JPS58164007A - Encoding storage and regenerating device of signal - Google Patents

Abstract

PURPOSE:To easily obtain a regenerating signal having a good quality, which is brought to tangent approximation, by integrating a ratio of an amplitude value in digital data and a value showing a sampling period, and executing interpolation of a signal. CONSTITUTION:For instance, a voice is converted to an electric signal by a microphone, and thereafter, is sampled by a sampling period of an A/D converter ADC, and its amplitude value is quantized. It is controlled by a controlling circuit CCT, and is stored in the second memory M2 as data consisting of a group of the amplitude value and information related to a zero point interval. In case of regeneration, this data group is read out and is provided to a D/A converter DAC, the digital amplitude value and the information related to the zero point interval are converted to an analog value, and a signal is applied to an interpolating circuit CP. The interpolating circuit CP consists of a divider and an integrator, and by the divider, a ratio of the amplitude value of the analog value and the information related to the zero point interval is calculated, and its output is integrated by the integrator, and is made approximate to its original AC signal by tangent approximation.

Description

[Detailed Description of the Invention] When an audio signal is encoded and transmitted or recorded and reproduced as a digital signal, conventional methods for reducing the amount of data as much as possible include logarithmically compressing the signal amplitude or calculating the difference. It is well known that various methods such as modulation or delta modulation have been adopted, but the conventional methods seek to reduce the amount of data in the amplitude direction. There was a problem in that the quality of the reproduced signal deteriorated due to quantization distortion.

When the amount of data decreases significantly, the applicant company does not calculate the decrease in the number of hits in the amplitude direction, but rather calculates it in the time axis direction. An encoding device with a variable sampling period that makes it possible to achieve a reduction in the original signal. An encoding device with variable sampling periods that can obtain digital data consisting of a sample value indicating the amplitude of a signal and a take indicating the sampling period for which the sample value is obtained. I have been making a proposal regarding this.

The present invention provides a sample value indicating the amplitude value of an IC, an original signal, and a sampling period for obtaining such a zero value, such as an IC such as a digital data encoded by the already proposed encoding device with a variable sampling period. If you decode the digital data consisting of a pair of take and , which indicate The purpose of this invention is to provide a signal encoding, storing and reproducing apparatus that can obtain the following information.The specific details of the signal encoding, storing and reproducing apparatus of the present invention will be described below with reference to the accompanying drawings. Explain the contents.

1 and 7 are block diagrams of exemplary embodiments of the signal encoding, storage and generation apparatus of the present invention, respectively;
In Figures 1 and 7, MIC is a microphone, L
PFr, LPFp is a low-pass filter, ADC is a Japanese food exchanger,
CCT is a control circuit including, for example, a microphone repeater, Ml is a first storage device (first memory, buffer memory), M2 is a second storage device (second memory), and DAC is The gap converter, CP is an interpolation circuit, SP is a speaker, and OP is an operation unit.
P has a record button Br.

A play button Bp and a stop button Bs are provided. In addition, the seventh
The CG line in the figure is a generator of ρ pulses.

Both of the devices shown in FIGS. 1 and 7 have a sampling period that is controlled by a control circuit CCT in accordance with the state of change on the time axis of the input analog signal. Perform variable encoding to create a digital data consisting of a sample value indicating the amplitude value of the original signal and a take indicating the sampling period of T to obtain the sample value, and then When internally stored, the second memory M2 is stored under the control of the control circuit CCT.
Then, the ratio between the amplitude value in the digital theta and the value representing the sampling period is determined by division, and the ratio between the amplitude value of the previous period and the value representing the sampling period is integrated to form a broken line. Although the apparatus is configured to obtain an approximated spontaneous signal, there is a difference in the specific content of encoding between the apparatus shown in the first figure and the apparatus shown in the seventh figure.
The apparatus j shown in the first figure will be explained.

Figure 2 shows how the sampling period changes in response to the state of change of the signal on the time axis when the signal is encoded by the device shown in Figure 1. This is a waveform diagram for illustrating and explaining how the signal is transmitted. In this second figure, Sa is an AC signal to be encoded, and the O-0 line in the diagram indicates the AC axis. It shows.

In FIG. 2, L1+V2+j3... is the time position where the AC signal Sa successively cuts the AC axis line 0-0, that is, the time position of successive zep points in the AC signal Sa, and , TI21 T231 T34, etc., are successively determined by the time position 1 . of the 0 point in the above-mentioned AC signal Sa. ．． 12... indicates the time interval between adjacent items (hereinafter sometimes referred to as zero point interval).

Now, the interval T, 2 between successive zero points in the AC signal 5avc
゜T231 T34... indicates a constant time value corresponding to 1/2 of the period of the AC signal when the AC signal is a sine wave signal with a constant frequency, but in the case of an audio signal, TI21 and T23 in Fig. 2 correspond to the signal contents.
1 T34... The time value changes.

By the way, conventionally, when sampling and quantizing an analog signal to generate a digital signal, a sampling pulse Ps having a constant time interval Tp shown on the AC axis line O-O in FIG. As is well known, according to the conventional method, even when taking the AC signal Sa in FIG. 2 as an example, as is clear,
The number of sampled values sampled at successive zero point intervals increases as the zero point interval becomes longer;
In the illustrated apparatus, the AC signal is encoded as a state in which approximately a constant number of samples are extracted at each zero point interval, regardless of the length of the interval between successive zero points in the AC signal. In this way, the amount of data can be reduced.

In other words, the time obtained by equally dividing each successive zero point interval in the AC signal 5alC by a predetermined number α (however, α is an integer greater than or equal to 2 and is a predetermined number) An alternating current signal tental encoding device capable of reducing the amount of data can be constructed such that the value is a sampling period for obtaining a sample value from a signal portion corresponding to each zero point interval. In order to divide α equally between each zero point, we need to divide α by 2.
It goes without saying that whether to use 3, 3, or 4 or more should be determined appropriately, taking into consideration the degree of fidelity required for the reproduced signal, the cost of the encoding device, etc.

In the device shown in Figure 1, the tental encoding device for AC signals detects successive points on the time axis in the AC signal to be subjected to tental encoding, and also detects points that are adjacent to each other on the time axis. Measure the time interval of one point, that is, the zero point interval, and divide the measured zero point interval into α equal parts (however, α
is a predetermined integer of 2 or more) as the sampling period, and from the signal part corresponding to the zero point interval (
Since α-1) sample values can be obtained, most of the operations can be performed by analog signal processing as long as the device configuration is such that each of the above operations can be performed satisfactorily. Alternatively, the device may be configured so that substantially all operations are performed by tental signal processing, but it is better to configure the device so that substantially all operations are performed by digital signal processing. The device configuration is shown as an example configuration in which all operations are performed by digital signal processing.

Next, the configuration and operation of the apparatus shown in the first figure will be explained.

In Figure 1, Microbon MIC converts sound waves into electrical signals (
Convert the audio signal) to a low-pass filter LPFrVc. In the example shown in the first figure, the signal source is a Microbon MIC, but it goes without saying that the signal source may be another audio signal generator.

The aforementioned low-pass filter LPFr is assumed to have a cut-off frequency of 4 KHz in the following description of the embodiment.

\ The audio signal, which has a frequency band of 4KT (2 or less) by the low-pass filter LPFr, is converted into a digital signal with a predetermined number of bits (8 bits in the following explanation) by the liquid converter ADC.The following explanation , the layer liquid exchanger ADC has a sampling frequency of 8KH.
The D converter ADC samples the input audio signal at a sampling period of vc +/5ooo seconds, quantizes the amplitude value of each sample, and converts it into 8 bits. into a digital signal.

The digital signal output from the layer liquid changer ADC is sent to a control circuit CCT that includes a microconbeater.
A digital signal with a reduced amount of data is obtained by performing predetermined signal processing under the control of the control circuit CC.
T, a first storage device M (first memory M1), a second storage device M2 (second memo! IM2), etc. In the following description, the first memory M,
is said to be a memory of 256 bytes, and the second memory M is said to be a memory of 64 bytes.

Next, with reference to the flowchart shown in FIG. 3, a description will be given of the encoding of signals that is sequentially performed by operating the recording button Br of the operating unit in the apparatus shown in FIG.

When the program starts (F start in Figure 3) by operating the record button Br on the operation unit OP, in step (1), the control circuit CCT for measuring the zero point interval is activated. The 8-hit counter is reset. 8 one after another from the converter M in step (2)
The digital signal of Hino 1 (1 byte) is stored in the 256-height first memory M1, and the first memory IJM
, IC Each time the 1-byte digital signal mentioned above is stored, the 8-hit counter for the zero point interval measurement is incremented by 1.

In step (3), the amplitude value of the original AC signal indicated by the stored 1-byte digital signal is stored in the first memory M.
It is determined whether or not, and if it is determined that it is not 0, the process returns to step (2), and if it is determined that it is zero, the process proceeds to step (4).

In step (4), the zero point interval indicated by the count value of the 8-hint counter for measuring the zero point interval is divided into α equal parts (however, α is a predetermined integer of 2 or more). Obtain numerical values that correspond to (α - 1) dividing points (for example, if the count value indicating the zero point interval is A, then α is 2).
．． When α is 3° and α is 4, the numerical value obtained in step (4) is, When α is the interval A, the numerical value obtained in step (4) is Δ, Hiro... As shown
1) α α items). In the following explanation, for the sake of simplicity, the case where the base is 2 is used as an example.

In step-j (5), read the take of the amplitude value of the first memory M1 at the address position of the first memory M, which corresponds to the count value 172 of the 8-bit counter for the zero point interval profit. , the take of the amplitude value and the value of 1/2 of the count value of the 8-bit counter for measuring the zero point interval described above are combined and stored in the second memory M2 of byte 64 (as described above). The take of the amplitude value read from the first memory M1 and the set value of the 8-hit counter for measuring the zero point interval, which indicates the time value of the zero point interval, are combined into 64 bytes. The count value of the above-mentioned counter to be used in conjunction with the take of the amplitude value read from the second memory M2 is as follows:
Since there is an advantage that the number of required bits is small, in the following explanation, the numerical value to be used in combination with the take of the amplitude value read from the first memory M is 8 hints for measuring the zero point interval. (The case where the value is 1/2 of the count value of the counter is described).

In step (6), the 8-hit color/reset for the open side of the zero point interval is reset to set the group value to 0, and the second
Memo! /M2Vc The 15-hit color provided for counting the number of data sets stored in the second memory M2Vc is counted up by one each time a new data set is stored in the second memory M2Vc. Then proceed to step (7). 1 step (
In 7), it is determined whether the second memory M2 has reached a full count, or whether the stop button B8 on the operation part oP has been pressed, and the state where the second memory M2 has reached a full count, or When it is detected that the stop button Bs is pressed, the program ends; otherwise, the process returns to step (2) vc.

In the above explanation, in order to simplify the explanation, it was described that the zero point of the signal is determined simply based on whether or not the amplitude value stored in the first memory M and IC is 0.
The amplitude values sequentially stored in the memory M of the D converter A
Since it is a discrete value obtained every sampling period in DC, the zero point of the AC signal is not always sampled, so it is stored sequentially in the first memory M1. If the polarity of the information indicating the amplitude value reversed is determined to indicate the zero point of the alternating current, a determining means should be used.

Furthermore, if it is determined in step (3) that the amplitude value of the original AC signal is zero, step (4)
If the number of (α-1) numbers obtained from the count value of the 8-bit counter for measuring the zero point interval is 2 or more, that is, α is 3 or more, in step (5) VC The take of the amplitude values to be read from the first memory M, may equally divide the ρ interval represented by the count value of 8 colors/each for the measurement of the zero point interval into α pieces. Each address position of the first memory M1 corresponding to each numerical value corresponding to (α-1) division points is sequentially read out, and The take of those amplitude values read out as
For each of these, a set of zero point interval and associated numerical value is made and stored in the second memory M2.

As is clear from the above explanation, the encoding by the apparatus shown in the first diagram is performed by setting the zero point interval in the AC signal to a predetermined number α (however, α is an integer greater than or equal to 2, and is not a predetermined number). The time values obtained when equally divided by (α−1) from the signal part corresponding to each zero point interval
) sample values, thereby reducing the amount of data. By adding information related to the zero point interval from which the sample value was obtained, the sample value and the zero point interval information are made into a set, which facilitates decoding.
It is designed so that you can get 1.

Next, decoding will be explained using the block diagram shown in the first figure and the two-chart shown in the fourth figure. First, when the program shown in the flowchart of FIG. 4 is started by operating the playback button Bp of the operation section OP in the device, in step (IP), a device is installed to count the number of data sets to be stored in the second memory M2. The 15-bit counter is reset and the process proceeds to step (2P). The 15-bit counter counts up by one each time one data set is read from the second memory M2.

In step (2P), information related to the amplitude value and the zero point interval (including the count value of the 8-hit counter for measuring the zero point interval) is divided into equal parts in the order stored in the second memory M2. The data set is read out from the set of the obtained value (in the example described above, the value is 1/2 of the count value of the 8-hit counter),
In 3P), the DA converter DAC converts the above-mentioned amplitude value into the amplitude value t1 of the γsρg quantity, and converts the information related to the above-mentioned zero point interval to the analog quantity of electrical quantity τl, and converts them into Interpolator CP [Let's give it.

FIG. 5 is a block diagram showing an example of the configuration of interpolation circuit 0 and the configuration of related parts. In FIG. 5, DIV is a divider, INT is an integrator, and DA converter DAC
, D/A 1 is a DAf converter which converts one amplitude value read from the second memory M2 and outputs an analog amplitude value 11, and D/A 2 is read from the second memory M2. This is a DA converter that converts information related to the ρ point interval into an analog electrical quantity τ1 and outputs it.

The interpolator σ has a divider D for its divider DIVK.
/AI output signal 11 and DA converter D/A2 output signal τl are given,
, 1 Xi-1 (1) The signal Xi is output by performing calculations as shown in equation (1).

The output signal X1 from the divider DIV is integrated by the integrator INT and output as a signal Yi.

Yi = Xi dt... (2) ゛Interpolation circuit CPK In the interpolation of the yaw signal, the slope θ of the interpolation is 2 in Fig. 6.
It is made to be equal to 1/τiK. The interpolated signal can be made to have a waveform that approximates the original AC signal by polygonal line approximation.

In order to provide time for signal interpolation by the interpolation operation explained with reference to FIGS. 5 and 6, that is, time divided into zero point intervals, step (4P) is replaced by step (2P). , step (3P), step (4P)
) The program play is created so that the elapsed time is equal to the time obtained by dividing the zero point interval time described above into six equal parts. The program play in step (4P) is shown as follows.

((The count value of the 8-bit counter for measuring the zero point interval corresponding to one of the zero point interval divided into α equal parts) ÷
(AD converter ADCIc sampling period) - (play time) )-11+step (2P) and step (3P)
time) + (fixed time during step (4P))... (3) In step (5P), the amplitude value is sign-inverted and output, and in step (6P), the previously mentioned step (
4P), now step (5P).

A program play is created in such a manner that the elapsed time of step (6P) and step (7P) corresponds to the above-mentioned equation (3).

In step (7P), it is checked whether the 15-hit counter that counts the number of data sets read from the second memory M2 has reached a full count, or whether the stop button BS has not been pressed. If the counter is at full count or if the stop button Bs is operated, the process ends; if not, the process returns to step (2P) and reads the next data set.

For example, if the average zero point interval is a count value of 4, the average zero point interval is 0.5 ms, so one byte is allocated to each of the amplitude value and information related to the zero point interval. If you try to memorize it by
If a 64-byte memory M2 is used as the second memory M2, approximately 16 seconds worth of audio signals will be stored and reproduced. In the apparatus shown in the first figure, the reproduced signal subjected to the polygonal approximation by the interpolation circuit CP is passed through the low-pass filter L.
A speaker 5PiC is provided through the PFp and reproduced sound is obtained.

Next, the apparatus shown in FIG. 7 will be explained. In the apparatus shown in FIG. 7, the signal to be encoded is a signal of a certain time length (one frame signal), and one
Sampling period Tc (sampling interval Tc
).

FIG. 8 is a waveform side diagram of the signal Sa before encoding, the 0-0 line in FIG. 8 is an AC axis line shown for reference, and in FIG. This is the time length of a signal portion (time length of one frame signal) divided into fixed time lengths on the time axis.

In the signal Sa[, the signal to each frame having a predetermined time length Tf crosses the AC axis O-0 line multiple times within the time length Tf, that is,
There are multiple zero points within the time length Tf, but the number of zero points in each one frame signal depends on the frequency components of the signal in each one frame. For example, the signal Sa shown in FIG.
To explain 1, there are 8 nine points in one frame signal from time t1 to t2, and from time t2 to t3.
There are four zero points in the one-frame signal up to this point, and below, for the successive l-frame signals on the time axis,
The numbers of zero points are 6, 3, and 4.

In the apparatus shown in FIG. 7, as described above, the signal portion of the signal Ra having a predetermined constant time length Tf, that is,
For each one-frame signal, the sample value sequence for that one-frame signal is determined by a sampling period Tc such that the time length is equally divided by a number related to the number of zero points existing in the l-frame signal. The aim is to achieve a reduction in the amount of data by performing encoding such that .

That is, as mentioned above, the number of zero points of the signal of one frame successively on the time axis like the signal Sa shown in FIG. 8 is 8.
, 4, 6, 3, and 4, for example, for a signal of one frame with 8 zero points,
In order to obtain a sample value sequence from the signal of one frame with a sampling period such that the time length Tf is equally divided by (8XK), for example, the signal of one frame with four zero points is , the number of zero points is 6 or For each one-frame signal, such as three, a sequence of sample values from each one-frame signal is obtained at a sampling period such that the time length Tf is equally divided by (6XK) or (3XK). Generally, 1
When the number of zero points in a frame signal is two, the time length Tf of that one frame signal is (ZXK
) A sequence of sample values can be obtained with a sampling period that is divided into equal parts, and unless the above-mentioned encoding method is used, recording internal operation with a reduced amount of data will not be possible. can be easily realized.

The data obtained by the encoding means as described above, that is, the sample value sequence from each one frame signal having a predetermined time length Tf has a specific relationship with the number of zero points 2 in the one frame signal. The number (ZXK)Ic with
Therefore, the data obtained by sampling with the sampling period obtained by dividing the time length Tf into equal parts is: that data, the sampling period increase Tc, the number N of sampled values in the l-frame signal, It is used for transmission or recording in pairs with information such as frame numbers.

Next, the configuration and operation of the device shown in FIG. 7 will be explained. The microphone MIC shown in FIG. 7 converts the sound wave into an electrical signal (audio signal) and provides it to a low-pass filter LPFric. In the device shown in Figure 7, a microphone M is used as a signal source.
Although an IC is used, the signal source may be a generator of other forms of audio signals or other signals.

In the following description of the embodiment, the low frequency n-wave device LPFr is assumed to have a cut-off frequency of 3 KHz.

The audio signal converted into a signal in the frequency band below 3KHz by the low-pass filter LPFr is converted into a digital signal with the required number of hints (8 bits in the following explanation) by the converter ADC, which includes a microphone μ computer. The above-mentioned D converter ADC receives the 8K output from the ripple pulse generator CG.
Layer liquid exchange is performed by pulses with a repetition frequency of Hz.

The digital signal output from the transducer ADC is an 8-hint digital signal in which the input audio signal is always sampled at a constant sampling period (1/8000 seconds in the example) and quantized. and it is the control circuit CCT
under the control of the first storage device M, (the first memory M1
, or buffer memory M+). The buffer memory M1 mentioned above is 51 in the following explanation example.
It is assumed to have a storage capacity of 2 bytes, which is divided into two parts each having half the storage capacity, and the two parts are used alternately for data writing and data reading.

Now, the recording operation of the apparatus shown in FIG. 7 is performed according to a program as shown in FIG. 9 by operating the record button Br on the operating section OP. When the record button Br in OP is operated, the program starts ("Start" in Figure 9).
) Then, in step (lr), the 9-bit sample counter, 8-hit zero point counter, and 16-bit frame counter provided in the control circuit CCT are reset.

Before the recording button Br is operated, that is, even before the "beginning" in the flowchart shown in FIG. ,
The interrupt operation in step (10r) is being carried out, and the digital signal output from the Japanese food changer ADC is sequentially stored in the buffer M, and the 9'-bit sample counter is counted up.

Also, before "E Hajime", the 9-pinto sample counter is repeatedly reset each time it reaches a full count.

In step (2r), the stored sample value is read from the buffer memory M, and the 9-bit sample counter is counted up by 1.

In step (3r), it is checked whether the sign of the sample value read in step (2r) is the same as the sign of the sample value immediately before it, and if there is no change in sign, it is determined that there is no point and step ( 2r), and when there is a change in sign, assume that the sample value read in step (2r) is the zero point, proceed to step (4r)K, and proceed to step (4r).
) counts up the 8-bit zero point counter by 1.

In step (5r), check whether the number of sample values sequentially read from the buffer memory M has reached 256 (or 512) using the count value of the sample counter of 9 hints,
The number of sample values read from buffer memory MI is 256.
(in other words, repeat steps (2r) to (4r) by 2
After repeating 56 times), proceed to step (6r) and continue reading from the buffer memory M until the number of sample values is 2.
If 56 days have not been reached, return to step (2r).

Here, as mentioned above, the number of sample values 256 successively output from the buffer memory M1 is determined by the sampling period ( 178000 seconds), the number of sample values obtained is 1, and the number 256 is 1.
This corresponds to the time length Tf of the frame signal.

In step (6r), using the count value Zc of the 8-hit zero point counter, a predetermined number, and the number 256 representing the time length Tf of the signal of one frame, the sample in the signal of the l frame is calculated. The sampling period Tc required to obtain the value sequence is calculated, and the number of samples N is also calculated.

Assuming that the sampling period Tc = 256/Zc - is the buffer memory M, the one with a storage capacity of 512 bytes is divided into two parts with a storage capacity of 1/2 as described above.
The time length of the l-frame signal is 32 milliseconds, and 25
Assume that a signal with 6 samples is being stored and read out, and now, for example, the number is set to 2, which is not the number mentioned above. If so, the sampling period Tc is Tc = 256/32X2 = 4, that is, 4/8000205 (milliseconds).

In the case of the above example, the number of samples N is 64, and for a signal of 1 frame 1/- frame 56 consisting of 256 samples and a time length of 32 ms, the number of samples N is 256 / Tc2. □=64.

Next, in step (7r), the sample values for each sampling period Tc are sequentially extracted from the buffer memory M for the one-frame signal to which the sampling period Tc is applied and the sample value sequence is taken out. In order to read out the 9-hint sample counter (address color/re), EN sample values are read out sequentially from the buffer memo IJM, using the count value at intervals of Tc as an address signal, and also the 16-hint frame. Counter count value FC frame number and sample IN
, the sampling period Tc and the N sample values are created and stored in the second storage device M2 (second memory M2), and step (8r) iC move on.

Step (8r) Teha, +6 hinoto frame counter is counted up by 1.

In step (9r), it is checked whether the frame counter of 16 hinoto has reached a full count or whether the stop button Ba has been operated, and whether the frame counter has reached a full count of 1-2 or the stop button Bs has been operated. If the condition is true, the process is over, and if not, the process returns to step (2r) and the above steps are repeated.

Frame counter count number, corresponding to 4FclC
12 bytes, 1 byte corresponding to the number of samples N, 1 byte corresponding to the sampling period Tc, and a 64-byte sample value string, the second memory has a storage capacity of 68 bytes for one frame signal. Since M2 is required,
Now, if a 64-byte memory is used as the second memory M2, 963 frames, that is, a signal of about 30 seconds or more, will be stored in the second memory M2.

As is clear from the explanation up to this point, the second memory M2
For each one-frame signal, data of sampling period Tc, sample value sequence, data of number of samples N, frame number Fc (count value Fc of frame counter), etc. are combined into one set. Although a digital signal is stored, the amount of data is greatly reduced compared to the original digital signal stored in the first memory M (buffer memory M+). The amount of data is reduced by the above-described encoding performed at the time of recording, making it possible to record and reproduce audio signals over a long period of time using a small capacity memory.

In order to perform recording and reproduction with even higher fidelity using the apparatus shown in Figure 7, it is meaningful to perform encoding based on the result of signal spectrum calculation as follows. .

That is, under the control of the control circuit CCT from the liquid exchanger ADC, the buffer memory M performs spectrum calculation on the stored signal of one frame, and from the calculation result, a certain threshold value in the signal of one frame is calculated. Find the highest frequency value in a signal component that is higher than the signal level (for example, 1/64 of the highest signal level), calculate the reciprocal of a value that is approximately twice the highest frequency value, and calculate it. value as a new sampling period Tc for that one frame signal
and by its sampling period Tc, (nofa memory M
, to obtain a sequence of sample values corresponding to the signal of one frame, and when such encoding is performed, it is necessary to obtain a sequence of sample values that corresponds to the signal of one frame. A reproduced signal with higher fidelity can be obtained than when the sampling period is set as follows.

In this case as well, using the new sampling period Tcvc obtained as described above, one sample value sequence and the data related to the sampling period Tc are selectively read out from Kunofame 31 MO IJM. The combined data is used for transmission, recording, and reproduction.

The decoding operation in the device shown in FIG. 7 is started by operating the playback button Bp of the operating section OP of the device, which is described with reference to FIGS. 1 and 4. This is similar to the decoding operation in Interpolation is performed by converting it into the related analog quantity τ1 and feeding it to the interpolation circuit CI), which produces a reproduced signal with a waveform that approximates the original signal waveform by polygonal approximation. is applied to the speaker SP through the .

As is clear from the above detailed explanation, the signal encoding, storage and reproducing apparatus of the present invention responds to changes in the original signal on the time axis in order to reduce the amount of data. A digital signal that is encoded in such a way that the sampling period is variable, and is made up of a sample value indicating the amplitude value of the original signal and data indicating the sampling period used to obtain the sample value. When decoding data, it is possible to easily obtain a reproduced signal approximated by a broken line by interpolating the signal by slow integration of the ratio between the amplitude value in the digital data and the value representing the sampling period. , according to the device of the present invention,
A high-quality reproduced signal can be easily obtained using a decoding means with a simple configuration.

[Brief explanation of drawings]

1 and 7 are boot diagrams of various embodiments of the device of the invention, FIGS. 2 and 8 are illustrative waveform diagrams, and FIGS. 3, 4, and 9 are The flowchart, Fig. 5 is a block diagram of an example configuration of an interpolation circuit, and Fig. 6 is an example characteristic diagram. MIC, = microphone ρ /, LPF r , LPF p
-・Low-pass filter, ADC...B converter, CG...Rep pulse generator, CCT..., control circuit including microphone ρ/pixel converter, OP ...Operation unit, DAC...Bi converter, IMl...
1st storage device (first memory, buffer memory),
M2... second storage device (second memory), cp...
, interpolation circuit,

Claims (1)

[Claims] By encoding the sampling period in such a way that the sampling period is variable according to the state of change in the time axis of the original signal, a sample value indicating the amplitude value of the original signal and a sampling period for obtaining the sample value can be set. This is an encoding storage/reproduction device for a signal that decodes a tenter □ rotator, which is paired with a take indicated by the digital data, and which calculates the ratio of the amplitude value and the value representing the sampling period in the digital data by dividing the ratio. A signal encoding storage/reproduction device capable of obtaining a reproduced signal approximated by a broken line, comprising means for determining the result obtained by the dividing means and means for integrating the result obtained by the dividing means.