JPH0248835A - Adaptive differential pcm system - Google Patents

Abstract

PURPOSE:To reduce the difference between a predictive signal and an actual signal by adopting the oversampling technology to generate the predictive signal at the next sampling time by primary linear prediction and shortening the period obtaining the predictive signal more than the actual one sampling period. CONSTITUTION:A signal fed to a signal processing system is fed to an adaptive inverse quantizer 17 and a predictive signal generating circuit 16, and an output signal of the circuit 16 and a reference signal generated in a reference signal generating circuit 19 are compared by a comparator 18. Then the result obtained by the comparator 18 is inputted to the inverse quantizer 17 and the circuit 16, where prescribed adaptive inverse quantization is applied and a signal nearly equal to an output signal from a subtractor 2 in a coding processing system is generated and the result is fed to an adder 20. Then an output from the adder 20 is given to a D/A converter 22 and given to a delay device 21 and the output signal of the delay device 21 is added to the output of the inverse quantizer 17 by the adder 20. Thus, the output of the adder 20 is made correspondent to the output of an A/D converter 6 and a signal outputted from the converter 22 to a terminal 2 is made correspondent to the analog signal fed to the terminal 1.

Description

DETAILED DESCRIPTION OF THE INVENTION (Field of Industrial Application) The present invention relates to an adaptive differential PCM system, particularly a digital signal transmission system, a wired transmission system using a telephone line.

The present invention relates to an adaptive differential PCM method suitable for audio recording/playback systems using semiconductor memories and the like.

(Prior Art) This is an adaptive differential PCM method used for transmitting, recording and reproducing digital signals.

Many different formats have been proposed over the years.
In the present applicant company, for example, Japanese Patent Publication No. 59-52585
In Japanese Patent Publication No. 60-53971, differential pulse code modulation (
A differential PCM) signal may be adaptively processed to perform bit shifting based on a predicted value obtained according to the signal level of the input differential signal, or a differential pulse code modulation (differential PCM) signal to be transmitted with a limited number of bits may be used. Differential PCM) The bit shift control of the signal is performed based on a predicted value that also uses the detection results of large amplitude states and high frequency states of the original analog signal. We have proposed a digital signal transmission system using differential PCM signals that can reduce gradient overload noise and granular noise specific to signals.

FIG. 11 is a block diagram showing the outline of the configuration principle of the digital signal transmission method using differential PCM signals described in the above-mentioned publication. In FIG. This is an input terminal for the original signal in the form of an analog signal, and the analog signal supplied to the input terminal 1 is input to an analog-to-digital converter (AD).
In step 6, Kabamotoization and quantization are performed and the digital signal
n, and is supplied to the subtractor 7 as the minuend signal Xn.

The subtractor 7 receives 11 more than the minuend signal Xn.
Since the signal Yn-1 corresponding to the signal before the 11th cycle is supplied as the subtraction signal, the subtracter 7 outputs (X
The difference signal dn represented by n - (Yn-1)) = d n is output and supplied to the adaptive quantizer 8, and the adaptive quantizer 8 adaptively processes the input difference signal dn. The signal P is adaptively quantized to a predetermined number of bits.
The adaptive processing in the adaptive quantizer 8 is performed based on the control signal supplied from the comparator 12 in the adaptive control circuit.

The output signal Pn from the adaptive quantizer 8 described above is:

It is supplied to the decoding system as an output signal of the encoding system via the transmission system or recording medium, and is also subjected to adaptive inverse quantization.
and a predicted signal generation circuit (PGC) 11.

FIG. 12 shows the predicted signal generation circuit (PGC) 11 shown in FIG. 11 described above (the predicted signal generation circuit (PGC) 18 used in the decoding system in FIG. 11 also has the same configuration). 12 is a block diagram showing a specific configuration example of the predictive signal generation circuit (PGC) shown in FIG. 12.
11 The terminals 3, 4.5, etc. in (16) are the first
1 Prediction signal generation circuit (PGC) 11 shown in Figure 1
．． 16 corresponds to terminals 3, 4.5.

The predicted signal generation circuit (PGC) 11 described above may have any configuration as long as it can generate a predicted value of the signal at the next sampling time;
Prediction signal generation circuit (PGC) 11 illustrated in Fig. 2
1, for example, No. 2 described in Japanese Patent Publication No. 60-53971
This is similar to the predicted signal generation circuit shown in the figure.

In the predicted signal generation circuit 11 shown in FIG. and is supplied to the subtracter 30 as a signal to be subtracted.

The above-mentioned subtracter 30 is connected to the subtracter 30 via the delay device 29 of one sampling period.
By being given as a subtraction signal to ) given to 12.

The prediction signal qn supplied to the comparator 12 is compared with a reference signal generated by the reference signal generation circuit 13 (the minimum reference signal is oooo, the maximum reference signal is 0111).

The comparator 12 receives the reference signal (the minimum reference signal is oooo,
The signal level of the maximum reference signal (0111) is compared with the predicted signal qn supplied thereto, and a signal ( Maximum reference signal 0111
A signal with a level exceeding the quantization width of the signal currently being transmitted (a signal with a level below the minimum reference signal 0000), or a signal with a level below the minimum reference signal 0000, or the compared predicted signal qn Maximum reference signal 0
A control signal Cn is supplied to the adaptive quantizer 8 according to the comparison result of whether the signal is between 111 and the minimum reference signal 0000.

Therefore, the adaptive quantizer 8 performs, for example, the following adaptive quantization operation. First, (1) as a result of the comparison in the comparator 12, the compared prediction signal q
If n is a signal between the maximum reference signal 0111 and the minimum reference signal 0000, the adaptive quantizer 8 operates under the same operating conditions as the previous operating conditions. continue.

(2) The predicted signal qn compared in the above comparison @12 is a signal whose signal level exceeds the quantization width of the signal currently being transmitted (a signal whose level exceeds the maximum reference signal 0111). In response to the comparison result, the adaptive quantizer 8 causes the quantization operation to be performed at the next sampling time to be performed with the quantization width changed by one step.

(3) The predicted signal qn compared in the above comparison @12 is a signal with a signal level lower than the quantization width of the signal currently being transmitted (a signal with a level lower than the minimum reference signal 0000). The adaptive quantizer 8 corresponds to the comparison result.
Then, the quantization operation performed at the next sampling time is performed with the quantization width changed by one step.

In addition to being supplied from the comparator 12 to the adaptive quantizer 8, the control signals described above are supplied to the adaptive inverse quantizer 9 via the inverter 10 corresponding to each of the cases (1) to (3) described above. be done.

The output signal Pn, which is supplied from the adaptive quantizer 188 to the decoding system via the transmission system or recording medium as an output signal of the encoding system as described above, is a prediction signal based on first-order linear prediction (slope prediction signal, differential The predicted signal) is adaptively quantized by an adaptive quantizer 8, transmitted, and supplied to a decoding system via a recording medium, etc., and is also supplied to an adaptive inverse quantizer II9.

The dequantization process performed on the signal Pn supplied to the adaptive dequantizer 9 described above is performed based on the control signal supplied from the comparison s12 in the adaptive control circuit via the inverter 1o, A signal dn substantially equal to the output signal dn from the subtracter 2 described above is generated and supplied to the adder 14.

The adder 14 adds the predicted signal Y(n-1) of one sampling period (1/fs) before and the signal dn to output the first predicted signal Yn, which is stored in the register. 1
Supply to 5. Said register 15 is supplied to it-
It operates as a delay device that delays one signal by one sampling period (1/fs) and outputs the delayed signal.

Next, the signal Pn supplied to the decoding system in FIG.
Adaptive inverse quantizer 17 and predictive signal generation circuit (PGC) 16
and will be supplied. The predicted signal generation circuit (PGC) 16 performs the same operation as described for the predicted signal generation circuit (PGC) 11 and outputs an output signal qn, which is supplied to the comparator 18.

In the comparison @is, the prediction signal generation circuit (PGC) 1 described above
The a signal qn of No. 6 is compared with a reference signal generated by the reference signal generation circuit 19 (minimum reference signal is 0000, maximum reference signal is 0111). The control signal Cn output from the comparator 18 is supplied to an adaptive inverse quantizer 17 and a predictive signal generation circuit (PGC) 16.

The adaptive dequantizer 17 performs a predetermined adaptive dequantization operation based on the control signal Cn supplied from the comparator 18, and the output signal dn from the subtracter 2 in the coding system described above is approximately equal to the output signal dn. Output the equal signal dn hat and feed it to the adder@20.

The output signal zn from the adder 20 is supplied to a digital-to-analog converter 22 and also to a delay device 21 which delays the signal by one sampling period, and the output signal of the delay device 21 described above is The adder 20 adds the signal to the output signal of the adaptive inverse quantizer 17 . The output signal zn from the adder 20 described above corresponds to the output signal Xn from the analog-to-digital converter 6 described above.

Output terminal 2 from the digital-to-analog converter 22 described above
The analog signal output to corresponds to the analog signal supplied to the input terminal 1.

(Problems to be Solved by the Invention) By the way, in the adaptive differential PCM method explained with reference to FIGS. 11 and 12, the predicted signal at the next sampling time (first-order linear prediction Because a predicted signal (slope predicted signal, differential predicted signal)) is generated, a problem has arisen in that, for example, near the peak value of the signal, the actual signal and predicted signal differ greatly.

FIG. 13 is a diagram for explaining the above problem,
In this FIG. 13, sequential sampling times tl, t2.
The solid line arrows shown at t3... indicate the actual signal magnitude.

Sequential sampling times tl, t2. The dotted arrows shown at t3, . . . indicate the magnitude of the predicted signal, which is the predicted signal magnitude at a point one sampling period later than the sampling time.

In FIG. 13, 1, for example, at time t3, time t4
The predicted signal q4, which is a prediction of the signal at time t4, differs greatly from the actual signal P4 at time t4, but if such a thing were to occur, it would be a hindrance to faithful signal transmission and recording/reproduction.

Improvement was required.

(Means for Solving the Problems) The present invention provides a solution to the large discrepancy between the predicted value and the actual value that conventionally occurs when the prediction is made based on the magnitude of the slope 1s before the normalization cycle. This is an adaptive differential PCM method in which the period for obtaining a predicted signal is made shorter than one actual sampling period to avoid occurrence of An adaptive differential PCM method that encodes a signal into an m-bit digital signal with a relationship n>m, and decodes the encoded m-bit digital signal into an n-bit digital signal, wherein the input signal and a subtracter that obtains the difference between the input signal and a signal corresponding to a signal one sampling period before, and an output signal of the subtracter is inputted to generate an adaptively quantized signal,
an adaptive quantizer that outputs the signal as an output signal of the encoding system; and a first circuit that dequantizes the output signal of the adaptive quantizer to generate a signal corresponding to the output signal of the subtracter. an adaptive inverse quantizer, an addition amount for adding the output signal of the first adaptive inverse quantizer and a signal corresponding to a signal one sampling period before the input signal; A delay device that delays the output signal by one sampling period, and prediction signal formation in a state where sampling frequency conversion is performed twice (however, K is 2 or more) on the output signal of the above-mentioned adaptive quantization amount. a first adaptive control circuit that generates a control signal for adaptive quantum processing, including means for generating a signal for forming a predicted signal, and means for generating a predicted signal by first-order linear prediction using the signal for forming a predicted signal; The signal generated by the coding system and the coding system described above is supplied as an input signal, and a sampling frequency conversion of twice (K is 2 or more) is performed on the input signal. a second adaptive control circuit for generating a control signal for adaptively controlling the meridian amount; a second adaptive inverse quantizer to which the signal is supplied as an input signal and whose operation is controlled by the output signal of the second adaptive control circuit; and an output of the second adaptive inverse quantizer. It is comprised of an adder to which a signal is supplied as one input signal, and means for supplying a signal obtained by delaying the output signal of the adder by one sampling period to the adder as the other input signal. An adaptive differential PCM method consisting of a decoding system that encodes an n-bit digital signal obtained by sampling and quantizing an analog signal into an m-bit digital signal with a relationship n>m, and This is an adaptive differential PCM method that decodes an m-bit digital signal into an n-bit digital signal, and obtains the difference between an input signal and a signal corresponding to a signal 1 sample period before the input signal. a sound attenuator, an adaptive quantizer that inputs the output signal of the subtracter, generates an adaptively quantized signal, and outputs it as an output signal of a coding system; and an output of the adaptive quantizer. a first adaptive inverse quantizer that inversely quantizes a signal to generate a signal corresponding to the output signal of the subtracter; an output signal of the first adaptive inverse quantizer and the input signal; (1) an adder that adds a signal corresponding to a signal before the Enomoto-Ihi period; a delay device that delays the output signal of the adder by one sampling period; The signal for each sampling period is divided into K
means for generating a predicted signal forming signal that is interpolated at points (2 or more).

a first adaptive control circuit that generates a control signal for first-order linear predictive control using the predictive signal forming signal; A predicted signal is formed in a state where the input signal is supplied as an input signal, and the signal for each sampling period in the input signal is interpolated at points (K is 2 or more) according to Newton's backward interpolation formula. a second adaptive control circuit that generates a primary line control signal using the predicted signal forming signal; and a signal generated by the coding system that is supplied as an input signal. a second adaptive inverse quantizer whose operation is controlled by the output signal of the second adaptive control circuit;
1 an adder to which the output signal of the second adaptive inverse quantizer is supplied as an input signal; and a signal obtained by delaying the output signal of the adder by one sampling period using a delay device. The adaptive differential PCM method consists of a decoding system configured with a means for supplying the other input signal to the device, and an n-bit digital signal obtained by sampling and quantizing an analog signal, with the relationship n>m. An adaptive differential PCM method that encodes an m-bit digital signal and decodes the encoded m-bit digital signal into an n-bit digital signal, wherein a subtracter that obtains the difference between the signal and the corresponding signal;
an adaptive quantizer that inputs the output signal of the subtracter, generates an adaptively quantized signal, and outputs it as an output signal of the encoding system; and an adaptive quantizer that dequantizes the output signal of the adaptive quantizer. a first adaptive inverse quantizer that generates a signal corresponding to the output signal of the above-mentioned subtracter; and 1 sampling of the output signal of the first adaptive inverse quantizer and the above-mentioned input signal. an adder that adds a signal corresponding to a signal of a previous period; a delay device that delays the output signal of the adder by one sampling period; and a next sampling time of the output signal of the adaptive quantizer. An interpolation prediction signal generation circuit that generates a prediction signal by interpolating the signal just before the interpolation according to Newton's forward interpolation formula, and a first adaptive control circuit that generates a control signal for controlling the interpolation prediction described above. The encoded system and the signal generated by the aforementioned encoding system are supplied as input signals, and the signal immediately before the next sampling time in the input signals is interpolated according to Newton's forward interpolation formula to generate a predicted signal. a second adaptive control circuit that generates a control signal for adaptively controlling the interpolation predictor to generate the interpolation predictor; a second adaptive control circuit whose operation is controlled by the output signal of the second adaptive control circuit;
an adaptive inverse quantizer, an adder to which the output signal of the second adaptive inverse quantizer is supplied as an input signal, and a delay device to delay the output signal of the adder by one sampling period. The present invention provides an adaptive differential PCM system comprising a decoding system constituted by means for supplying the signal obtained by adding the signal to the adder as the other input signal.

(Example) Hereinafter, the adaptive differential PCM method of the present invention will be described in detail with reference to the accompanying drawings. 1, 5, and 8 are block diagrams of six different embodiments of the adaptive differential PCM method of the present invention, and FIGS. 2 to 4 are block diagrams for explaining the embodiment shown in FIG. 6 and 7 are diagrams for explaining the embodiment shown in FIG. 5, and FIGS. 9 and 10 are diagrams for explaining the embodiment shown in FIG. 8.

First, the adaptive differential PCM of the present invention shown in FIG.
In order to avoid the large discrepancy between the predicted value and the actual value, which conventionally occurs when prediction is made based on the magnitude of the slope at the beginning of the sampling period R, the method is
An n-bit digital signal obtained by sampling and quantizing an analog signal by using oversampling technology so that the period for obtaining a predicted signal is shorter than one actual sampling period is Adaptive differential PCM that encodes the encoded m-bit digital signal into an n-bit digital signal; and decodes the encoded m-bit digital signal into an n-bit digital signal.
In this method, a subtracter 7 obtains a difference dn between an input signal Xn and a signal Yn-1 corresponding to a signal one sampling period before the input signal, and an output signal dn of the subtracter 7 is inputted. an adaptive quantizer 8 that generates an adaptively quantized signal Pn and outputs it as an output signal of the encoding system; A first adaptive inverse quantizer 9 that generates a signal dn corresponding to the output signal dn, and one sample of the output signal dn of the first adaptive inverse quantizer 9 and the input signal Xn. The signal Yn- corresponding to the signal before the cycle
1, a delay device 15 that delays the output signal of the adder 14 by one sampling period, and a delay device 15 that delays the output signal of the adder 14 by one sampling period;
In order to adaptively control a means (23) for generating a predicted signal forming signal that has been subjected to sampling frequency conversion (K is 2 or more) and a device using the above predicted signal forming signal. a first adaptive control circuit (11
~13), and the signal Pn generated by the above coding system is supplied as an input signal, and sampling is performed twice (however, K is 2 or more) as much as the input signal. means (25) for obtaining a frequency-converted predicted signal forming signal; and a second adaptive control circuit (16, 18, 18) for generating a control signal for adaptively controlling the predicted signal. 19) and a second adaptive inverse quantizer to which the signal Pn generated by the coding system is supplied as an input signal and whose operation is controlled by the output signal of the second adaptive control circuit. 17, an adder 20 to which the output signal dn of the second adaptive dequantizer 1117 and the output signal dn of the second adaptive dequantizer are supplied as input signals; addition 1
This is an adaptive differential PCM system consisting of a decoding system constituted by a means for supplying a signal obtained by delaying the output signal zn of 820 by one sampling period by a delay device 21 to the addition l120 as the other input signal. .

In the adaptive differential PCM method of the present invention shown in FIG. 1, the components corresponding to the conventional adaptive differential PCM method described with reference to FIG. The same drawing numerals are used as the drawing numerals shown in FIG.

In FIG. 1, 1 is an input terminal for an original signal in the form of an analog signal that is to be transmitted, recorded, or reproduced, and the analog signal supplied to the input terminal 1 is input to an analog-to-digital converter (AD). In step 6, sampling and quantization are performed to form a digital signal Xn, which is supplied to a subtracter 7 as a minuend signal Xn.

Since the subtractor 7 is supplied with the signal Yn-1, which corresponds to a signal one sampling period earlier than the minuend signal Xn, as the subtraction signal, the subtractor 7 receives (X n
−(Yn−1))=difference signal dn denoted by n
is outputted and supplied to the adaptive quantizer 8, which adaptively processes the input difference signal dn to produce a signal Pn that is adaptively quantized to a predetermined number of bits.
However, the most adaptive processing in the adaptive quantizer 8 is performed based on the control signal Cn supplied from the comparator 12 in the adaptive control circuit.

The output signal Pn from the adaptive quantizer 8 described above is supplied to the decoding system via the transmission system or recording medium as an output signal of the encoding system, and is also supplied to the prediction signal generation circuit (PGC).
) 11, and an adaptive inverse quantizer 9 and an adaptive quantizer 8
It is also supplied to the Fs conversion circuit 23, which is configured to generate a predicted signal forming signal that has been subjected to sampling frequency conversion twice (K is 2 or more) to the output signal of (In the following description with reference to FIGS. 2 to 4, the case where K is 5 is exemplified).

The above-mentioned Fs conversion circuit 23 (the same goes for the Fs conversion circuit 25 provided in the decoding system) is a block 31 in FIG.
For example, the block 32 is configured by a digital filter having the functions shown in the block 32 and the block 32.

That is, the Fs conversion circuit 23 described above (also the F8 conversion circuit 25 provided in the decoding system) converts the signal from the adaptive quantizer 8 as shown in block 31 in FIG. Between two consecutive output signals on the time axis, for example, as shown in FIG. 3(a), the output signal Pn-1 from the adaptive quantizer 8 and the output signal Pn-1 When a sampling frequency conversion of 5 times (an example when K in K times is 5) is performed on the output signal of the adaptive quantizer 8 between the output signal Pn that appears after one sampling period from , the one sampling period mentioned above is (K
−1)=0 is inserted into each of the four equally divided time positions as shown in FIG. 3(b), and then as shown in block 32 in FIG. The output signal Pn from the adaptive quantizer 8 is passed through a low-pass filter to prevent aliasing distortion (see (a) to (c) in FIG. 4). -1 and the output signal Pn that is cursed after one sampling period from the output signal Pn-1, the time position equally divided into (K-1) times 4 is shown in FIG. ) as shown in P'5n-1,
P'5n-2, P'5rr3. A signal interpolated by each signal of P'5n-4 is output and supplied to a predictive signal generation circuit (PGC) 11 having a function as shown by block 33 in FIG.

FIG. 12 shows the predicted signal generation circuit (PGC) 11 shown in FIG. 1 described above (the predicted signal generation circuit (PAC) 16 used in the decoding system in FIG. 12 is a block diagram showing a specific example of the configuration of the prediction signal generation circuit (PGC) 1 shown in FIG. 12.
Terminal 3.4.5 etc. in 1 (16) corresponds to terminal 3.4.5 shown in predictive signal generation circuit (PGC) 11.16 shown in FIG.

The predicted signal generation circuit (PGC) 11 described above may have any configuration as long as it can generate a predicted value of the signal at the next sampling time.
Prediction signal generation circuit (PGC) 11 illustrated in Fig. 2
1 For example, No. 2 described in Japanese Patent Publication No. 60-53971
The prediction signal generation circuit 11 shown in FIG. 12 is similar to the prediction signal generation circuit shown in the figure.
, the output signals Pn, P' from the Fs conversion circuit 23
5n-1, P'5n-2, P'5n-3, P'5n-4
For each signal of By being given as a subtraction signal to the subtracter 30, 1 is sent from the subtracter 30 to the a terminal 5.
Predicted signal qn at the next sampling time by linear prediction
That is, a signal qn represented by the following equation (1) is sent out as an output signal and given to the comparator (level comparison@) 12.

The prediction signal qn supplied to the comparator 12 is compared with a reference signal generated by the reference signal generation circuit 13 (minimum reference signal is 0000, maximum reference signal is 0111).

The comparator 12 uses the reference signal (the minimum reference signal is ooooo) supplied from the reference signal generation circuit 13 described above.
, the maximum reference signal is 0111) and the predicted signal qn supplied thereto, and the signal level of the compared predicted signal qn exceeds the quantization width of the signal currently being transmitted. (Maximum reference signal 011
1), or a signal with a level below the quantization width of the signal being transmitted (a signal with a level below the minimum reference signal 0000), or a compared predicted signal. A control signal Cn is supplied to the adaptive quantizer 8 according to the comparison result of whether qn is a signal between the maximum reference signal 0111 and the minimum reference signal 0000.

Therefore, the adaptive quantizer 8 performs, for example, the following adaptive quantization operation. First, (1) As a result of the comparison in the comparator 12, the compared prediction signal qn is determined to be the maximum reference signal. If the signal is determined to be between 0111 and the minimum reference signal 0000, the adaptive quantizer 8 continues to operate under the same operating conditions as before.

(2) The predicted signal qn compared by the comparator 12 described above is a signal whose signal level exceeds the quantization width of the signal currently being transmitted (a signal whose level exceeds the maximum reference signal 0111). In response to the comparison result, the adaptive quantizer 8 causes the quantization operation to be performed at the next sampling time to be performed with the quantization width changed by one step.

(3) The predicted signal qn compared by the comparator 12 described above is a signal with a signal level lower than the quantization width of the signal currently being transmitted (a signal with a level lower than the minimum reference signal aaoo). The adaptive quantizer 8 corresponds to the comparison result.
Then, the quantization operation performed at the next sampling time is performed with the quantization width changed by one step.

In addition to being supplied from the comparator 12 to the adaptive quantizer 8, the control signals described above are supplied to the adaptive inverse quantizer 9 via the inverter 10 corresponding to each of the cases (1) to (3) described above. be done.

The output signal Pn supplied from the adaptive quantizer 8 to the decoding system via the transmission system or recording medium as an output signal of the encoding system as described above is a predicted signal (
adaptive quantizer 8 (gradient prediction signal, differential prediction signal)
The signal is adaptively quantized and transmitted, supplied to a decoding system via a recording medium, etc., and also supplied to an adaptive inverse quantizer 9.

The dequantization process performed on the signal Pn supplied to the adaptive dequantizer 9 described above is performed based on the control signal supplied from the comparator 12 in the adaptive control circuit via the inverter 10. , generates a signal dn substantially equal to the output signal dn from the subtracter 2 described above, and supplies it to the adder 14.

The adder 14 adds the predicted signal Y(n-1) from one sampling period (1/fs) ago and the signal dn, outputs the next predicted signal Yn, and stores it in the register. 1
Supply to 5. The register 15 operates as a delay device that delays the signal supplied thereto by 1@normal period (1/fs) and outputs the delayed signal.

Next, the signal Pn supplied to the decoding system in FIG. 1 is supplied to an adaptive inverse quantizer 17 and a predictive signal generation circuit (PGC) 16. The predicted signal generation circuit (PGC) 16 performs the same operation as described for the predicted signal generation circuit (PGC) 11 and outputs an output signal qn, which is supplied to the comparator 18.

The comparator 18 uses the predicted signal generation circuit (PGC) 1 described above.
The output signal qn of 6 is compared with a reference signal generated by the reference signal generation circuit 19 (minimum reference signal is 0000, maximum reference signal is 0111). The control signal Cn output from the comparator 18 is supplied to an adaptive inverse quantizer 17 and a predictive signal generation circuit (PGC) 16.

In the adaptive inverse quantization amount 17, a predetermined adaptive inverse quantization operation is performed by the control signal Cn supplied from the comparator 18, and subtraction [! outputting a signal dn which is approximately equal to the output signal dn from 2;
It is supplied to the adder 2o.

The output signal zn from the adder 20 is supplied to a digital-to-analog converter 22 and also to a delay device 21 that delays the signal by i11 conversion periods, and the output signal of the delay device 21 described above is is added to the output signal of the adaptive inverse quantizer 17 in the adder 20. The output signal zn from the adder 20 described above corresponds to the output signal Xn from the analog-to-digital converter 6 described above, and is also output from the digital-to-analog converter 22 to the output terminal 2. The analog signal provided corresponds to the analog signal supplied to the input terminal 1.

In this way, the adaptive difference P of the present invention shown in FIG.
In the CM method, oversampling technology is used to generate a predicted signal at the next sampling time (predicted signal (slope predicted signal, differential predicted signal) by first-order linear prediction) by first-order linear prediction. Since the period to be calculated is set to be shorter than the actual sampling period, the predicted value and the actual value that conventionally occur when prediction is made based on the magnitude of the slope before the sampling period. A large discrepancy between the two can be avoided as shown in FIG. 3(d).

Next, the adaptive differential PCM of the present invention shown in FIG.
In order to avoid the large discrepancy between the predicted value and the actual value, which conventionally occurs when prediction is made based on the magnitude of the slope before the sampling period,
Using Newton's backward interpolation formula, interpolation is performed so that the period for obtaining a predicted signal is shorter than one actual sampling period, and the n-bit digital signal obtained by sampling and quantizing the analog signal is an adaptive differential P that encodes a related m-bit digital signal and decodes the encoded m-bit digital signal into an n-bit digital signal;
In the CM system, a subtracter 7 that obtains a difference dn between an input signal Xn and a signal Yn-1 corresponding to a signal one sampling period before the input signal, and an output signal dn of the subtracter 7 are inputted. an adaptive quantizer 8 that generates an adaptively quantized signal Pn and outputs it as an output signal of the encoding system; and a subtracter 7 that dequantizes the output signal Pn of the adaptive quantizer 8 and outputs it as an output signal of the encoding system. a first adaptive inverse quantizer 9 that generates a signal dn corresponding to the output signal dn of the first adaptive inverse quantizer 9; Signal Y corresponding to the signal before the sampling period
an addition 1114 for adding n-1, and the adder 14
A delay device 15 delays the output signal of the adaptive quantizer 8 by one sampling period, and a signal for each sampling period of the output signal of the adaptive quantizer 8 is divided into a number of signals according to Newton's backward interpolation formula (where K is 2 or more). ) generates a control signal for adaptively controlling the means 34 for generating a predicted signal forming signal that is interpolated at points ) and the unifier 9 using the above predicted signal forming signal. First adaptive control circuit (1
A coding system constituted by (However, K is 2
a means 35 for generating a predicted signal forming signal that is interpolated at the point (above); and a means 35 for generating a predicted signal by linear prediction using the predicted signal forming signal described above; A second adaptive control circuit (1) generates a control signal for adaptively controlling the quantizer 17.
6, 18° 19) and the signal generated by the aforementioned encoding system is supplied as an input signal, and the operation is controlled by the output signal of the aforementioned second adaptive control circuit (16, 18, 19). a second adaptive inverse quantizer 17,
an adder 2o to which the output signal of the second adaptive inverse quantizer 17 is supplied as an input signal; and the adder 20
This is an adaptive differential PCM system consisting of a decoding system constituted by a means for supplying a signal obtained by delaying the output signal of by one sampling period by a delay device 21 to the adder 20 as the other input signal.

In FIG. 5, 1 is an input terminal for an original signal in the form of an analog signal that is to be transmitted, recorded and reproduced, and the analog signal supplied to the input terminal 1 is input to an analog-to-digital converter (AD). In step 6, sampling and quantization are performed to form a digital signal Xn, which is supplied to a subtracter 7 as a minuend signal Xn.

Since the subtracter 7 is supplied with the signal Yn-1, which corresponds to a signal one sampling period earlier than the minuend signal Xn, as the subtraction signal, the subtracter 7 outputs (X n
-(Y n-1)) = d The difference signal dn represented by dn is output and supplied to the adaptive quantizer 8, and the adaptive quantizer 8 adaptively processes the input difference signal dn. The signal Pn is adaptively quantized to a predetermined number of bits.
The adaptive processing in the adaptive quantizer 8 is performed based on the control signal Cn supplied from the comparator 12 in the adaptive control circuit.

The output signal Pn from the adaptive quantizer 8 described above is supplied to the decoding system via the transmission system or recording medium as an output signal of the encoding system, and is also supplied to the prediction signal generation circuit (PGC).
) 11, and the output signals of the adaptive inverse quantizer 9 and the adaptive quantizer 8 for each sampling period according to Newton's backward interpolation formula (where K is 2
(In the following description with reference to FIGS. 6 and 7, (The above example shows the case where K is 4).

The interpolation circuit 34 described above (also the interpolation circuit 35 provided in the decoding system) processes each sample in the output signal of the adaptive quantizer 8, as shown in block 36 in FIG. The signals Pn-1, Pn-2- for each cycle are expressed as u
= (X-Xo)/h, N If (x)=f
o+ (uΔf−1)+ [(u(u+1)Δ”fo/
21 +...-+r(u(u+1)=(u −n +
1) Δf-n/n! ] -(2) Calculate by substituting the interpolation polynomial based on the step difference shown in the above equation (2), that is, Newton's backward interpolation formula, and convert the output signal Pn-1 into 1#A, for example. Interpolated values P'5n-1, P'5n-2, P'5n-2,
P'5n-3 and P'5n-4 are output and supplied to a predictive signal generation circuit (PGC) 11 having a function as shown by block 37 in FIG.

FIG. 12 shows the predicted signal generating circuit (PGC) 11 shown in FIG. 1 described above (the predicted signal generating circuit (PGC) 16 used in the decoding system in FIG. 5 is also the same) 12 is a block diagram showing a specific example of the configuration of the prediction signal generation circuit (PGC) 1 shown in FIG. 12.
Terminals 3.degree.4.5, etc. in 1 (16) correspond to terminals 3 and 4.5 shown in the predictive signal generation circuit (PGC) 11.16 shown in FIG.

The predicted signal generation circuit (PGC) 11 described above may have any configuration as long as it can generate a predicted value of the signal at the next sampling time.
Prediction signal generation circuit (PGC) 11 illustrated in Fig. 2
For example, No. 2 described in Japanese Patent Publication No. 60-53971
The prediction signal generation circuit 11 shown in FIG. 12 is similar to the prediction signal generation circuit shown in the figure.
, the output signals Pn, P'5n from the interpolation circuit 34
-1, P' 5n-2, P' 5n-3, P'
For each of the 5n-4 signals, the output signal from the absolute value circuit 27 is multiplied by 8 in the coefficient circuit 28 and supplied to the subtracter 30 as a signal to be subtracted, and the delay unit 2 with an I11 period
9 as a subtraction signal to the subtracter 30, the output terminal 5 from the subtracter 30 receives the predicted signal qn at the next sampling time based on first-order linear prediction, that is, the next signal. A signal qn represented by equation (1) is sent out as an output signal and given to the comparator (level comparator) 12.

The prediction signal qn supplied to the comparator 12 is compared with a reference signal generated by the reference signal generation circuit 13 (minimum reference signal is 0000, maximum reference signal is 0111).

The comparator 12 receives the reference signal (the minimum reference signal is 0000,
The maximum reference signal is 0111) and the predicted signal qn supplied to it are compared in signal level, and the signal level of the compared predicted signal qn exceeds the quantization width of the signal currently being transmitted ( Maximum reference signal 0111
(a signal with a level above the minimum reference signal 0000), or a signal with a signal level below the quantization width of the signal currently being transmitted (a signal with a level below the minimum reference signal 0000), or

The compared prediction signal qn is the maximum reference signal 011
A control signal Cn is supplied to the adaptive quantizer 8 according to the comparison result of whether the signal is between 1 and the minimum reference signal 0000.

Therefore, the adaptive quantizer 8 performs, for example, the following adaptive quantization operation, namely, first.

(1) As a result of the comparison in the above comparison $12,
The compared prediction signal qn is the maximum reference signal 011
If the signal is determined to be between 1 and the minimum reference signal 0000, the adaptive quantizer 8 continues to operate under the same operating conditions as before.

(2) The predicted signal qn compared by the comparator 12 described above is a signal with a signal level that exceeds the quantization width of the signal currently being transmitted (a signal with a level that exceeds the maximum reference signal 0111). In response to the comparison result, the adaptive quantizer 8 causes the quantization operation to be performed at the next sampling time to be performed with the quantization width changed by one step.

(3) The predicted signal qn compared by the comparator 12 described above is a signal with a signal level lower than the quantization width of the signal currently being transmitted (a signal with a level lower than the minimum reference signal 0000). The adaptive quantizer 8 corresponds to the comparison result.
Then, the quantization operation performed at the next sampling time is performed with the quantization width changed by one step.

The control signal described above is supplied from the comparator 12 to the adaptive quantizer 8 corresponding to each of the cases (1) to (3) described above.
Additionally, it is supplied via an inverter 10 to an adaptive inverse quantizer 9 .

The output signal Pn supplied from the adaptive quantizer 8 to the decoding system via the transmission system or recording medium as an output signal of the encoding system as described above is a predicted signal (
adaptive quantizer 8 (gradient prediction signal, differential prediction signal)
The signal is adaptively quantized and transmitted, supplied to a decoding system via a recording medium, etc., and also supplied to an adaptive inverse quantizer 9.

The dequantization process performed on the signal Pn supplied to the adaptive dequantizer 9 described above is performed based on the control signal supplied from the comparison @12 in the adaptive control circuit via the inverter 10. , generates a signal dn substantially equal to the output signal dn from the subtracter 2 described above, and supplies it to the adder 14.

The adder 14 adds the predicted signal Y(n-1) from one sampling period (1/fs) ago and the signal dn, outputs the next predicted signal Yn, and stores it in the register. 1
Supply to 5. The register 15 operates as a delay device that delays the signal supplied thereto by the IIJI regularization period (1/fs) and outputs the delayed signal.

Next, the signal Pn supplied to the decoding system in FIG. 5 is supplied to the adaptive inverse quantizer 17 and the interpolation circuit 35. The interpolation circuit 35 described above performs the same operation as the interpolation circuit 34 described above in the encoding system, and its output signal is supplied to the prediction signal generation circuit (PGC) 16. The predicted signal generation circuit (PGC) 16 performs the same operation as described for the predicted signal generation circuit (PGC) 11 and outputs an output signal qn, which is supplied to the comparator 18.

The comparator 18 uses the predicted signal generation circuit (PCyC) described above.
The output signal qn of 16 is compared with the reference signal generated by the reference signal generation circuit 19 (minimum reference signal is oooo, maximum reference signal is 0111). Comparator 18
The control signal Cn outputted from the control signal Cn is supplied to an adaptive inverse quantizer 17 and a predictive signal generation circuit (PGC'') 16.

The adaptive dequantizer 17 performs a predetermined adaptive dequantization operation based on the control signal Cn supplied from the comparator 18, and the output signal dn from the subtracter 2 in the coding system described above is approximately equal to the output signal dn. It outputs an equal signal dn hat and supplies it to the adder 20.

The output signal zn from the adder 20 is supplied to the digital-to-analog converter 22 and also to the delay device 21 which gives the signal a delay of i11 conversion periods, and the output signal from the delay device 21 is added. The output signal of the adaptive inverse quantizer 17 is added to the output signal of the adaptive inverse quantizer 17. The output signal zn from the adder 20 described above corresponds to the output signal Xn from the analog-to-digital converter 6 described above, and is also output from the digital-to-analog converter 22 to the output terminal 2. The analog signal provided corresponds to the analog signal supplied to the input terminal 1.

In this way, the adaptive difference P of the present invention shown in FIG.
In the CM method, Newton's backward interpolation formula is used to generate a predicted signal (predicted signal (slope predicted signal, differential predicted signal) by first-order linear prediction) at the next sampling time by first-order linear prediction. Since the period for obtaining the predicted signal is made shorter than one actual sampling period by the interpolation operation performed in Large discrepancies between predicted and actual values can be avoided as shown in FIG.

Next, the adaptive differential PCM of the present invention shown in FIG.
In order to avoid the large discrepancy between the predicted value and the actual value, which conventionally occurs when prediction is made based on the magnitude of the slope before the sampling period,
Newton's forward interpolation formula is used to interpolate so that the period for obtaining the predicted signal is shorter than one actual sampling period, and the n-bit digital signal obtained by sampling and quantizing the analog signal is an adaptive differential P that encodes a related m-bit digital signal and decodes the encoded m-bit digital signal into an n-bit digital signal;
In the CM system, a subtracter 7 that obtains a difference dn between an input signal Xn and a signal Yn-1 corresponding to a signal one sampling period before the input signal, and an output signal dn of the subtracter 7 are inputted. an adaptive quantizer 8 that generates an adaptively quantized signal Pn and outputs it as an output signal of the encoding system; and a subtracter 7 that dequantizes the output signal Pn of the adaptive quantizer 8 and outputs it as an output signal of the encoding system. a first adaptive inverse quantizer 9 that generates a signal dn corresponding to the output signal dn of the first adaptive inverse quantizer 9; II! l! an adder 14 that adds the signal Yn-1 corresponding to the signal before the main conversion period; a delay device 15 that delays the output signal of the adder 14 by one sampling period; and the adaptive quantizer 8. an interpolation prediction signal generation circuit 39 that generates a prediction signal by interpolating the signal immediately before the next sampling time in the output signal according to Newton's forward interpolation formula; and a control signal for generating the above-mentioned interpolation prediction signal. A coding system constituted by a first adaptive control circuit 12.13, and a signal generated by the coding system is supplied as an input signal, and an interpolated predicted signal generation circuit 41 that generates a predicted signal by interpolating a signal according to Newton's forward interpolation formula; a second adaptive control circuit 18-19 that generates a control signal for the interpolated predicted signal; a second adaptive inverse quantizer 17 to which the signal generated by the quantizer is supplied as an input signal and whose operation is controlled by the output signal of the second adaptive control circuit; An adder 2, O to which the output signal of the adaptive inverse quantizer 17 is supplied as an input signal, and a signal obtained by delaying the output signal of the adder 20 by one sampling period in a delay device 21, This is an adaptive differential PCM system consisting of a decoding system constituted by means for supplying the other input signal to the adder 20.

In FIG. 8, 1 is an input terminal for an original signal in the form of an analog signal that is to be transmitted, recorded and reproduced, and the analog signal supplied to the input terminal 1 is sent to an analog-to-digital converter (AD) 6. is sampled and quantized into a digital signal Xn, which is sent to the subtractor 7 as a minuend signal Xn
Supplied as.

Since the subtracter 7 is supplied with the signal Yn-1 corresponding to the signal one sampling period before the minuend signal Xn as the subtraction signal, the output of the subtracter 7 is (X n
-(Yn-1)) = difference signal dn denoted by dn
is supplied to the adaptive quantizer 8 via the limiter 38. The adaptive quantizer 8 adaptively processes the input difference signal dn and outputs it as a signal Pn that has been adaptively quantized to a predetermined number of bits. This is performed based on the control signal On supplied from the comparator 12 in the control circuit.

The output signal Pn from the adaptive quantizer 8 described above is:

It is supplied to the decoding system as an output signal of the encoding system via a transmission system or a recording medium, and is also supplied to an adaptive inverse quantizer 9,
It is also supplied to an interpolated predicted signal generation circuit 39 that generates a predicted signal by interpolating the signal immediately before the next sampling time in the output signal of the adaptive quantizer 8 described above according to Newton's forward interpolation formula. The interpolation prediction signal generation circuit 39 (also the interpolation prediction signal generation circuit 41 provided in the decoding system) described above includes an adaptive quantizer 8 as shown in block 43 in FIG. The difference is calculated from the signals Pn-1, Pn-2, . . . for each sampling period in the output signal of The previous interpolated value, u
= (X-Xo)/h as N I (X)=f
o+uΔfo+ [(u(u-1)Δ”fo/21]+
−−+[:(u(u−1)”(u −n +1)Δfo
/nl] - (3) Calculate the interpolation polynomial based on the step difference shown in equation (3) above, that is, Newton's forward interpolation formula, to calculate the interpolated value P'5n+ immediately before the next sampling time.
4 as a prediction signal qn as shown by block 45 in FIG.

The predicted signal qn supplied to the comparison 1812 is compared with a reference signal generated by the reference signal generation circuit 13 (the minimum reference signal is oooo, the maximum reference signal is 0111) in the comparison @12.

The comparator 12 receives the reference signal (the minimum reference signal is 0000,
The signal level of the maximum reference signal (0111) is compared with the predicted signal qn supplied thereto, and a signal ( Maximum reference signal 0111
or the signal level is below the quantization width of the signal currently being transmitted (a signal at a level below the minimum reference signal oooo), or the compared predicted signal qn is Maximum reference signal 0
A control signal Cn is supplied to the adaptive quantizer 8 according to the comparison result of whether the signal is between 111 and the minimum reference signal 0000.

Therefore, the adaptive quantizer 8 performs, for example, the following adaptive quantization operation. First, (1) As a result of the comparison in the comparator 12, the compared prediction signal qn is determined to be the maximum reference signal. If the signal is determined to be between 0111 and the minimum reference signal 0000, the adaptive quantizer 8 continues to operate under the same operating conditions as before.

(2) The predicted signal qn compared by the comparator 12 described above is a signal whose signal level exceeds the quantization width of the signal currently being transmitted (a signal whose level exceeds the maximum reference signal 0111). In response to the comparison result, the adaptive quantizer 8 causes the quantization operation to be performed at the next sampling time to be performed with the quantization width changed by one step.

(3) The predicted signal qn compared by the comparator 12 described above is a signal with a signal level lower than the quantization width of the signal currently being transmitted (a signal with a level lower than the minimum reference signal oooo). The adaptive quantizer 8 corresponds to the comparison result.
Then, the quantization operation performed at the next sampling time is performed with the quantization width changed by one step.

In addition to being supplied from the comparator 12 to the adaptive quantizer 8, the control signals described above are supplied to the adaptive inverse quantizer 9 via the inverter 1o corresponding to each of the cases (1) to (3) described above. be done.

The output signal Pn supplied from the adaptive quantizer 8 to the decoding system via the transmission system or recording medium as an output signal of the encoding system as described above is a predicted signal (
adaptive quantizer 8 (gradient prediction signal, differential prediction signal)
The signal is adaptively quantized and transmitted, supplied to a decoding system via a recording medium, etc., and also supplied to an adaptive inverse quantizer 9.

The dequantization process performed on the signal Pn supplied to the adaptive dequantizer 9 described above is performed based on the control signal supplied from the comparator 12 in the adaptive control circuit via the inverter 10. , generates a signal dn substantially equal to the output signal dn from the subtracter 2 described above, and supplies it to the adder 14.

In the adder 14 described above, II! 11 cycle (1/fs
) The previous predicted signal Y(n-1) and the signal dn are added to output the next predicted signal Yn, which is supplied to the register 15 via the limiter 40. The register 15 receives the signal supplied thereto for one sampling period (1/
It operates as a delay device that delays the output by an amount (fs).

Next, the signal Pn supplied to the decoding system in FIG. 8 is supplied to the adaptive inverse quantizer 17 and the interpolation prediction signal generation circuit 41. The interpolation prediction signal generation circuit 41 described above performs the same operation as the interpolation prediction signal generation circuit 39 described above in the coding system, and the output signal qn is output from the comparator 1.
8.

The comparator 18 uses the predicted signal generation circuit (PGC) 1 described above.
The output signal qn of 6 is compared with a reference signal generated by the reference signal generation circuit 19 (minimum reference signal is 0000, maximum reference signal is 0111). The control signal Cn output from the comparator 18 is supplied to the adaptive inverse quantizer 17.

The adaptive dequantizer 17 performs a predetermined adaptive dequantization operation based on the control signal Cn supplied from the comparator 18, and the output signal dn from the subtracter 2 in the coding system described above is approximately equal to the output signal dn. It outputs an equal signal dn hat and supplies it to the adder 20.

The output signal zn from the adder 20 is supplied to the digital-to-analog converter 22 via the limiter 42, and
The output signal of the delay device 21 is supplied to the delay device 21 which gives a delay of one sampling period to the signal, and the output signal of the delay device 21 is added to the sum Il!
It is added to the output signal of the adaptive inverse quantization 8117 at I20. Then, the output signal z from the adder 20 described above
n corresponds to the output signal Xn from the analog-to-digital converter 6 described above, and the analog signal output from the digital-to-analog converter 22 to the output terminal 2 is supplied to the input terminal 1. It is compatible with analog signals.

(Effects of the Invention) The adaptive differential PCM method of the present invention causes a large discrepancy between the predicted value and the actual value, which conventionally occurs when prediction is performed based on the magnitude of the slope before the sampling period. This is an adaptive differential PCM method in which the period for obtaining a predicted signal is made shorter than one actual sampling period to prevent this from occurring.

An n-bit digital signal obtained by sampling and quantizing an analog signal is encoded into an m-bit digital signal with a relationship n>m, and the encoded m-bit digital signal is converted into an n-bit digital signal. Adaptive differential PC for decoding
M method, which includes a subtracter that obtains the difference between the input signal and a signal corresponding to a signal one sampling period before the input signal, and an output signal of the subtracter that is input to generate an adaptively quantized signal. and an adaptive quantizer that outputs it as an output signal of the encoding system, and a second device that dequantizes the output signal of the adaptive quantizer to generate a signal corresponding to the output signal of the subtracter. a first adaptive inverse quantizer; an adder for adding the output signal of the first adaptive inverse quantizer and a signal corresponding to the signal of the input signal i11 cycles before;
A delay device delays the output signal of the adder by one sampling period, and a sampling frequency conversion of twice (however, 2 or more) is performed on the output signal of the adaptive quantizer. A first adaptive control circuit that generates a signal for forming a prediction signal of a state, and a first adaptive control circuit that generates a control signal for adaptively controlling the primary signal using the signal for forming a prediction signal of the state. The signal generated by the coding system and the coding system described above is supplied as an input signal, and the input signal is subjected to sampling frequency conversion of twice (however, 2 or more). means for obtaining a signal for forming a predicted signal; a second adaptive control circuit for generating a primary line control signal using the signal for forming a predicted signal; and a signal generated by the coding system as an input signal; a second adaptive inverse quantizer whose operation is controlled by the output signal of the second adaptive control circuit; Adaptive differential PCM method consisting of a decoding system constituted by an adder that adds the output signal of the adaptive inverse quantizer of No. 2 to a signal delayed by one sampling period using a delay device, and an analog signal. The n-bit digital signal obtained by sampling and quantizing the n>m
This adaptive differential PCM method encodes the encoded m-bit digital signal into an n-bit digital signal with a relationship between an input signal and one of the input signals. A subtracter obtains the difference between the signal corresponding to the signal before the sampling period, and the output signal of the subtracter is inputted to generate an adaptively quantized signal, which is output as the output signal of the encoding system. an adaptive quantizer; and a first device that dequantizes the output signal of the adaptive quantizer to generate a signal corresponding to the output signal of the subtracter.
with an adaptive inverse quantizer.

an adder that adds the output signal of the first adaptive inverse quantizer and a signal corresponding to a signal one sampling period before the input signal; A delay device that delays the output signal of the above-mentioned adaptive quantizer for each sampling period is interpolated at a number of points (however, 2 or more) according to Newton's backward interpolation formula. A code constituted by means for generating a signal for forming a predicted signal, and a first adaptive control circuit for generating a control signal for controlling in a first-order linear predictive manner using the signal for forming a predicted signal. The signal generated by the coding system and the coding system described above is supplied as an input signal, and the signal for each sampling period in the input signal is divided into several (however, two or more) signals for each sampling period in the input signal according to Newton's backward interpolation formula. means for generating a predicted signal forming signal interpolated at points; a second adaptive control circuit generating a primary line control signal using the predicted signal forming signal; and 5 the above-mentioned encoding. a second adaptive inverse quantizer to which the signal generated by the system is supplied as an input signal and whose operation is controlled by the output signal of the second adaptive control circuit; An adaptive decoding system consisting of an adder that adds the output signal of the quantizer and a signal obtained by delaying the output signal of the second adaptive inverse quantizer by one sampling period using a delay device. An n-bit digital signal obtained by sampling and quantizing an analog signal using the differential PCM method is encoded into an m-bit digital signal with a relationship n>m, and the encoded m-bit digital signal is Adaptive differential PC that decodes into n-bit digital signal
M method, the input signal and the input signal I! ! A subtracter that obtains the difference between the signal and the signal corresponding to the unification period, and the output signal of the subtracter is inputted to generate an adaptively quantized signal, which is output as an output signal of the encoding system. an adaptive quantizer; a first adaptive inverse quantizer that dequantizes the output signal of the adaptive quantizer to generate a signal peak corresponding to the output signal of the subtracter; an adder that adds the output signal of the adaptive inverse quantizer No. 1 and a signal corresponding to a signal one sampling period before the input signal; and an adder that delays the output signal of the adder by one sampling period. an interpolation prediction signal generation circuit that generates a prediction signal by interpolating the signal immediately before the next sampling time in the output signal of the adaptive quantizer according to Newton's forward interpolation formula; a first adaptive control circuit that generates a control signal for controlling a second adaptive control circuit that generates a control signal for adaptively controlling an interpolation predictor that generates a predicted signal by interpolating the signal immediately before the sampling time according to Newton's forward interpolation formula; a second adaptive inverse quantizer to which the signal generated by the coding system is supplied as an input signal and whose operation is controlled by the output signal of the second adaptive control circuit; An adder to which the output signal of the adaptive inverse quantizer is supplied as an input signal on one side, and a signal obtained by delaying the output signal of the adder by 1111 periods in a delay device to the adder as the other input signal. and a decoding system constituted by supply means, and the adaptive differential PCM method of the present invention is capable of reducing the discrepancy between the predicted signal and the actual signal. According to the method of the present invention, the problems of the conventional method described above can be satisfactorily solved.

[Brief explanation of the drawing]

1 and 5 and FIG. 8 are block diagrams of six different embodiments of the adaptive differential PCM method of the present invention, and FIGS. 2 to 4 are diagrams for explaining the embodiment shown in FIG. 6th
FIG. 7 is a diagram for explaining the embodiment shown in FIG.
9 and 10 are diagrams for explaining the embodiment shown in FIG. 8, FIG. 11 is a block diagram of a conventional example of the adaptive differential PCM method, and FIG. 12 is a configuration example of a predicted signal generation circuit. FIG. 13 is an explanatory diagram for explaining the problem. 1. Input terminal for an original signal in the form of an analog signal used for recording and reproduction, etc.; 2. Output terminal; 3-5・
...Input terminal, 6...Analog-to-digital converter (A
D), 7, 30... subtractor, 8... adaptive quantizer,
9.17...Adaptive inverse quantizer, 10, 24, 26...
・Inverter, 11.16...Prediction signal generation circuit (P
GC), 12...13, 19... Reference signal generation circuit, 14.20... Adder, 15, 21, 29... Delay device, 18... Comparator, 22... Digital・Analog converter (DA), 27... Absolute value circuit, 28...
・Coefficient circuit 34.35...Interpolation circuit, 38, 40.4
2...Limitsuta. 39°41...Interpolation prediction signal generation circuit.

Claims (1)

[Claims] 1. An n-bit digital signal obtained by sampling and quantizing an analog signal is encoded into an m-bit digital signal with a relationship n>m, and the encoded m-bit digital signal is An adaptive differential PCM method for decoding a signal into an n-bit digital signal, comprising: a subtracter that obtains a difference between an input signal and a signal corresponding to a signal one sampling period before the input signal; and the subtracter described above. an adaptive quantizer that inputs the output signal of the adaptive quantizer, generates an adaptively quantized signal, and outputs it as an output signal of the encoding system; and an adaptive quantizer that dequantizes the output signal of the adaptive quantizer and performs the subtraction described above. a first adaptive inverse quantizer for generating a signal corresponding to the output signal of the quantizer;
an adder for adding the output signal of the adaptive inverse quantizer and a signal corresponding to a signal one sampling period before the input signal; and a delay for delaying the output signal of the adder by one sampling period. means for generating a signal for forming a predicted signal in a state where K times (however, K is 2 or more) sampling frequency conversion has been performed on the output signal of the adaptive quantizer; A first apparatus for generating a control signal for adaptively controlling the adaptive quantizer and the first adaptive inverse quantizer, including means for generating a predicted signal by first-order linear prediction using the signal forming signal. A coding system constituted by an adaptive control circuit No. 1 and a signal generated by the above coding system are supplied as input signals, and the signal generated by the above-mentioned coding system is supplied as an input signal. a means for obtaining a signal for forming a predicted signal in a state where sampling frequency conversion has been performed; and a means for generating a predicted signal by first-order linear prediction using the signal for forming a predicted signal; a second adaptation control circuit that generates a control signal for adaptively controlling the encoder; and a second adaptation control circuit that is supplied with the signal generated by the coding system as an input signal; a second circuit whose operation is controlled by the output signal of the control circuit;
an adaptive inverse quantizer, an adder to which the output signal of the second adaptive inverse quantizer is supplied as an input signal, and a delay device to delay the output signal of the adder by one sampling period. Adaptive differential PCM method 2 consists of a decoding system configured with a means for supplying the input signal to the adder as the other input signal, and an n-bit digital signal obtained by sampling and quantizing an analog signal. An adaptive differential PCM method that encodes an m-bit digital signal with a relationship n>m, and decodes the encoded m-bit digital signal into an n-bit digital signal. A subtracter obtains the difference between the signal corresponding to the signal one sampling period before, and the output signal of the subtracter is inputted to generate an adaptively quantized signal, which is used as the output signal of the encoding system. a first adaptive inverse quantizer that inversely quantizes the output signal of the above-mentioned adaptive quantizer to generate a signal corresponding to the output signal of the above-mentioned subtracter; The first
an adder for adding the output signal of the adaptive inverse quantizer and a signal corresponding to a signal one sampling period before the input signal; and a delay for delaying the output signal of the adder by one sampling period. For forming a predicted signal in a state where the signal for each sampling period in the output signal of the adaptive quantizer described above is interpolated at K points (however, K is 2 or more) according to Newton's backward interpolation formula. Adaptive quantizer and first adaptive inverse quantizer, including means for generating a signal and means for generating a predicted signal by first-order linear prediction using the predicted signal forming signal. A coding system constituted by a first adaptive control circuit that generates a control signal for control, and a signal generated by the coding system is supplied as an input signal, and each sampling in the input signal is A means for generating a predicted signal forming signal in which a signal for each cycle is interpolated at K points (K is 2 or more) according to Newton's backward interpolation formula, and using the above predicted signal forming signal. a second adaptive control circuit for generating a control signal for adaptively controlling the second adaptive inverse quantizer; a second adaptive inverse quantizer to which the signal generated by the second adaptive inverse quantizer is supplied as an input signal and whose operation is controlled by the output signal of the second adaptive control circuit; an adder to which the output signal of the converter is supplied as one input signal; and means for supplying a signal obtained by delaying the output signal of the adder by one sampling period to the adder as the other input signal. Adaptive differential PCM method 3, which consists of a decoding system configured by The adaptive differential PCM method decodes the encoded m-bit digital signal into an n-bit digital signal, the difference between the input signal and a signal corresponding to a signal one sampling period before the input signal. , an adaptive quantizer that inputs the output signal of the subtracter, generates an adaptively quantized signal, and outputs it as an output signal of a coding system; a first adaptive inverse quantizer that inversely quantizes an output signal to generate a signal corresponding to the output signal of the subtracter;
an adder for adding the output signal of the adaptive inverse quantizer and a signal corresponding to a signal one sampling period before the input signal; and a delay for delaying the output signal of the adder by one sampling period. an interpolation prediction signal generation circuit that generates a prediction signal by interpolating the signal immediately before the next sampling time in the output signal of the adaptive quantizer according to Newton's forward interpolation formula; and the interpolation prediction signal generation circuit that generates a prediction signal. and a first adaptive control circuit that generates a control signal for adaptively controlling the adaptive quantization amount and the first adaptive inverse quantizer based on the output signal of the circuit. system and
An interpolation prediction signal generation circuit that is supplied with the signal generated by the above-mentioned coding system as an input signal and generates a prediction signal by interpolating the signal immediately before the next sampling time in the input signal according to Newton's forward interpolation formula. and a second adaptive control circuit that generates a control signal for adaptively controlling the second adaptive inverse quantizer based on the output signal of the interpolation prediction signal generation circuit; a second adaptive inverse quantizer to which the signal generated by the second adaptive control circuit is supplied as an input signal and whose operation is controlled by the output signal of the second adaptive control circuit; An adder to which the output signal of the inverse quantizer is supplied as one input signal, and a signal obtained by delaying the output signal of the adder by one sampling period by a delay device to the other input signal. an adaptive differential PCM method comprising a decoding system configured by means for