GB1591059A

GB1591059A - Digital signal processing method and apparatus

Info

Publication number: GB1591059A
Application number: GB5380777A
Authority: GB
Original assignee: Sony Corp
Current assignee: Sony Corp
Priority date: 1976-12-27
Filing date: 1977-12-23
Publication date: 1981-06-10
Also published as: DE2758276C2; JPS6136311B2; DE2758276A1; AU521114B2; CA1097819A; JPS5381222A; IT1193270B; AU3188377A; ATA933377A; FR2375782A1; AT380608B; NL7714443A; FR2375782B1

Description

(54) DIGITAL SIGNAL PROCESSING METHOD AND APPARATUS (71) We, SONY CORPORATION, a Japanese Body Corporate of 7-35 Kitashinagawa 6-chome, Shinagawa-ku, Tokyo 141, Japan do hereby declare the invention for which we pray that a patent may be granted to us, and the method by which it is to be performed, to be particularly described in and by the following statement: This invention relates to the processing of digital signals in such a way that there is a high probability of detecting and correcting errors in certain stages of the processing. In particular, the invention relates to deriving primary pulse code modulation (PCM)multi-bit' digital signals from an analog signal, deriving secondary PCM signals from the same values of the analog signal as were included into the primary PCM signals, delaying either the primary or secondary signals, and combining groups of the delayed signals in interleaved' sequence with groups of the relatively undelayed signals.

It has been proposed in our co-pending patent applications, Nos. 4946/77 and 7410/77 (Serial Nos. 1556092 and 1524811), to process analog audio signals on a video tape recorder (VTR). These applications, along with a third co-pending application No. 51864/77 (Serial No. 1579199) describe techniques for encoding the audio signals by means of PCM techniques and arranging the resultant pulse signals to correspond to a video format so as to be handled easily by a VTR.

The analog signals are sampled at frequency at least substantially twice as high as the highest information frequency embodied in such signals. A convenient sampling rate is three times the horizontal line frequency rate of the video signal, because this makes the resultant sampled signal commensurate with video horizontal synchronizing signals incorporated into the pseudo-video format.

VTRs are capable of handling audio signals with excellent fidelity, but there is still an occasional problem of loss of signal due to a drop out or to a noise burst. Since the quality'of reproduction possible with the equipment of the type described hereinafter is of such high' quality as to be consistent with commercial standards, it is important to avoid even occasional errors in the processed signal.

According to the present invention there is provided a method of'processing first multi-bit digital signals grouped into digital words, comprising the steps of selecting, from each digital word of the first multi-bit signals, a secondary group of bits consisting of only predetermined ones of the bits of the respective said digital word, each digital word and the secondary group of bits thereof comprising related signal groups; delaying one of said signal groups relative to the other of said groups and combining the delayed signal groups in interleaved sequence with subsequent, relatively undelayed, signal groups to form composite, sequential signals.

A second aspect of the invention provides aparatus for processing a primary multi-bit digital signal, said apparatus comprising an input to receive said primary signal into said apparatus; a first gate to generate a secondary multi-bit digital signal containing only part of said primary signal; a delay connected to one of said input and said first gate to delay the signal from a selected one thereof relative to the signal from the other thereof by a predetermined interval of time; a second gate connected to said delay and to said other of said input and said first gate to conduct, alternately, a delayed signal and an undelayed signal; a first time-compressor connected to said input to receive and store said primary signal at one rate and to extract said primary signal from storage at a higher rate in first groups spaced apart in time; a second time-compressor connected to said first gate to receive and store said secondary signal at said one rate and to extract said secondary signal from storage at said higher rate in.second groups spaced apart in time; and third and fourth gates connected to said first and second time compressors respectively to extract said groups of said primary signal in alternation with said groups of said secondary signal.

Thus a multi-bit digital signal, referred to as a primary signal, and a secondary multi-bit digital signal on which the same information is encoded in the same PCM arrangement, except that the lowest bit orders of the primary signal may be omitted from the secondary signal, are arranged in corresponding, or related, groups, each of which may comprise a digital word signal. Either the primary signal is delayed relative to the secondary signal or the secondary signal is delayed relative to the primary signal, and groups are then interleaved to form a sequential multi-bit signal. After this sequential signal has been processed, related groups may be brought back into coincidence, and one of the signals selected for further processing. This selection is made on the basis of which signal is most error free at that point. In the embodiment described below, if both of the signals are error free, a predetermined one of them is selected for further processing while if neither signal is error free, a preceding, temporarily stored signal may be processed a second time instead of processing either of the erroneous signals.

In order to determine which signal is error free, cyclic redundancy check code signals are obtained for both the primary and secondary PCM signals prior to processing, and these cyclic redundancy check code signal (CRC) are interleaved with the primary and secondary signals to make up a complete sequential signal. In determining whether there is an error in either of the processed signals, the primary and secondary signals may be measured for coincidence, and if they do coincide, the present invention provides that one of them, usually the primary signal, be chosen for further processing. Simultaneously, the CRC signals for the respective primary and secondary signals may be used to determine whether either the primary or secondary signal is error free. If only one of these signals is error free, which means that the two signals would not be found to be in coincidence, the error free signal is automatically selected for further processing. Each signal selected for further processing is retained in what may amount to a temporary memory circuit to be used yet again if neither of the next succeeding pair of primary and secondary signals is error free.

Continued utilization of the same signal a second time is consistent with the fact that it is likely to be less distracting if the output signal does not vary at all for an additional increment of time than if it varies sharply.

In order to provide time for inserting the secondary signal groups and the CRC signals in sequence with the primary signal groups, both the primary and secondary signals are subjected to time compression. This may be effected by applying these signals to memory circuits at one clocking rate and reading them out of the memory circuits at a different clocking rate. For ease of handling the signals in subsequent circuits, such as the coincidence circuit, it is preferably that the time compression of the primary signals be exactly equal to the time compression of the secondary signals. Since both of these signals contain substantially the same information, except, possibly, for information encoded in the least significant bits of the primary signal, the time compression should be approximately 2:1. In fact, by omitting enough of the least significant bits of the primary signal to correspond to the bits that are required to transmit the CRC signals, the sum of the truncated secondary signal groups plus the two sets of CRC signals may be made exactly equal to the number of bits of the primary signal groups so that time compression of exactly 2:1 will result in a uniformly dense sequential pulse signal.

A further time compression is required to allow space between certain of the pulses to insert horizontal and vertical synchronizing and equalizing signals. In addition, since it is desireable to provide an integral number of complete, or composite, groups in each horizontal video interval, it is also desirable to include word synchronizing pulses between each pair of composite groups. The term "composite" is used to indicate that such a group includes a primary signal digital word, the CRC signal for that word, a secondary digital word spaced in time from the primary digital word, and the CRC signal for the secondary signal.

The invention will be futher described with reference to the accompanying drawings, in which: Figure I is a block diagram of a signal processing system according to the present invention.

Figure 2 is a block diagram of an encoder section of the circuit in Figure 1.

Figures 3A-3F are symbolic representations of digital signals illustrating the operation of the circuit in Figure 1 in accordance with the present invention.

Figure 4 is a block diagram of a time compressor for use in the circuit in Figure 2.

Figure 5 is a waveform diagram of signals obtained in the operation of the circuit in Figure 4.

Figure 6 is a schematic diagram in block form of a CRC encoder and gate suitable for use in the circuit in Figure 2.

Figure 7A-7D are waveforms illustrating the operation of two embodiments of the circuit in Figure 6.

Figure 8 is a block diagram of a decoder suitable for use in the circuit in Figure 1.

Figure 9 is a schematic and symbolic circuit representation of CRC decoder circuits as used in the circuit in Figure 8.

Figures lOA-lOC are waveforms obtained in the operation of the circuit in Figures 9.

Figure 11 is a schematic and symbolic diagram of a logic circuit in Figure 8.

Figure 12 is a symbolic representation of digital signals as applied to the coincidence detector in Figure 8.

Figure 13 is a truth table representing logic conditions in the operation of the circuit in Figure 8.

Figure 14 is a schematic diagram of a gate and memory circuit for use in the circuit in Figure 8.

Figure 15 is a symbolic representatuion of signals as processed according to this invention to illustrate error reduction.

One of the coding concepts to be used in the following disclosure is known as the cyclic redundancy check code (CRC). The mathematical aspects of the CRC will be described first in terms applicable to the embodiment that follows.

Cyclic Redundancy Check Code The CRC code is generally expressed by a polynomial F(x) with indeterminant x and coefficients from an n bit code (any, a2, --- a1, aO) as follows, F(x) = an 1xn~l + a,.2x"2 + ---- + aO For example, if the 5 bit code (10011) is expressed by the polynomial F(x), then F(x) = X4 + X + l This polynomial is called the polynomial over Galois field of 2.

The encoding and decoding of the CRC code is essentially characterized by a division algorithm such that the code polynomial F(x) is divided by the generator polynomial G(x).

Now assuming that the code polynomial of degree (k-l) for a k bit code is expressed as M(x) and the generator polynomial of degree (n-k) as G(x), the division algorithm is as follows, M(x)xn-k = G(x) Q(x) + R(x) in which Q(x) is the quotient polynomial and R(x) is the remainder polynomial having a greatest degree of (n-k-l). It should be noted that the encoded code polynomial V(x) comprises the code polynomial M(x) xnk and the remainder polynomial R(x) added to the former polynomial. Therefore, the encoded polynomial V(x) has degree (n-k) and is given as follows: V(x) = M(x)x + R(x) = G(x)Q(x) This means that the encoded polynomial V(x) is divisible by the generator G(x).

Next, if a noise signal, which is expressed by a polynomial E(x), is introduced into the code polynomial V(x) during transfer, the code polynomial V'(x) at the decoding side is expressed as follows; V'(x) = V(x) + E(x) If no error is introduced therein, E(x) = 0. Then, V'(x)=V(x) and hence the polynomial V'(x) is divisible by the polynomial G(x).

However, if the polynomial V'(x) is not divisible by the generator polynomial G(x) in the decoder, causing a remainder polynomial R'(x) to be generated, the polynomial V'(x) is regarded as having an error bit. Then, the polynomial V'(x) is given as follows; V'(x) = G(x) Q'(x) + R'(x) The polynomial V(x) should be divisible by the generator polynomial G(x), so that the remainder polynomial R'(x) must be the remainder in the dividing algorithm of dividing the polynomial E(x) by the generator G(x). Accordingly, it is apparent that the remainder polynomial R'(x) is a factor showing whether or not the code polynomial V'(x) contains the error bits. Such a remainder R'(x) is called a syndrome.

One example will be shown in the condition of n=7, k=4 and the generator polynomial G(x) = x + x + 1, (1) M(x) = x3 + 1=(1001) Mx)3x3=x + x3 M(x)x = G(x)Q(x) + R(x) Rx) x + x (2) V x) = M(x)x3 + R(x) = x6 + x3 + x2 + x=(1001110) (3) E(x) = x5 = (0100000) (4) V'(x) = V(x) + E(x) = X6 + X5 + x + x + x=(1101110) V'(x) = G(x)Q'(x) + R'(x) (5) = R'(x) = x + x + 1 =(111) The basic circuit of the CRC code encoder and decoder comprises a dividing circuit with the divisor G(x) which generates the the remainder, not the quotient. The dividing circuit is essentially formed by a shift register, each stage of which is preceded by a modulo 2 adder that adds, on a modulo 2 basis (which means counting to the base 2 without carry), the output of the preceding stage and output of the shift register according to whether the appropriate element of the polynomial is gj=1 or gi=0 in the divisor G(x) = gn-1xn-1 +gn-2xn-2 +...... + g2x + g1x + g0.

Now, the generator polynomial G(x) in the above example is given as follows: G(x) = x3 + x + 1 Accordingly, the dividing circuit of the polynomial G(x) includes a three-stage shift register with feedback loops from the output to modulo 2 adders at the input and between the first and second stages. The clocking conditions in each shift register stage and the calculation example are shown; (i) E(x) = 0 V(x) = X6 + X3 + X2 + X

TABLE I

Conditions in shift registers clock input Do D1 D2 (initial condition) 0 0 0 tl 1 1 0 0 t2 0 0 1 0 t3 0 0 0 1 t4 1 0 1 0 t5 1 1 0 1 t6 1 0 0 0 t7 0 0 0 0 The remainder (ii) E(x) = x5 V' (x) = X6 + x5 + X 3 + X2 + X X3 + X2 + x + 1 X + x + 1 + x6 + x5 + X3 + X2 + X x6 +X4 + X3 X5 +X4 + x2 + X x5 +X3 + X2 X4 + X3 + X X4 + X2 + X X3 + X2 X3 + x + 1 the remainder x + x + 1 TABLE 2

Conditions in shift registers Clock input Do D1 D2 (initial condition) 0 0 0 tl 1 1 0 0 t2 1 1 0 0 t3 0 0 1 1 t4 1 0 1 1 t5 1 0 1 1 t6 1 0 1 1 t7 0 1 1 t7 0 1 1 The remainder Accordingly, the contents of the shift registers show whether or not the transferred code contains error bits.

The circuit in Figure 1 includes a VTR 1 which may be, for example, of the type referred to in the companion applications, supra. The VTR has an input terminal 1i and an output terminal 1o.

The system is arranged for use with sterophonic audio signals, although it can be used with other types of signals. As set up for sterophonic audio signals, it includes two input terminals 2L and 2R, to which the left and right audio signal channels may be applied, respectively. The input terminal 2L is connected to a low pass filter 3L which is connected, in turn, to a sample-and-hold circuit 4L. The output signals of the sample-and-hold circuit 4L. The output signals of the sample-and-hold circuit is connected to an analog-to-digital (A-D) circuit 5L, the output of which is connected to a parallel-to-serial converter 6.

The input terminal 2R is connected to the parallel-to-serial converter 6 through an identical circuit including a low pass filter 3R, a sample-and-hold circuit 4R and a A-D circuit SR.

The output of the converter 6 is connected to an encoder 7 which will be described in greater detail hereinafter in conjunction with Figures 2-7. Following the encoder 7 is a time compressor 8 that provides additional time compression of the output signal of the encoder to allow for insertion of synchronizing signals in a synchronizing signal adder circuit 9 that follows the time compressor 8. The output of the synchronizing signal adder circuit 9 is connected to the input terminal li of the VTR 1.

The circuit to this point includes the elements used in recording a signal in the VTR 1.

For playing back signals previously recorded, the output terminal 1o of the VTR 1 is connected to a synchronizing signal eliminator circuit 10 that extracts the synchronizing signals and utilizes them in the control of the VTR in the normal manner. The output of the synchronizing signal eliminator circuit 10 is connected to a time expansion circuit 11 that returns the spacing between successive pulse signals to an even amount and closes the gaps that were provided for insertion of the synchronizing signals. The output of the time expansion circuit 11 is connected to a decoder 12 that performs the converse of the function of the encoder 7 and will be described hereinafter in greater detail, particularly in connection with the circuits in Figures 8-14.

The output of the decoder 12 is connected to a serial-to-parallel circuit 13, which has two output terminals. One of the output terminals is connected to a digital-to- analog (D-A) circuit 14L and the other to a D-A circuit 14R for the left and right audio channels respectively. The output of the D-A circuit 14L is connected through a low pass filter 15L to an output terminal 16L, and the output of the D-A converter 14R is similarly connected through a low pass filter 15R to an output terminal 16R.

The conventional reference oscillator circuits and clocking, synchronizing, and gating signal circuits utilized in conjunction with the sample-and-hold circuits, the A-D and D-A converters, the parallel-to- serial and serial-to-parallel converters, the encoder and decoder, the time compressor and expander circuits, the synchronizing signal adder and eliminator circuits, and the VTR are all standard and need not be described in detail.

In the operation of the circuit in Figure 1, audio signals to be processed are sampled at a suitably high rate in the sample-and-hold circuits 4L and 4R. It is convenient and satisfactory for that rate to be three times the repetition rate of a horizontal video synchronizing signal, or about 47.25KHz. At each sampling, the respective A-D circuits 5L and SR provide sixteen-bit PCM signals to the converter 6. The latter may be a 32 position shift register clocked at a sufficiently high speed to read out all 32 bits of information applied to the converter from the two A-D converters 5L and SR. The resulting multi-bit digital signal is represented in Figure 3A as including a left channel signal with sixteen bits ranging from a most significant bit M to a least significant bit L and a right channel signal that also includes sixteen bits and ranges from a most significant bit M to a least significant bit L. The interval of time required to extract the digital signal represented in Figure 3A from the converter 6 is equal to the sampling time and therefore, in this embodiment, is H/3, where H is the horizontal line interval of a video signal. The entire multi-bit digital signal represented in Figure 3A may be considered to be a digital word signal, or digital word group.

Digital word signals like that in Figure 3A are generated at a constant rate so that three such digital words substantially completely fill up a horizintal line interval. In order to allow time for a comparison signal, the constant flow of signals similar to that in Figure 3A from the converter 6 to the encoder 7 must be subjected to time compression. This is illustrated in the circuit in Figure 2.

Figure 2 includes an input terminal 21 connected to a first time compression circuit 22.

The output of the time compression circuit 22 is connected to a CRC encoder 23 and to a gate circuit 24 to which the output of the CRC encoder 23 is also connected.

The input terminal 21 is likewise connected to a gate circuit 25, the output of which is connected to a second time compression circuit 26. The output of the latter is connected to a second CRC encoder 27 and to another gate circuit 28. The output of the encoder 27 is also connected to the latter gate circuit. The output of the gate circuit 28 is connected through a delay means 29 to a gate circuit 30 to which the output of the gate circuit 24 is also connected. The gate circuit 30 has an output terminal 31.

The time compression circuit, or time compressor, 22 reduces the length of time required to transmit the digital word signal shown in Figure 3A. A circuit to accomplish this is shown in Figure 4 as comprising a double pole input switch 32 connected to inputs of two 32 bit shift registers 33 and 34. The outputs of each of the shift registers 33 and 34 are connected to the terminals of another switch 36. The shift register 33 has a write clock input terminal 33W and and read clock input 33R, and the shift register 34 has similar write and read terminals 34W and 34R, respectively.

The bits of the signal shown in Figure 3A are applied continuously to the switch 32 and are connected, one word at a time, to the shift registers 33 and 34 alternately. To accomplish this, the arm of the switch 32 is switched from one of its output terminals to the other at the end of each 32 incoming bits. The information is clocked in to the shift register to which the arm of the switch 32 happens to be connected to any instant by the write clock signal at a frequency equal to the frequency at which the bits are generated in the converter 6 in Figure 1, but they are read out by the read clock signal at a rate twice the rate at which they are written in. The arm of the switch 36 is connected to whichever one of the two shift registers 33 and 34 is being operated in its "read" mode at any given time. Figure 5 shows that the 32 bit PCM signal applied to the arm of the switch 32 is, by virtue of being read out at twice the speed that it is written in, compressed by a factor of 2:1. Thus instead of having the 32 bits occupy the full interval of time from one sample to the next, the bits are bunched together as shown in Figure 5 leaving unused intervals of time each of which is one half the total interval occupied by the signal shown in Figure 3A. The second, or comparison signal, and two sets of CRC signals can be inserted into the resulting unused intervals.

Figure 6 shows the CRC encoder which may be used either as the encoder 23 or as the encoder 27. The circuit for each is the same and is connected in accordance with the CRC equation G(x) = X4 + X + 1. The encoder has a signal input terminal 37 connected to one input terminal of an exclusive-OR gate 38. The output of the gate 38 is connected to the D input terminal of a D-type flip-flop 39. The output of this flip-flop is connected to one input terminal of another exclusive-OR gate 41, the output terminal of which is connected to the D input terminal of a D-type flip-flop circuit 42. The latter is connected in sequence to two other D-type flip-flop circuits 43 and 44. A clock input terminal 46 is connected to the clock terminals of all four of the flip-flops 39, 42, 43, and 44. The output terminal of the flip-flop 44 is connected through an AND gate 47 back to second input terminals of each of the exclusive-OR gates 38 and 41. The output terminal of the flip-flop 44 is also connected through another AND gate 48 to an output terminal 49. Two input terminals.51 and 52 are gating signal input terminals to control the AND gates 47 and 48, which make up the gate circuit 24 connected to the encoder 23 or the gate circuit 28 connected to the en-coder 27.

The operation of the circuit in Figure 6 will be discussed in connection with the gating signals shown in Figures 7A-7D. The signals in Figures 7A and 7B are those that are applied to the gate circuit 24 while the signals in Figures 7C and 7D are those that are applied to the gate circuit 28.

The clusters of 32 bit PCM signals are applied to the input terminal 37 of the encoder 23.

During the time this 32 bit signal is being applied, the signal in Figure 7A is applied to the input terminal 51 to enable the AND circuit 47 to act as a direct connection from the output of the flip-flop 44 back to the second input terminal of each of the exclusive-OR gates 38 and 41. This enables the encoder 23 to operate as a shift register with feedback in accordance with the equation G(x) = X4 + X + 1, as described previously in connection with the cyclic redundancy check code. After the 32 bit interval, which may be considered a digital word, the AND gate 47 is disabled to prevent any further signal feedback from the output of the flip-flop 44 to the exclusive-OR gates 38 and 41. At the same time no further input signals are applied to the terminal 37. During the next four bit intervals, the gating signal in Figure 7B is applied to enable the AND gate 48, and during this interval, the flip-flops 39 and 42-44 unload through the output terminal 49. As shown the input terminal 37 is directly connected to the output terminal 49 so that, during the 32 bit intervals while the AND gate 48 is disabled, the PCM input signal consisting of 32 bits is transmitted through the short circuit to the output terminal 49. The CRC pulses are added sequentially in the immediately ensuing four bit interval. These bits are clocked at the same rate as the input signal applied to the terminal 37. This is also the same clocking rate as the read clock applied to the terminals 33R and 34R in the time compressor in Figure 4. Thus, all of the pulses at the output terminal 49 of the circuit in Figure 6 are at the same repetition rate, and they occur over a 36 bit interval, which is the sum of the two intervals shown in Figures 7A and 7B. The complete signal is represented in Figure 3B as comprising a time-compressed 32 bit portion that corresponds to the signal in Figure 3A, and a four bit portion that includes the CRC signal. Since the signal in Figure 3A has been referred to as a digital word, the signal in Figure 3B may be referred to as an expanded digital word.

The total time difference between the signal in Figure 3A and that in Figure 3B is the time required for 28 bits. As suggested previously, this interval can be uniformly filled by a 28 bit signal comprising 24 bits of information and another four bit CRC signal. The 24 bit signal corresponds to the most significant bits of the original information signal applied to the input terminal 21 in Figure 2, and this signal is illustrated symbolically in Figure 3C as comprising 12 bits for the left channel and 12 bits for the right channel. This is a truncation of the original signal shown in Figure 3A, but all that it represents in physical terms is a relatively small reduction in the dynamic range of operation of the system. Since the range of operation represented by even a 12 bit signal is still quite good, the loss of additional range is almost unnoticeable.

The derivation of the truncated signal from the original input signal at the terminal 21 is made by the gate circuit 25. This truncated signal is then compressed by a time compressor 26 identical with the compressor 22, and it is actually the compressed signal that is represented in Figure 3C. The CRC signal for the compressed truncated signal is generated in the encdder 27 and the gate circuit 28 by applying the gating signals shown in Figures 7C and 7D to the circuit in Figure 6.

While the signal with a signal into coincidence with the Bj l8 signal. The A1 - 18 signal from the delay means 53 is connected to the input of a CRC decoder 54, and the Bj 18 signal from the gate 52 is applied to asimilar CRC decoder 55. The output signal of the delay means 53 is also applied to the gate 56, and the output of the latter circuit is connected to a time expansion circuit 57. In a similar manner, the By 18 signal from the gate 52 is applied to a gate 58, and the output of the latter gate is connected to time expansion circuit 59. The output of the time expansion circuit 57 is connected through a gate 60 that truncates the signal to the same degree that the B signal was previously truncated. In addition the output of the time expansion circuit 57 is connected to an output gate 61. In a similar manner the output of the time expansion circuit 59 is connected to another input terminal of the gate 61 and to a coincidence detector 62 that also receives the truncated output signal of the gate 60. Output signals from the CRC decoders 54 and 55 and from the coincidence detector 62 are applied to a logic circuit 63 to produce an output gate control signal connected to the gate 61 and to a "hold" signal.

The CRC decoders 54 and 55 in the circuit in Figure 8 are shown in greater detail in Figure 9. Each of these circuits is quite similar to the CRC encoder circuit 23 shown in Figure 6. The decoder 54 includes an input terminal 66 to which the Aj signals are applied.

This terminal is connected to one of the input terminals of an exclusive-OR gate 67 the output of which is connected to the D input terminal of a D-type flip-flop 68. The output of the D flip-flop is connected to one of the input terminals of another exclusive-OR gate 69, and the output of the latter is connected to the D input terminal of the first of three successive d-type flip-flops 71-73. The output of the last flip-flop stage 73 is connected directly back to a second input terminal of the exclusive-OR gate 67 and to the second input terminal of the exclusive-OR gate 69. The decoder 54 also has a gate signal input terminal 74 connected to an AND circuit 76, the output of which is connected to the clock signal input terminals of each of the flip-flops 68 and 71-73. The clock signal, itself, is applied by way of a clock signal input terminal 77 to the other input terminal of the AND gate 76. The output terminals of the four flip-flops 68 and 71-73 are connected to a common OR circuit 79, the output terminal of which is identified by reference numeral 81.

The CRC decoder 55 includes an input terminal 82 connected to one input terminal of an exclusive-OR gate 83, the output of which is connected to the D terminal of a D-type flip-flop 84. The output of the flip-flop 84 is connected to one input terminal of an exclusive-OR gate 86, and the output of the exclusive-OR gate 86 is connected to the D input terminal of the first of three D-type flip-flops 87-89 connected in sequence. The output of the flip-flop 89 is connected back to second input terminals of the exclusive-OR gates 83 and 86. A gating signal input terminal 91 is connected to a second input terminal of an AND gate 92, the other input terminal of which is connected to the clock signal input terminal 77. The output of the AND gate 92 is connected to the clock terminals of each of the flip-flops 84 and 87-89. The output terminals of all four of the flip-flops 84 and 87-89 are connected to input terminals of an OR gate 93 that has an output terminal 94.

The input signal Aj applied to the input terminal 66 should be, unless it contains an error, exactly like the Aj signal of the composite digital word represented in Figure 3E. This section of the digital word is 36 bits long and so the gate signal illustrated in Figure 10A is applied to the gate signal input terminal 74 to enable the AND gate 76 to allow clock signals applied to the input terminal 77 to clock 36 bits of information into the decoder 54, starting just after each word synchronizing HD shown in Figure 10C. If there are no errors in the signal Aj, the signal a at the output terminal 81 will be "0", but if there are any errors, the signal a will be "1".

In a similar manner, the gates signal shown in Figure 10B is applied to the gating signal input terminal 91 to allow 28 bits of information to be clocked into the decoder 55 starting one bit after the word synchronizing pulse HD. If there are no errors in the Bj signal applied to the input terminal 82, the output signal ss at the output terminal 94 will be "0", but if there are errors, the signal ss will be "1".

The signals Aj 18 and Bi~ls, which may be referred to as extended word signals because they include CRC components, are applied to the gate circuits 56 and 58, respectively.

These gate circuits allow only the basic information signals to pass through and delete the respective CRC signals appended to the basic information signals. Thus, at the output of the gate circuit 56, the signal should correspond exactly to the first 32 bits of the signal illustrated symbolically in Figure 3B and at the output of the gate circuit 58, the signals shduld correspond to the signal illustrated symbolically in Figure 3C, assuming, in each instance, that no error has entered either of the information signals.

The respective information signals, stripped of the CRC pulses, are re-expanded by the respective time expansion circuits 57 and 59. Thus, at the output of the time expansion circuit 57, the signal should be exactly like the signal symbolically illustrated in Figure 3A, if there is no error in the processed signal. The signal at the output terminal of the time expansion circuit 59 should be similar to the signal in Figure 3A except that it has a total of 24 bits instead of 32 bits.

The output signals of the time expansion circuit 57 and 59 are compared for coincidence in the coincidence detector 62. However since the output signal of the expansion circuit 59 has only the 24 most significant bits of the 32 bit output signal of the expansion circuit 57, the latter signal may be truncated in the gate 60 to delete its eight least significant bits so that the two signals applied to the coincidence detector 62 will each have the same number of bits of the highest order. These two signals must have the same number of bits, if they contain information for two stereophonic signals, because such signals contain two MSB signals and each of these MSB signals in the primary signal must be compared with the corresponding two MSB signals of the secondary signal.

The two signals to be compared in the coincidence detecting circuit are illustrated in Figures 12A and 12B. The signal in Figure 12A is the secondary signal, which was originally generated in truncated form to include only the highest orders of bits from M to L' of the pimary signal. Since this signal, as symbolically represented in Figure 12A, contains only information bits and no CRC bits, it is simply labeled as signal B. The truncated primary signal, which now has the same number of bits as the secondary signal B is referred to as the signal A' to distinguish it from the full range, or non-truncated, primary signal A. If the two signals B and A' coincide bit for bit, the coincidence detector 62 will produce an output signal y having a value of "1". However, if the two signals B and A' do not coincide, the value of the output signal y will be "0".

The full range primary signal A at the output of the time expansion circuit 57 and the truncated secondary, or comparison, signal B at output of the time expansion circuit 59 are applied to separate terminals of the gate circuit 61, which selects one of these two signals to be passed through the gate circuit to an output terminal 64 for further processing.

Essentially, the gate 61 is like a switching circuit that connects either the expansion circuit 57 or the expansion circuit 59 to the output terminal 64.

The output signals a and P from the CRC decoders 54 and 55 and the output signal y from the coincidence detector 62 are all applied to the logic circuit 63 to generate therein signals to control the operation of the gate 61. The logic circuit is shown in greater detail in Figure 11. It includes four input terminals 96-99 to receive the signals a, ss, and y and the word synchronizing signal HD, respectively. The terminals 96 and 99 are connected to an AND gate 101 the output of which is connected to an inverter 102 and to one input terminal of each of two AND gates 103 and 104. The output terminals 97 and 99 are connected to two input terminals of another AND gate 106, the output terminal of which is connected to the other input terminal of the AND gate 104 and the input of an inverter 107.

The input terminals of the inverter 102 is connected to one of the input terminals of a NAND gate 108 and to one of the terminals of a second NAND gate 109. The output terminal of the AND gate 106 is connected to the other input terminal of the NAND gate 109. The output terminals of the two NAND gates 108 and 109 are connected to two input terminals of a third NAND gate 110, The output terminal of the NAND gate 110 and the input terminal 98 are connected to the two input terminals of an OR gate 111, and the input terminal 98 is also connected to the input terminal of an inverter 112. The output terminal of the OR gate 111 and the input terminal 99 are connected to the two input terminals of another NAND gate 113, the output terminal of which is connected to the set terminal of a flip-flop 114. The input terminal 99 is connected to the reset terminal of this flip-flop and to the reset terminals of two other flip-flops 115 and 116 as well as to one of the input terminals of an AND gate 117.

The other input terminal of the AND gate 117 is connected to the output terminal of the inverter 112, and the output terminal of the AND gate 117 is connected to one of the input terminals of each of two NAND gates 118 and 119. The output terminals of the AND gates 103 and 104 are connected, respectively, to second input terminals of the NAND gates 118 and 119 respectively, and the output terminals of these NAND gates are connected, respectively, to the set input terminals of the flip-flop 115 and 116. The three flip-flops 114-116 have output terminals 121-123, respectively.

The circuit of the gate 61 controlled by the logic circuit 53 is shown in somewhat more detail in Figure 14. It includes two input terminals 126 and 127 connected, respectively, to output terminals of the time expansion circuits 57 and 59, respectively. In series with the leads from the input terminals 126 are two electronic switching circuits 128 and 129, which are controlled by the flip-flops 114 and 115, respectively, as indicated by the labels on the arrows adjacent the symbols of the switching circuit.

The switching circuits 128 and 129 are connected together to the input terminal of a delay or memory device 131 having sufficient capacity to store enough signal bits to equal one full range primary signal A. For example, the device 131 may be a 32 bit shift register. The output of the device is connected to one terminal of a double-throw switching circuit 132, the arm of which is connected to the output terminal 64. The other terminal of the switching circuit 132 is directly connected to the output terminals of the switching circuits 128 and 129.

As indicated by the arrow adjacent the symbol of the switching circuit 132, the state of conductivity of that switching circuit is controlled by the flip-flop 116.

The logical relationship between the signals a, ss, and y and the conditions of the signals transmitted from the gate 61 to the output terminal 64 is summarized in the truth table in Figure 13. When the signal a has the value "1" after comparing one digital word of the truncated primary signal A' with a corresponding digital word of the secondary signal B, the truth table indicates that the signal A will be transmitted through the switching circuit 132 to the output terminal 64, no matter whether the signals a and ss are "0" (which is probable if the value of signal a is "1") or "1".

When a = 0, and y = 0, indicating that there are no errors in the primary signal A, the switching circuit 128 is also closed and the switching circuit 132 connects the primary signal A directly to the output terminal 64. This condition exists whether or not the signal ss also has the value "0".

On the other hand, if a = 1, indicating that there are errors in the primary signal A, but ss = 0, indicating that there are no errors in the secondary signal B instead of the full range signal A only means that the smallest changes in amplitude represented by the least significant bits will not be present in the signal at the terminal 64, but the omission of these lowest order bits for one digital word or even for several words will be virtually unnoticeable.

If there are errors in both signals A and B, a = ss = 1. In such rare instances, both the switching circuits 128 and 129 will be open-circuited until the next digital word is measured.

However, instead of suddenly dropping the digital signal at the output terminal 64 to a zero condition for one digital word interval, which might cause a sudden high-amplitude negative signal to be produced by the D-A converters 14L and 14R, the switching circuit 132 is actuated to connect the output terminal of the memory or delay device 131 to the output terminal 64. The device 131 has received whichever word signal was carried through the switching circuit 132, and has replaced each word signal with the next word signal as long as one of the switching circuits 128 or 129 was conductive. When neither is conductive the device 131 still has the last-used word signal in it and, can unload this signal to the output terminal via the switching circuit 132. This simply means that for a very short interval of time, the output signal will remain constant. This is less likely to be objectionable than if the. output signal value were allowed to vary sharply.

As an alternative to providing the storage device 131, the circuit may be arranged to interrupt the supply of clock pulses to the D-A converters in Figure 2 for one digital word interval if a = ss = 1, thereby effectively holding the output signal level relatively constant briefly.

The operation of the circuit in Figure 11 to obtain the signals to control the switching circuits 128, 129 and 132 in Figure 14 will now be described.

Various circuit point within the circuit in Figure 11 are labeled alphabetically for ease of description of the operation of the circuit. The three signals a, ss, and y establish the operating conditions of the circuit 53, while the word synchronizing pulses applied to the terminal 99 initiate each operation. That is, it is only when a word synchronizing signal pulse is applied to the input terminal 99 and from there to each of the reset terminals of the flip-flops 114 - 116 that the output terminals 121-123 will be forced to take on the values determined by the logical "0" or "1" values of the signals a, ss, and y.

If the truncated primary signal A' exactly corresponds to the secondary signal B in the coincidence detector 62, thereby making the signal y = 1, this signal point g at the output of the OR gate 111 will have value "1", no matter what the output value of the NAND gate 110 happens to be. As a result, the circuit point h at the output of the NAND gate 113 and at the set terminal of the flip-flop 114 will also be "1" until the word synchronizing signals applied to the other input terminal of the NAND gate 113 raises that input terminal to the value of "1", thereby dropping the output signal at the circuit point h to "0" for the duration of the word synchronizing pulse. This causes the output terminal 121 to have the value "1". As previously stated, this closes the switching circuit 128 in Figure 14 and causes the switching circuit 132 to conduct the full range signal A through to the output terminal 64.

In the foregoing conditions, when y = 1, the inverter 112 inverts this to the value "0" thus disabling the AND gate 117 and holding the circuit point m at "0". This prevents either of the NAND gates 118 or 119 from responding to a word synchronizing pulse to set either of the flip-flops 115 and 116. As a result, both of the output terminals 122 and 123 have the value "0".

When y = 0, it indicates that there is an error either in the truncated primary signal A' or in the secondary signal B compared therewith. If it is assumed that the error is in the secondary signal B, the signal B will have the value "1" and the signal a will have the value "0". As a result, the circuit point a will have the value "0" and the voltage level at the circuit point b at the output of the AND gate 106 will correspond to the word synchronizing signal applied to the input terminal 99. The inverter 102 will invert the value "0" at the circuit point a to the value "1" at the circuit point c. The inverter 107 will invert the word synchronizing pulse signal at the circuit point b and apply this inverted synchronizing pulse to the NAND gate 108. The conditions of the circuit points c and d will cause the output of the NAND gate 108 at the circuit point e to follow the word synchronizing pulse. At the same time, the conditions of the circuit points b and c at the input terminals to the NAND gate 109 will cause the output of that NAND gate at the circuit point f to be the inverse of the word synchronizing pulse signal, and therefore, opposite to the signal at the point e at all times. Thus, one or the other of the input terminals of the NAND gate 110 will always be at the value "0", and therefore the output terminal of this NAND gate will always be "1" in response to such input conditions. This will cause the output of the OR gate 111 following the NAND gate 110 to have the value "1" at the circuit point g. This represents the same condition that prevailed when the value of the signal y was "1", and so the output terminal 121 of the flip-flop 114 will be at the value "1", causing the switching circuit in Figure 14 to be closed so as to transmit the signal A to the output terminal 64.

Since the output of the AND gate at point a is "0", both of the AND gates 103 and 104 will have "0" values at their output terminals and, hence, at the circuit points i and j. These conditions keep the output values of the NAND gates 118 and 119 at the points k and I at "1" so that the output terminals 122 and 123 of the flip-flops 115 and 116 remain at "0", keeping the switching circuits 129 and 132 in Figure 14 in the conditions shown.

The condition just described in detail in which a = 0, y = 0, and ss = 1 is Condition II in the following Table in which the initials W.S. stand for word synchronizing pulse and the symbol W.S. stands for the inverted word synchronizing pulse. The previous condition in which y = 1 is Condition V in the Table.

If there is an error in the primary signal A but none in the secondary signal B, the value of a will be "1" and the value of B will be "0". The value of y will be "0". This is Condition III in the Table and therefore need not be described in words.

If both the primary and secondary signals A and B have errors so that a = (3= 1 and o = 0, the logic circuit will operate according to Condition IV in the Table, and as previously described will cause the switching circuit 132 in Figure 14 to draw a replacement from the storage device 131. In effect this makes three signals available to the output terminal 64 without delay; signal A as a first choice, signal B as a second choice, and the signal stored in the device 131 as a third choice.

TABLE Circuit Points Conditions I II III IV V 96( input) 0 0 1 1 97( input) 0 1 0 1 99( input) 0 0 0 0 1 a 0 0 W.S. W.S.

b 0 W.S. 0 W.S.

c 1 1 W.S. W.S.

d 1 W.S. 1 W.S.

e 0 W.S. W.S. W.S.

f 1 W.S. 1 1 g 1 1 W.S. W.S. 1 h W.S. W.S. 1 1 O O 0 W.S. O O 0 O 0 W.S.

k 1 1 W.S. 1 1 1 1 1 1 W.S. 1 m W.S. W.S. W.S. W.S. 0 121 1 1 0 0 1 122 0 0 1 0 0 123 0 0 0 1 0 Condition I represents a state in which analysis of the CRC code in each of the decoders 54 and 55 indicates that there is no error in either the A signal nor the B signal, so that a = B = 0. Yet the statement y = 0 means that the most significant bits of the two signals do not coincide so that at least one of them must be in error. This requires that the error be such as to make the error cause the information signal or cause both the CRC signal and information signal to shift precisely to a new value that will be decoded as a permissible, and therefore correct, value. Such a situation, while possible, has a very low probability of occurrence. The logic circuit 63 is arranged under this condition to select the primary signal A to be passed through the gate 61 to the output terminal 64.

Figure 15A shows the effect of this invention in overcoming a noise burst that extends over three composite digital words including the primary signals Ai through Ai+2 and the secondary signals By 18 through Bj 1s grouped within this interval. As indicated in Figure 3F, this is one horizontal line interval.

As shown in Figure 15B, after the composition word signals have been separated and the error-free primary signals Aj.18 through Aj 15 brought into the same time intervals with the erroneous secondary signals By 18 through Bi-15, the logic circuit 53 can easily select the error-free primary signals for further processing, for example in the serial-to- parallel converter 13 and on beyond.

In the same way, when the erroneous signals A through Aj+3 have been brought into the same time intervals with the error-free related secondary signals Bi through Bj+3, the logic circuit 63 can easily select the error-free secondary signals for further processing. Thus, the advantage of the invention is obtained.

WHAT WE CLAIM IS: 1. A method of processing first multi-bit digital signals grouped into digital words, comprising the steps of selecting, from each digital word of the first multi-bit signals, a secondary group of bits consisting of only predetermined ones of the bits of the respective said digital word, each digital word and the secondary group of bits thereof comprising related signal groups; delaying one of said signal groups relative to the other of said groups and combining the delayed signal groups in interleaved sequence with subsequent, relatively undelayed, signal groups to form composite, sequential signals.

2. The method of Claim 1 wherein the step of delaying one of said signal groups comprises delaying said secondary group relative to the digital word in each of said related signal groups.

3. The method of Claim 1 in which said digital signals comprise pulse code modulated multi-bit signals with a range of higher to lower bit orders between a most significant bit and a least significant bit, and the method includes the step of selecting a plurality of higher bit orders of said digital word to be said secondary group of bits.

4. The method of Claim 3 and including the steps of time-compressing the composite sequential signals.

5. The method of Claim 3 and including the steps of time-compressing said digital word and separately time-compressing said secondary group of bits before combining said delayed signal groups with said relatively undelayed signal groups.

6. The method of Claim 5 and including the steps of adding a group of first cyclic redundancy check code signals to the time-compressed delayed signal groups and a group of second cyclic redundancy check code signals to the time-compressed relatively undelayed signal groups before combining said time-compressed delayed signal groups and said time-compressed relatively undelayed signal groups.

7. The method of Claim 6 wherein said step of delaying said one of said signal groups includes delaying such groups by the length of time occupied by one of said timecompressed relatively undelayed signal groups plus the time occupied by the first cyclic redundancy check code signals therefore plus an integer N times a full interval occupied by a complete set of said time-compressed relatively undelayed signals and said first and second cyclic redundancy check code signal, wherein N 2 0.

8. The method of Claim 7 wherein all of said time-compressed signals are compressed by a ratio of 2:1.

9. The method of Claim 6 wherein each of said first redundancy check code signals comprises P bits immediately following each said time-compressed delayed signal group, and each group of said second redundancy check code signals comprises P bits immediately following each said time-compressed relatively undelayed signal group.

10. The method of Claim 9 wherein the total number of bits in each said secondary group of bits is 2P less than the total number of bits in each said digital word.

11. The method of Claim 10 wherein the number of bits in each said digital word is 32, the number of bits of each secondary group of bits is 24 and the total number of bits in said cyclic redundancy check code signals is 8.

12. The method of Claim 1 and including the steps of time-compressing said digital word and said secondary group of bits before combining the digital word and the secondary group of bits processing the combined signals, and re-expanding the processed signals to their original time durations.

13. The method of Claim 12 and comprising selectively delaying the processed signals to re-establish the original relative timing between re-expanded digital word and the secondary group of bits.

14. The method of Claim 1 wherein said composite sequential signals include a pluarlity of first multi-digit cyclic redundancy check code binary signals, each related to one of said digital words and preceding the next group of said secondary groups of bits and a plurality of second multi-digit cyclic redundancy check code binary signals, each related to one of said secondary group of bits and each following a group of said secondary groups of bits and preceding the next group of said digital words, said method being characterised by comparing the delayed and relatively undelayed signals and choosing for further processing the one of said compared signals which is most error free.

15. The method of Claim 14 and including the steps of storing the chosen signal until the time of making the next comparison and if, at that time, neither of the compared signals is error free, choosing the stored signal for further processing another time.

16. The method of Claim 14 wherein the step of choosing the one of said compared signals which is most error free comprises choosing said digital words if both of said compared signals have no errors.

**WARNING** end of DESC field may overlap start of CLMS **.

Claims

**WARNING** start of CLMS field may overlap end of DESC **. circuit 63 can easily select the error-free secondary signals for further processing. Thus, the advantage of the invention is obtained. WHAT WE CLAIM IS:

1. A method of processing first multi-bit digital signals grouped into digital words, comprising the steps of selecting, from each digital word of the first multi-bit signals, a secondary group of bits consisting of only predetermined ones of the bits of the respective said digital word, each digital word and the secondary group of bits thereof comprising related signal groups; delaying one of said signal groups relative to the other of said groups and combining the delayed signal groups in interleaved sequence with subsequent, relatively undelayed, signal groups to form composite, sequential signals.

17. Apparatus for processing a primary multi-bit digital signal, said apparatus

comprising an input to receive said primary signal into said apparatus; a first. gate to generate a secondary multi-bit digital signal containing only part of said primary signal; a delay connected to one of said input and said first gate to delay the signal from a selected one thereof relative to the signal from the other thereof by a predetermined interval of time; a second gate connected to said delay and to said other of said input and said first gate to conduct, alternatively, a delayed signal and an undelayed signal; a first time-compressor connected to said input to receive and store said primary signal at one rate and to extract said primary signal from storage at a higher rate in first groups spaced apart in time; a second time-compressor connected to said first gate to receive and store said secondary signal at said one rate and to extract said secondary signal from- storage at said higher rate in second groups spaced apart in time; and third and fourth gates connected to said first and second time compressors respectively to extract said groups of said primary signal in alternation with said groups of said secondary signal.

18. The apparatus of Claim 17 and including a first check code generator connected to said input to receive said primary signal to generate a first check code signal based on each of said first signal groups; a second check code generator connected to said first gate to receive said secondary signal to generate a second check code signal based on said secondary signal groups; said third gate being connected to said first check code generator; said fourth gate being connected to said second check code generator, said second, third and fourth gates being effective to transmit, in sequence, an interval of predetermined length of said primary signal, an interval of predetermined'length of said first check code signal, an interval of predetermined length of said secondary signal, and an interval of predetermined length of said second check code signal as a composite digital word signal.

19. The apparatus of Claim 18 and including a fifth gate for separating each interval of said primary signal and each interval of said first check code signal from each interval of said secondary signal and each interval of said second check code signal; a first decoder connected to said decoding gate to decode said primary signal and said first check code signal to generate a first logic signal having a value based on the existence of error in the primary signal portion of said digital word signal; a second decoder connected to said fifth gate to decode said secondary signal and said second check code signal to generate a second logic signal having a value based on the existence of error in the secondary signal portion of said digital word signal; and switches connected to said first and second decoders and to an output circuit to be controlled by said first and second logic signals to connect either said primary signal or said secondary signal to said output circuit.

20. The apparatus of Claim 17 and including first check code generator connected to said input to receive said primary signal and to generate a first check code signal based on each of said first signal groups; a second check code generator connected to said first gate to receive said secondary signal and to generate a second check code signal based on said secondary signal group; said third gate being connected to said input and said first check code generator to transmit to said second gate, in sequence, an interval of predetermined length of said primary signal, and an interval of predetermined length of said first check code signal; said fourth gate being connected to said first gate and said second check code generator to transmit to said second gate, in sequence, an interval of predetermined length of said second check code signal; a fifth gate for receiving the composite signal produced by said second gate to separate said primary signal from said secondary signal; a second delay selectively connected to said fifth gate to bring each said secondary digital word and the primary digital word related thereto in information content substantially into time coincidence; and a coincidence detector connected to said second delay and to said fifth gate to compare each said secondary digital word to the primary digital word related thereto in information content and to generate a logic signal having a predetermined value if at least a predetermined portion of each of the compared digital signals is identical.

21. The apparatus of Claim 20 wherein said secondary digital signal includes only the higher order bits of said primary digital signal related in information content, and said apparatus comprises a sixth gate to truncate each said primary digital signal to match the bit orders in said secondary digital signal, the matched bit orders comprising said predetermined portion.

22. The apparatus of Claim 20 and includes switches connected to said coincidence detector to select said primary digital signal if the compared digital signals are identical.

23. The apparatus of Claim 22 and including a decoder connected to said fifth gate to generate a second logic signal having a predetermind value if said primary digital signal is error-free; and a logic circuit connected to said decoder and to one of said switches to be actuated by said second logic signal to select said primary digital word if said primary digital signal is error-free.

24. The apparatus of Claim 23 wherein said second delay is connected in series between said fifth gate and said decoder.

25. The apparatus of Claim 23 and including a second decoder connected to said fifth gate to generate a third logic signal having a predetermined value if said secondary digital signal is error-free, said logic circuit being connected to said second decoder to cause one of said switches to be actuated by said third logic signal to select said secondary digital signal if said primary digital signal has an error and said secondary digital signal is error-free.

26. The apparatus of Claim 25 and including a storage device to store and to repeat a previously selected one of said digital signals, said logic circuit being responsive to said first-named, said second, and said third logic signals to actuate one of said switches to select the previously selected one of said digital signals if there are errors in the primary and secondary digital signals controlling the generation of said second and third logic signals.

27. A method of processing digital signals, substantially as hereinbefore described with reference to the accompanying drawings.

28. Apparatus for processing a digital signal substantially as hereinbefore described with reference to and as shown by the accompanying drawings.