807,024. Automatic volume and tone control. KOKUSAI DENSHIN DENWA CO. Ltd. Sept. 16, 1955 [Sept. 18, 1954 (2) ; Nov. 6, 1954; April 25, 1955], No. 26637/55. Class 40 (4) In an automatic compensation system for the transmission characteristics of a signal channel, a compensating network operated in accordance with measurements made upon a transmitted reference signal comprises a combination of controllable amplifiers and delay lines. The invention is especially applicable to systems liable to " fading." First embodiment. As shown in Fig. 4 incoming speech signals together with the transmitted pulses modified according to the channel transmission characteristic are fed to terminal III. The circuit IV accepts the pulses and constructs signals for feeding to circuit V which is controlled in an inverse way to the variations in the transmission characteristic. The signal from III is thus corrected by circuit V and the output fed to a blanking circuit VI controlled by a pulse generator VII synchronized by pulses received from circuit IV to remove the pulses from the speech output which is fed to terminal VIII. Circuits IV and V shown together in more detail in Fig. 5 operate as follows. The incoming pulses synchronize a pulse generator Op, Fig. 2, whose output is fed to a series of delay lines D arranged to produce gating pulses i for gates Go ... G L at delays - (i = o ... L), 2W where W is the bandwidth of the channel. The impulse response of the channel is thus obtained at successive instants by gates Go ... G L and fed to low-pass filters LFo ... LF L to produce D.C. outputs at IIa o ... IIa L . This network is shown as FD in Fig. 5. These outputs are used to control amplifiers A = ai(t) (i = o . . . L) serially connected to delay lines D = i/2W, as shown in Fig. 1 to form a network simulating the transmission characteristic of the channel K(y), where This network is shown in Fig. 5 as K o (y). Its output is fed to a circuit C1 similar to FD and fed with pulses from a generator 0 synchronized to generator Op (Fig. 2). However, alternate outputs are inverted so that when these are fed to a circuit similar to that of Fig. 1 odd powers of y have their signs reversed so that a function K o (- y) is produced. The series combination of K o (y) and K o (- y) (lower circuit) produces an output K 1 (-y<SP>2</SP>). Likewise the channel characteristic in series with K o (- y) (upper circuit) produces an output K 1 (- y<SP>2</SP>). The process is repeated using a circuit C2 by means of which the odd amplifiers are omitted and the even amplifiers have alternate positive and negative amplification factors and so on. It is demonstrated mathematically that if a sufficient number of such stages is used then the output of the last stage of the upper chain is substantially proportional to the original signal prior to transmission over the channel, the factor of proportionality being determined by the lower chain. The output of the lower chain is therefore used to control the gain of an amplifier G of the upper chain in such a way as to compensate for changes in the factor of proportionality. In a second embodiment the impulse response is used to determine the frequency characteristic of the compensating system, Fig. 7a. The impulse responses from Fig. 2 are used to control amplifiers A = a i fed by signal f = if o . The combined amplifier outputs K o (y) are fed to a sign converting circuit C1 consisting of band-pass filters and inverters which invert alternate frequencies to produce an output K o ( Y ). The signals K o (y), K o (- y) are fed to multipliers X 1 X_ 1 to produce products K 1 (y<SP>2</SP>) and K 1 (- y<SP>2</SP>). The operation is repeated using further sign converters C2, C3 ... to form functions Kr(y<SP>2r</SP>) and Kr(- y<SP>2r</SP>). The product K 1 (- y<SP>2</SP>)K 2 (- y22) ... . K N - 1 (- y2<SP>N-1</SP>) is then formed by multipliers X - 1 , X _ ... X- (N - 1) and the output subjected to automatic gain control at G. The resulting amplitude outputs for the various frequency inputs if o constitutes the impulse response of the compensating network and by means of this response a network of the type described in Specification 782,004 can be controlled to provide the necessary correction. In a third embodiment it is demonstrated that the frequency characteristic of the inverse network may be expressed as The circuit of Fig. 8b may be used to form x and may be formed by the circuit of Fig. Sa in which the amplifiers are connected to the delay lines in reverse order as compared with Fig. 1. The circuits for x are then introduced at XA, Fig. 8c, to produce a compensating circuit having the required frequency characteristic. In a fourth embodiment, Fig. 9a, the impulse response of the compensating network is constructed. If the impulse response of the channel is given by then the impulse response of the compensating network is given by where R(o) and x<SP>1</SP> are the D.C. and A.C. components of |K(t) |<SP>2</SP>. [ K(t) |<SP>2</SP> is constructed by means of generators f = ifo, amplifiers A = a i and A<SP>1</SP> = a L - i together with a suitable multiplier Y<SP>1</SP>. The arrangement of squaring circuits X I , X n ... X M , frequency doublers H 1 , H 2 ... H M and adders AD 1 ... AD M + 1 in Fig. 9a may thus be seen to construct the impulse response function indicated above. This is used as described in Specification 782,004 to form the required inverse network. In a further embodiment a measuring signal comprising a plurality of equally spaced test frequencies all of equal amplitude is transmitted. Thus in a system having a carrier frequency fo a high signal comprising frequencies fo + mfp (m = O ... M) all of equal amplitude is transmitted. At the receiver RE, Fig. 11, a harmonic generator HG produces frequencies fo + mf p to feed respective demodulators Do ... D M connected to filters LF. The D.C. outputs of the latter inversely control the amplification of amplifiers AM o ... AM M which function to amplify the outputs of filters Fo ... F M which function to select suitable speech bands around the frequencies fo + mfp. These filters are constituted by suitable combinations of amplifiers or attenuators and delay lines. The filters are fed with the incoming signal so that the combined output t of the amplifiers consists of a signal with its distortion compensated. This arrangement is suitable where the duration of the impulse response T<SP>1</SP> of the compensating network is smaller than that, T, of the channel. For the case where T<SP>1</SP> is greater than T the modification of Fig. 12 may be used, in which a device D comprising demodulators and filters LF as in Fig. 11 produces outputs A o ... A M corresponding to the various frequencies and these are scanned at S1 ; the output thereof is fed to a low-pass filter LPF whose continuous output over the scanning period represents the frequency response of the channel. This output is distributed over leads A o 1 ... A L <SP>1</SP> by a distributer S2 operating at a faster speed so as to give a larger number of interpolated values of the frequency response. These are fed to amplifiers AM o ... AM L fed by filters in the manner described for Fig. 11.