871,992. Stereophonic transmission systems. ELECTRIC & MUSICAL INDUSTRIES Ltd. Oct. 4, 1957 [Oct. 13, 1956; July 6, 1957], Nos. 31251/56 and 29331/57. Divided out of 21465/57. Class 40 (4). In a stereophonic sound transmission system there is derived from sounds an audio-frequency signal of relatively wide frequency range representative of the sounds and a control signal which is representative of the directions or positions of the sources of the sounds, is substantially independent of the loudness of the sounds, and is such that it may be employed in conjunction with the audio-frequency signal to energize spaced sound reproducers to give stereophonic reproduction of the sounds, the rate at which the control signal can vary to represent different directions or positions being such that although the control signal is confined to a relatively narrow range of low frequencies it extends into the audio-frequency range to such an extent that the reproduction can give the effect of a plurality of sound sources sounding virtually simultaneously. Transmitting equipment. Fig. 1.-A pair of coincident directional microphones 1, 2 provide left and right signals V L and V R , respectively, which are applied to pre-amplifiers 3, 4, tone and volume control circuits 5, 6 and adjustable gain amplifiers 7, 8. The outputs of amplifiers 7 and 8 are summed in amplifier 9 and differenced in amplifier 10 to give signals V L + V R and V L - V R , which signals are applied respectively to phase-shift networks 11 and 12 to produce a relative delay of 90 degrees between corresponding components over the full frequency range. The outputs of networks 11 and 12 are then added in amplifier 13 to produce a composite signal V A representative of the wave shape of the sounds picked up by the microphones 1 and 2 which is suitable for both stereophonic and monophonic reproduction. A composite signal may be derived by direct addition of V L and VR, but the method described above is preferred. The signal V L from the tone and volume control circuit 5 is also applied to an amplifier 14 the response of which rises approximately linearly with frequency and the output V<SP>1</SP>L is applied to a phase-splitting amplifier 15 which delivers signals V<SP>1</SP>L and - V<SP>1</SP>L to a push-pull amplifier 16 and to a rectifying arrangement 17, comprising a first fullwave rectifier 18 which peak rectifies the signals V<SP>1</SP>L and - V<SP>1</SP>L and a second full-wave rectifier 19 which peak rectifies the signal V<SP>1</SP>L after it has been phase advanced by 90 degrees and phase delayed by 90 degrees. Such polyphase rectification reduces the proportion of the original signal V<SP>1</SP>L in the output so that subsequent smoothing is facilitated. The outputs of rectifiers 18 and 19 are combined in a smoothing filter 20 the output of which, V<SP>11</SP>L, is representative of-theenvelope shape of V<SP>1</SP>L. The filter 20, which may begin to cut off at about 100 c.p.s., delivers the output V<SP>11</SP>L to a cathode follower 21 and thence, if desired, to a further filter 22 and cathode follower 23. The filter 22, if employed, weights the signal V<SP>11</SP>L in favour of transients. The output of 21, or 23 as the case may be, is applied to the junction of resistors 24, 25. A signal V<SP>11</SP>R is similarly derived from VR and is applied to the lower end of resistor 30 which is equal in value to resistor 25. The circuits 26 and 27, termed logger circuits, derive the logarithm of applied signals. To the first is applied V<SP>11</SP>L and to the second (V<SP>11</SP>R + V<SP>11</SP>L)/2, the output of the second being phase inverted by circuit 31 the output of which represents log( 1/V<SP>11</SP> R + B<SP>11</SP> L ). This output is added to that of circuit 26 (viz. log V<SP>11</SP>L) by cathode followers 32, 33 and resistors 34, 35 to produce a signal representing log( V<SP>11</SP> L /V<SP>11</SP> R + V<SP>11</SP> L ) which passes via a D.C. coupled amplifier 36 to an amplifier 37 having an exponential response characteristic which produces a signal repre- senting (V<SP>11</SP> L /V<SP>11</SP> R + V<SP>11</SP> L ). This signal, which is substantially independent of the loudness of the sound, passes via a cathode follower 38 to a ring modulator 39 where it modulates a 15 kc/sec. carrier derived from oscillator 40. The modulated carrier, after amplification in 40a, is added to the signal VA in the adding circuit 41 and is then used to control the cutting stylus 42 of disc recording apparatus. The signal (V<SP>11</SP> L /V<SP>11</SP> R + V<SP>11</SP> L ) is modulated on the 15 kc/sec. carrier to prevent its reproduction as an audio-frequency signal. The rectifiers 18 and 19 may be interconnected so as to exaggerate the differential significance of the control signal. Reproducing apparatus, Fig. 2.-Signals derived from a gramophone disc 50, which is a replica of that cut by the stylus 42 (Fig. 1), by a pick-up head 51 are applied in parallel to two amplifiers 52 and 53 leading to the audio frequency section and to the control signal section, respectively, of the reproducing equipment. The output of amplifier 52 passes via a cathode follower 53a to a low-pass filter 54 which suppresses the carrier modulated with the control signal and delivers the signal VA to a tone control circuit 55 and thence to an amplifier 56. The output signal VA then modulates a carrier wave of, say, 850 kc/sec. derived from an oscillator 60, in a balanced modulator 59, the sideband components from which are applied in parallel to two radio-frequency amplifiers 61 and 62 the gains of which are controlled by signals derived from the control signal. The output of the amplifier 53 which has a band-pass response centred on 15 kc/sec. passes via the band-pass filter 63, which removes any components of VA to a phase-splitting amplifier 64 the outputs of opposite polarity from which are demodulated in 65 and 66 to derive outputs representing (V<SP>11</SP> L / V<SP>11</SP> R + V<SP>11</SP> L ) and (1- V<SP>11</SP> L / V<SP>11</SP> R + V<SP>11</SP> L ), respectively, the signal representing unity in the output of 66 being set up by applying a suitable bias to that demodulator. These outputs are applied to logger circuits 67 and 68 which produce two gain control signals of the form log(V<SP>11</SP> L / V<SP>11</SP> R + V<SP>11</SP> L ) and log(V<SP>11</SP> R / V<SP>11</SP> R + V<SP>11</SP> L ) which control the gain of amplifiers 61 and 62. These two amplifiers have exponential response characteristics and give outputs of the form V A (V<SP>11</SP> L / V<SP>11</SP> R + V<SP>11</SP> L ) and V A (V<SP>11</SP> R / V<SP>11</SP> R + V<SP>11</SP> L ), V A being in the form of sideband components of the 850 kc/sec. carrier. These outputs are combined in amplifiers 70 and 71 with carrier oscillations from oscillator 60 via the amplifiers 72 and 73 which impart phase delay such as to ensure that at the amplifiers 70 and 71 the carrier oscillations are in correct phase for combination with the outputs of 61 and 62. The signals in the two channels are then demodulated by 74 and 75 and fed as audio-frequency signals via amplifiers 76 and 77 to the spaced loud-speaker groups 78 and 79. It is stated that the use of a control signal which extends into the audio-frequency range to a significant extent allows a realistic impression to be obtained of a plurality of spaced sound sources sounding vertically simultaneously, as constituted for example by an orchestra. The control signal which is recorded as the ratio (V<SP>11</SP> L / V<SP>11</SP> R + V<SP>11</SP> L ) may be employed in this form in reproducing apparatus with any form of gain control characteristic and can give satisfactory results in commercial sound reproducers. The use of the logarithm of the ratio in conjunction with an exponential gain control characteristic is, however, preferred because it gives highly linear gain control. Modified form of reproducing equipment, Fig. 3.-In order to avoid the difficulty of matching the gain control characteristics of two channels, only one of the channels leading to the loudspeakers 78, 79 is controlled in gain by the control signal. From Fig. 2 it will be seen that the output envelope of amplifier 61 is repre- sented by V A (V<SP>11</SP> L / V<SP>11</SP> R + V<SP>11</SP> L ) and that of 62 by V A (1 - V<SP>11</SP> L / V<SP>11</SP> R + V<SP>11</SP> L ), and hence that the signal to be fed to 79 may be obtained by subtracting the output of 61 from its input. Accordingly, the output of the band-pass filter 63 is passed by an amplifier 100 (Fig. 3) to the single demodulator 65 which reproduces the control signal at low frequency, from which the logger circuit 67 derives the signal log(V<SP>11</SP> L / V<SP>11</SP> R + V<SP>11</SP> L ) for gain control of amplifier 61. The audiofrequency signal from the demodulator 74 is applied via the cathode follower 103 to the audio amplifier 76 (as in Fig. 2), and also to a differencing circuit 107. The amplifier 62 of Fig. 2 is replaced by a delay device 104 consisting of an amplifier identical to 62 but having no automatic gain control. Manual gain control may be produced by the bias control 105 to establish initial balance between the channels. The carrier frequency is added to the output of 104 in the amplifier 71 whereafter the demodulator 75 produces VA suitably delayed to be passed by the cathode follower 106 to the differencing circuit 107 which produces the desired output signal for the second channel. This modified form of the reproducing equipment is claimed in Specification 871,994. Detailed circuits of the rectifying arrangement 17 of Fig. 1 and of the balanced modulator 59 of Fig. 2 are described in the Specification.