but this situation almost never occurs for short-term integration because even if the musicians are playing the same instrument in unison .there always tends to exist lapses or differences in pitch, tempo, attack, decay, etc. As a result, each of individual phasors .707R* and .707L' integrated over F * the short-term would tend to take on its negative or positive finite value, while the corresponding phasors L^, and R,, being incoherent even over the short-term, tend to take on a zero value, even for short-terra integration, provided the integration extends at least over a period equal to the inverse of the lowest frequency being analyzed. The improved logic now to be described is based upon the recognition of the just-described* properties of the signals and application of the principles to a practicable implementation.

In accordance with the invention, the instantaneous amplitudes of the signals delivered to the four loudspeakers are controlled by logic circuitry contained within the dashed line enclosure 87 in such a manner that a listener is given a substantially perfect illusion of four separate independent sources of sound. The two signals represented by phasor diagrams 44 and 46 delivered ' 37872 2 respectively, vhich are designed to cut off all information belov approximately 150 Ha. While the cutoff frequenc may be altered over a very vide range vithout seriously a fecting the operation of the circuit to follow, it has been observed that very little directional information is contained at frequencies below roughly 150 Hz, and that by eliminating all frequencies belov this value permits the short-term integration in the analyzer (to be described) to be carried out over an interval of only approximately 6.7 milliseconds. The transmission characteristics of the networks 92 and 94 are preferably as shown in PIG, 3, being substantially flat over a frequency range between about 150 Hz and 2000 Hz, and increasing at this point by about lOdB or the higher frequencies. Thus, the two networks exhibit transmission characteristics which resemble the equal loudness contour of the human ear at moderate loudness level and^ver the audio range of interest, except for an added sharp cutoff below approximately 150 Hz* A set of equal loudness contours are illustrated and described in an article by applicant and Emil Torick entitled "Researches in Loudness Measurement", IEEE TRANSACTIONS^ ON AUDIO AND BLECTROACOUSTICS. Vol. AU-14, No. 3, pp. 141-151» 1966* The function of networks 92 and 94 is to so shape the signals delivered to the circuitry to follow that the signal switching logic will place the respective Lg, their proper channels on the basis of their relative loudness* rather than their energy content* For example networks having weighting curves corresponding to such loudness contours would preclude a low frequency, high energy signal, such as would be -e*e?gy-r-&i5'i*a,t produced by a drum from incorrectly switching a higher frequency, lower energy, signal produced by a piccolo, for example. While networks having the characteristics illustrated in FIG. 3 are preferred, most of the advantages of the invention may be achieved without increasing the transmission of the network at the higher end of the frequency band, or, indeed, it may even be attenuated above approximately 1000 Hz as shown by the dashed-line in FIG. 3.

- After shaping by the networks 92 and 94 the signals are applied to a pair of gain control amplifiers 96 and 98, the gains of which are made to closely track each other by application of a common control signal to their . respective gain control terminals 100 and 102.

The two outputs of amplifiers 96 and 98 are separated into four separate outputs by networks 104 and 108 in the same manner as the four outputs are produced by circuits 48 and 50. These two sets of four outputs thus resemble each other in musical content, but the former set is held at a uniform level , despite variations in the dynamic -/range of the record (as modified by circuits 92 and 94) , by the action of gain control amplifiers 96 and 98. To achieve this constant output level, the four signals are rectified by rectifiers 110, 112, 114 and 116, respectively, and summed by isolating resistors 118, 120, 122 and 124 to develop a sum signal across a common resisto 126 connected between a bus 128 common with the resistors and ground. The voltage developed across resistor 126 is applied over conductor 128 to the unction 130 between amplifiers 96 and 98, respectively, which are operative in response thereto to keep the average voltage across resistor 126 substantially constant.

The gain control action is enhanced by connecting four capacitors 132, 134, 136 and 138 across the resistors 118, 120, 122 and 124, respectively, whereby the rectified voltage represents the envelope of the wave rather than its instantaneous value. The automatic gain control action maintains the sum of the voltages across the resistors 118, 120, 122, and 124 constant because the voltage across the relatively smaller resistor 126 is the sum of the four rectified voltages.

While half-wave rectifiers are illustrated, and the circuit will function satisfactorily with half-wave rectification, it should be understood that full-wave rectification is preferable and is used. in the preferred embodiment. While the j st-described gain control function is not .absolutely essential for the proper functioning of the balance of the logic, applicants have found that it provides somewhat better performance and its use is preferred.

An important aspect of the invention is that the signals delivered by amplifiers 96 and 98 and the networks 104 and 108 (whether or not subjected to automatic gain- control) are autocorrelated , as by multiplication in a particular way and subsequent integration. More particularly, the output signals from amplifiers 96 and 98 are applied via conductors 142 and C-1061- applied via conductors 148 and 146, respectively, to a second multiplier 150. The output of multiplier 144 is coupled over conductor- 152 to an R-C integrating circuit consisting, for example, of resistor 154 and capacitor 156 and the output of multiplier 150 is applied over conductor 158 to a similar integrating cirucit consisting of resistor 160. and capacitor 162. The values of resistors 154 and 160 and capacitors 156 and 162 are so related that the integrators have a turnover point at approximately the cutoff point of networks 92 and 94, namel at approximately 150 Hz, so as to provide integration of all signals at 150 Hz and above.

The function of the multipliers and associated integrators will be evident from a comparison of the outputs of the two integrators for the example previously, discussed; i.e., when the input signal consists of only a "front left" signal, L , and a "right front" signal, F R . It is seen from the waveform associated with the F output of integrator 154-156, which handles the product of the -uricorrelated voltages 1 and R^, is very nearly zero; it is not precisely zero only because the integrator is not a perfect integrator. The output of integrator 160-162, on the other hand, varies plus and minus depending upon the synchronization of the parts played by the two musicians. Over a long period of time the output of integrator 160-162 is also zero, but if the integration is carried out for about 6.7 milliseconds, corresponding to 150 Hz, the instantaneous high frequency perturbations As an alternative to the RC integrators, suitable low-pass filters with rejection above about 150 Hz. may be used. The use of such filters would eliminate the actual A.C. signals and their instantaneous products, and would transmit only the low frequency control signal corresponding to the correlation function.

The outputs of integrators 154-156 and 160-162 are respectively applied to a pair of similar wave shaping networks 164 and 166, the purpose of which is to limit the excursion of the output voltages from the integrators to a reasonable figure. The transmission characteristic of the networks may take a variety of shapes, but in general, circuits having the characteristic depicted in FIG. 4 would be used. The output voltage maxima E* and E" of the shapino network are adjusted m m . " to levels suitable for use with the electrical elements to follow. The output voltages from the shaping networks 3,64 and 166 are rectified by corresponding rectifiers 170 and 176 to charge capacitors 174 and 180, respectively. The charge on the capacitors is allowed to discharge through resistors 172 and 178, respectively. The values of the resistors and capacitors are so chosen that the discharge time constant is of the order of 100 milliseconds, it being understood, however, that the choice of time constant may vary over a very large range without significantly affecting the performance of the decoder. While the rectifiers 170 and 176 are shown as half-wave rectifiers, it will be understood that C-1061 t The outputs of rectifiers 170 and 176 are used to control the gain of the gain control amplifiers which precede the four loudspeakers 72, 74, 76 and 78. To this end, the outputs of the' rectifiers are applied to a first subtracting circuit 182, in plus-minus relationship, and to a second subtracting network 184, in a minus-plus relationship. The output of network 182 is applied in parallel to the gain control terminals 80 and 82 of gain control amplifiers 66 and 68, respectively, while the output of network 184 is applied in parallel to the gain control elements 84 and 86 of amplifiers 64 and 70, .respectively.

The performance of the decoder will now be ana-lyzed by again assuming that the input consists of the same input voltages and R^, as before. The output of , rectifier 170, depicted by the associated wave form, is near zero, whereas, the output of rectifier 176, as is apparent from the depicted wave form, initially has a positive output. Consequently, the output of summing circuit 182 is negative, and since it is applied as a control signal to gain control amplifiers 66 and 68, their gain is diminished, or, depending upon the adjustment of the gain, may be completely cut off. At the same time, the output of network 184 would be positive, thereby enhancing the gains of amplifiers 64 and 70.

Therefore, the two signals L' and R' appear only at the F F corresponding loudspeakers 72 and 78, positioned at the left front and right front corners of the listening " " - Conversely, if the incoming signals delivered by the pickup 26 were composed o-f only the "back" signals, .707LB and .7071^, then the output of recitifier 176 would be at or near zero, while a significant voltage would appear at the output of rectifier 170. Under these conditions, the output of summing device 182 would be positive, while the output of summing device 184 would be negative, thereby causing the gain of amplifiers 66 and.68 to be increased and the gain of amplifiers 64 and 70 to be diminished, or completely cut off, which, of course should be the result if the input "back" signals are to appear as signals L" and R" .

B B It will be recognized by ones skilled in the art that it is not essential that the shaping networks 164 and 166 be placed immediately following the integrators but may be placed at any appropriate location elsewhere in the circuit, for example, immediately following the summing networks 182 and 184, respectively, or may be eliminated by provision of suitable gain characteristics in the gain control amplifiers 64, 66, 68 and 70.

As a further refinement, but not essential to the operation of the switching logic of the invention, it may be desirable, as taught in applicant's co-pending Israe 37418 application Ser: No. -4-4-/-¾¾ -, to connect in circuit, m advance of one. or all of applifiers 64, 66, 68 and 70, additional networks for the purpose of placing the acoustical output phasors of loudspeakers 72, 74, 76 and 78 in an optimum phase position; or the amplifiers 64 It will be evident from the foregoing that there is provided a system for reproducing and presenting on four independent loudspeakers four channels of information carried as two composite signals on a two-track medium. An important aspect of the decoding apparatus which gives it its capability of creating a substantially perfect illusion of sound proceeding from four separate sources is the concept of sensing, using autocorrelation techniques/ which channels have the predominant signal and switching to those channels, while at the same time attenuating the* signals in the other channels, to give the illusion of four separate channels of information.

While the designations Lp, R^, L and Rg have been used in the specification and claims to identify signals corresponding to particular loudspeaker locations and a particular matrixing scheme, it is to be understood that this has been done as a convenience in describing the operation of the invention in a particular arrangement of matrix and loudspeakers , and not in a limiting sense. That is, it will be recognized by ones skilled in the art that the described principles can be utilized in reproducing apparatus employing an arrangement in which four. loudspeakers are positioned at locations which may not be accurately characterized as "left front", "right front", etc. and supplied from a decoder based on any suitable matrix scheme, and that in this case signal designations corresponding to the locations of the loudspeakers would be used.

Claims (5)

C-10C1 CLAI S :

1. In apparatus adapted to receive a first composite signal including..a dominant signal, Lp, and two subdominant signal components, LB and RB, and a second composite signal including a dominant signal, Rp, and two subdominant signals, Lfi and Rg, and including additive and subtractive networks operative to derive from said first and second composite signals a third composite signal including, a dominant signal LB and two subdominant signals Lf and Rp and a fourth composite signal including a dominant signal, g, and two subdominant signals, Lp and p, and means for applying said first, second, third, and fourth composite signals to respective first, second, third and fourth gain control amplifiers for reproduction over four corresponding loudspeaker circuits, a control circuit for. enhancing the realism of the four channel sound produced by said apparatus, said control circuit comprising: a pair of phase shift networks operative to position said first and second composite signals in phase-quadrature prior to. application thereof to said additive and subtractive networks, first and second circuit means connected to respectively receive said quadrature first and second composite signals and said third and fourth signals and operative to produce first and second alternating current signals respectively representing the autocorrelation of said first and second and of said third and fourth composite signals over a predetermined time inter- C-1061 first and second rectifying and time constant- defining-means respectively connected to receive and operative to convert said first and second alternating current signals to first and second unidirectional signals, circuit means operative to subtract said first and second unidirectional signals from each other, and * · gain control means for differentially applying said subtracted signals to said first and second gain control amplifiers and to said third and fourth gain control amplifiers, thereby to control the gain of said amplifiers in concert with the instantaneous presence of said composite signals.

2. Apparatus according to claim 1 wherein said first and second circuit means comprises first and second multiplying circuits operative continuously to multiply said first and second and said third and . fourth composite signals, respectively, and first and second integrating means respectively operative to inte- grate the outputs of said first and second multiplying circuits over said predetermined time interval, thereby ' to produce integrated products of said first and second and of said third and fourth composite signals.

3. Apparatus according to "claim1wherein said predetermined time interval is approximately 6.7 milliseconds. c-iodi

4. '4. In apparatus adapted to receive a first composite signal including a dominant signal, Lf/ and two subdominant signal components , and , and a second composite signal including a dominant signal, R^, and two subdominant signals , and R^, and including first additive and subtractive networks operative to derive from said first and second composite signals a third composite signal including a dominant signal, L^, and two subdominant signals, and R^, and a fourth composite signal including a dominant signal, R^, and two subdominant signals, Lp and , and means for applying said first, second, third and fourth composite signals to respective first, second, third and fourth gain control amplifiers for reproduction over four corresponding loudspeaker circuits , a control circuit for enhancing the realism of the four channel sound produced by said apparatus, said control circuit comprising: a pair of phase-shift networks operative to position said first and second composite signals in phase-quadrature prior to application thereof to said additive and subtractive networks, . circuit means including second additive and subtractive networks connected to receive and operative to derive 'from said quadrature first and second composite signals fifth, sixth, seventh and eighth signals substantially corresponding in information content to said first, second, third and fourth composite signals, respectively, first and second multiplying and integrating C-10G1 V and respectively operative to produce a first integrated product of said fifth and sixth signals and a second integrated product of said seventh and eighth signals, first and second rectifying and time constant-defining circuits respectively connected to receive and operative to convert said. first and said second integrated productsto first and second unidirectional signals subtracting circuit means operative to subtract said first and second unidirectional signals from each other, and gain control means for differentially applying said subtracted signals to. said first and second gain control amplifiers and to said third and fourth gain control amplifiers, thereby to control the gains of said amplifiers in concert with the instantaneous presence of said composite signals. C-1061

5. Apparatus according to claim 4 further including in circuit between said phase-shift networks and said circuit means first and second signal-shaping networks for modifying said' quadrature first and second composite signals prior to application to said second additive and subtractive networks, said signal-shaping networks having like transmission characteristics approximating the equal loudness contour of the human ear at moderate loudness level and over the audio range of interest, v For the Applicants Br. Yi zhak Hose