TW484259B - Compression method and device, compression and expansion system, and recording medium - Google Patents

Abstract

The present invention relates to a compression method and device, compression and expansion system, and recording medium, especially relates to the compression and expansion way of continuous analog signal or digital signal. The solution is: in the compression side, the compression data is obtained by using the combination of the timing data of the time inteval of each sampling point and the discrete amplitude data of each sampling data, wherein the points 102a to 102f whose absolute values of their differential values are less than the specified value among the analog signal 101 in the compression side. At the same time, in the expansion side, the interpolation data between two amplitude data is obtained by using the amplitude data and timing data contained in said compression data according to the two amplitude data and the timing data inbetween of the continuous two sampling points, so that the expansion data is obtained, so that the signal remaining on a time base can be processed without the need for frequency conversion when the signal on the time base is compressed.

Description

484259 A7 B7 V. Description of the invention (1) Technical field (please read the precautions on the back before filling out this page) The present invention relates to compression methods and devices, expansion methods and devices, compression expansion systems, recording media, and particularly to continuous Compression and expansion of analog or digital signals. 2. Description of the Related Art Conventionally, in the case of transmitting or storing a signal with a large amount of information such as an image signal or an audio signal, compression and expansion are performed for the purpose of reducing the amount of information transmitted or prolonging the storage time of a storage medium. signal. Generally, in the case of compressing an analog signal, the analog signal is first sampled and digitized in accordance with a specified sampling frequency, and the obtained digital data is compressed. Printed by the Consumer Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs, for example, on the compression of video signals or audio signals, using the DCT (Discrete-Cosine-Transform: time-frequency transform filter), etc. After the data, the compression method in the frequency region is used. As a compression method for sound signals, D P CM (Different Pulse Code M o d u 1 a t i ο η: Differential Pulse Code Modulation), which is often used on telephone lines, is also used for this purpose. In addition, the compression method of D P C M is a method of encoding the difference between adjacent samples when the waveform is sampled. In addition, as a method for performing time / frequency conversion, there are also methods using a sub-channel filter, a wave filter, or an MDCT (Modified Discrete Cosine Transform). As an encoding method using this method, MP EG (Moving Picture Image Experts Group This paper size applies Chinese National Standard (CNS) A4 specification (mm) -4- 259 259 A7 B7 V. Description of invention (2): Moving day image expert group audio. In addition, the most widely used image compression system is generally known by this MP E G standard. The expansion processing of the data compressed in accordance with the compression method described above is basically performed by the opposite operation of the compression processing of the compression method. That is, the compressed digital data is processed by the frequency / time conversion and the frequency region. After the signal is converted into a signal in the time zone, the original digital data is reproduced by applying a specified expansion process. Moreover, the original data thus obtained is digitally-analog-converted as required, and is output as an analog signal. However, in the above-mentioned conventional compression and expansion methods, the signals on the time axis are converted into the signals on the frequency axis for compression, time / frequency conversion during compression, and frequency / time conversion during expansion, etc. Processing becomes necessary. Therefore, there is a problem that the processing becomes complicated, and the structure for realizing this purpose becomes very complicated. This not only shortens the processing time required for compression and expansion, but also becomes an important reason for making the device compact. In general, considering the situation of data compression and expansion, how to improve the compression rate and how to improve the quality of reproduced data is an important issue. However, in the above-mentioned conventional compression and expansion methods, there is a desire to increase the compression ratio of video signals or sound heads, and the quality of the images or sounds to be reproduced when the compressed data is expanded, and vice versa. There is a problem that the quality of the reproduced sound and the compression ratio of the video signal or the sound signal become low. Therefore, in order to achieve this paper size, the Chinese National Standard (CNS) A4 specification (210X297 mm) is applicable ~ (Please read the precautions on the back before filling this page) 1 ^^. Order printed by the Intellectual Property Bureau of the Ministry of Economic Affairs and Consumer Cooperatives -5- 484259 A7 B7 V. Description of the invention (3) It is extremely difficult for both parties to improve the compression ratio and the quality of the reproduced data. (Please read the precautions on the back before filling out this page.) This invention is to solve this problem. Those who have completed this kind of problem aim to simplify the compression and expansion processing of the signal, while shortening the processing time, and also simplifying the structure for realizing this use. Another object of the present invention is to provide a new compression and expansion method that can obtain higher compression rate and higher quality reproduction data. Another object of the present invention is to make it easier to perform compression and expansion processing without using a table. DISCLOSURE OF THE INVENTION In order to solve the above-mentioned problem, on the compression side of the first invention, the signal of the compression target is sampled at a time interval of differential. Absolute 値 being a point below the specified ，, and the discrete amplitude data and display of each sample point The combination of timing data at time intervals between specimen points is obtained as compressed data. Printed by the Employees' Cooperative of the Intellectual Property Bureau of the Ministry of Economics Here, you can also oversample the signal of the compression target, and sample the oversampled data at the interval between which the differential absolutely does not reach the designated point. Furthermore, the above-sampled data may be processed to generate average sample data of continuous specimens. Furthermore, in the development side of the first invention, the compressed data composed of a combination of amplitude data of a specified specimen point from which a signal of a compression target is extracted and timing data showing a time interval between each specimen point is used by Interpolate between the amplitude data of consecutive specimen points and the timing data between them. Interpolate between the amplitude data with the time interval displayed by the above timing data. The paper size applies the Chinese National Standard (CNS) A4 specification (210 &gt); < 297 mm) -6-484259 A7 V. Description of the Invention (4) Information to obtain expanded information. Here, a sampling function obtained from the amplitude data of two consecutive sample points and the timing data between them can also be used to obtain the interpolation data between the two amplitude data. According to the first invention configured as described above, when the signal on the time axis is compressed, time / frequency conversion is not performed, and processing is performed on the frequency axis, but processing can be performed on the time axis. It is also possible to process the compressed data in the time axis. This makes it possible to simplify the processing of compression and expansion, shorten the processing time, and simplify the configuration for this. In addition, compressed data is transmitted from the compression side, and in the case of reproduction on the expansion side, through simple interpolation operations on the time axis, the compressed data that is input to the expansion side can be sequentially processed and reproduced, so that real-time action can be realized. . Further, in this embodiment, among the sampling points, only the data of the specimen point can be obtained as compressed data, and a high compression rate can be achieved. This specimen point is equivalent to the part of the fold back in the signal of the compression target, and all include the minimum points necessary to reproduce the original data by interpolation on the expansion side. The original data has good reproducibility and can be obtained. High-quality recycled materials. In the compression side of the second invention, the digital data of the basic waveform corresponding to the input π discrete data 値 is synthesized by oversampling and moving average operation or convolution operation, and the digital After interpolation, the above-obtained digital interpolation is sampled at the time interval when the differential absolute value becomes a minimum point, and the discrete amplitude data of each specimen point and this paper scale are applied to the Chinese National Standard (CNS) A4 Specifications (210X297 mm (please read the precautions on the back before filling out this page) Order printed by the Intellectual Property Bureau Employee Consumer Cooperatives of the Ministry of Economic Affairs 484259 Printed by the Intellectual Property Bureau Employee Consumer Cooperatives of the Ministry of Economic Affairs A7 B7 V. Invention Description (5) A combination of timing data indicating the time interval between each specimen point is obtained as compressed data. In the development side of the second invention, the amplitude data and timing data included in the compressed data are used, based on two consecutive specimens. The two amplitude data of the points and the timing data between them are obtained by interpolating the interpolation data between the above two amplitude data to obtain the expanded data. According to the second invention configured as described above, when the data on the time axis is compressed, time / frequency conversion is not performed, and processing on the frequency axis can be performed on the time axis. Furthermore, when the data compressed on the time axis is expanded, Data can also be processed on the time axis. With this, it is possible to simplify the processing of compression and expansion to shorten the processing time, and also to simplify the structure. Furthermore, the compressed data is transmitted from the compression side. In the case of expansion on the expansion side, it is also possible to process the compressed data inputted to the expansion side in order to reproduce it in a simple interpolation operation on the time axis, so that real-time action can be achieved. In this way, high compression ratio can be achieved by only using the data of the specimen point as compressed data. This specimen point is equivalent to the part of the signal that is folded back in the compression target signal. Because of the minimum points necessary to reproduce the original data, the reproducibility of the original data is good, and high-quality reproduced data can be obtained. In the second invention, when oversampling and convolution operations are performed on digital data to obtain an interpolation 値, in order to obtain a certain interpolation 只, only a limited number of discrete data may be considered, and no rounding is generated. Because of the error, the correct interpolation can be obtained. Therefore, the paper size used for this compression is applicable to the Chinese National Standard (CNS) A4 specification (210X297 mm) -8- (Please read the precautions on the back first (Fill in this tribute again) 484259 Printed by the Consumers ’Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs A7 B7 V. Invention description (6) In the case of shrinking processing, the materials reproduced on the unfolding side can improve the reproducibility of the original materials before compression In the compression side of the third invention, the inputted digital data is differentiated, and the point where the polarity of the differential chirp changes is detected as the specimen point, and the discrete compression amplitude data of each specimen point is obtained with the specified binarization. Integer digital data, the combination of the compressed amplitude difference data which is the result of calculating the difference between the compressed amplitude data and the timing data indicating the time interval between each specimen point is used as a pressure. Obtained information. In the development side of the third battalion, the multiple integrals are evenly sampled with the above-mentioned compressed amplitude difference data, and a moving average operation is performed on the integral 値, and the moving average operation 値 and timing obtained by using The data is obtained by interpolating the squared interpolated data between the amplitude data of each specimen point with the time interval displayed by the above-mentioned timing data, and taking this as expanded data. In another form of the third invention, on the compression side, the input digital data is rounded by the first frame, and the differential is changed by the first rounded digital data to change the polarity of the differential frame. The point is detected as a specimen point, and as the discrete compressed amplitude data of each specimen point, digital data obtained by rounding the second frame larger than the first frame is obtained. In another aspect of the third invention, on the compression side, the compression amplitude difference data and the timing data are converted into variable-length block data. In another form of the third invention, in the unfolding side, the compressed amplitude difference data that is oversampled by an even number of times is applied to the Chinese paper standard (CNS) A4 specification (210X297 mm) by the above paper with the timing data ) I (Please read the precautions on the back before filling this page) -9-484259 A7 Printed by the Consumers' Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs B7 V. Description of the invention (7) Each of the separated points shown in the specimen The position sign is reversed, and multiple integrals are obtained from the data rows thus obtained. In another aspect of the third invention, in the development side, multiple integrations and moving average calculations of the above-mentioned even-amplified compressed amplitude difference data are performed at each division of each specimen point. According to the third invention configured as described above, when the signal on the time axis is compressed, it is possible to perform processing on the frequency axis without performing time / frequency conversion, and to perform processing on the time axis. When the compressed data is unfolded, it can be processed on the time axis. This simplifies the processing of compression and expansion, can shorten the processing time, and can simplify the configuration to this. In addition, during the unfolding process, by using a simple square interpolation operation on the time axis, the input compressed data can be sequentially processed and reproduced without using table information, and real-time operations can be realized. According to the third invention, compressed data can be generated only from discrete data with little difference between the amplitude data of the sample point of the digital data, the polarity change, and the timing data indicating the time interval between the occurrence of each sample point. In addition, since the amplitude data of each specimen point is rounded up by a predetermined scale, the data length of the amplitude data can be reduced by one digit per character, and the amount of data can be greatly reduced here. Furthermore, in the third invention, instead of using the rounded amplitude data as the compressed data as it is, the differential data is obtained. As the compressed data, the number of bits necessary for the compressed data can be made smaller and more Reduce the amount of data. In addition, according to the other features of the third invention, the compressed amplitude difference data and timing data obtained as described above are coded into a variable-length block capital paper size that applies the Chinese National Standard (CNS) A4 specification (210X297 mm) (Please read the precautions on the back before filling in this page) -10- 484259 A7 B7 V. Description of the invention (8) Material, as the final compressed data, can further increase the compression rate. (Please read the precautions on the back before filling in this page.) Also, according to the third invention, the turning point in the signal of the compression target is detected as the specimen point, and the original data is reproduced by interpolation on the expansion side. The minimum necessary points are all included in the compressed data. Therefore, by improving the reproducibility of the original data, high-quality reproduced data can be obtained. In addition, according to the other features of the third invention, the differential point is appropriately rounded digital data to detect the specimen point, so that the position of noise components or unwanted signal components can not be detected as the specimen point. It is possible to detect only the correct position as the specimen point. This makes it possible to improve the reproducibility of the original data before compression with respect to the expanded data reproduced on the unfolding side. Printed by the Consumer Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs. According to the other features of the third invention, the compression amplitude difference data symbol that is evenly oversampled at the middle position of each division of each specimen point is reversed. When multi-integration and moving average calculations are performed on the data column whose symbols are reversed, the rounding error on the compression side can be canceled, and the digital waveform with a change in amplitude 更 can be reproduced more smoothly. In addition, according to another feature of the third invention, in the development side, the multiple integration is performed at each division of each specimen point, which can eliminate a cumulative error due to integration and can reproduce a more accurate digital waveform. Brief Description of Drawings Fig. 1 is a diagram for explaining a compression method according to the first embodiment. Fig. 2 is a diagram for explaining a developing method according to the first embodiment. This paper size applies the Chinese National Standard (CNS) A4 specification (210X297 mm) -11-484259 Α7 Β7 Printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs 5. Description of the invention (9) Figure 3 shows the moment shown in Figure 2 The interval from T 1 to T 2 is a diagram showing the interpolation principle of this embodiment. Fig. 4 is a diagram showing an example of a sampling function. Fig. 5 is a graph showing the relationship between discrete data and the interpolation between these. Fig. 6 is a diagram for explaining an interpolation operation expression of a specific example of data interpolation processing on the development side. FIG. 7 is a diagram showing the processing results for the case where oversampling and convolution operations are applied between the times T 1 and T 2 shown in FIG. 2. Fig. 8 is a diagram showing other processing results in a case where oversampling and convolution operations are applied. Fig. 9 is a block diagram showing a configuration example of a compression device according to the first embodiment. I 1 0 is a block diagram showing an example of the configuration of the average unitary interpolation data generating section in FIG. 9. _ 1 1 is a block diagram showing a configuration example of the timing synthesizer in FIG. 9 _ 1 2 is a block diagram showing a configuration example of a deployment device according to the first embodiment. _ 1 3 is a block diagram showing a configuration example of a compression device according to the second embodiment. ® 1 4 is a block diagram showing a configuration example of a deployment device according to the second embodiment. Fig. 15 is a waveform diagram showing a digital substrate used in the second embodiment. The size of this paper applies to China) A4 size (21 × 297 mm) (Please read the notes on the back before filling out this page) -12- 484259 A7 B7___ V. Description of the invention (1〇) Figure 16 illustrates the second implementation This figure shows the operation examples of type oversampling and convolution operation. (Please read the precautions on the back before filling out this page.) Figure 17 shows the function generated by the digital substrate waveform shown in Figure 15. Fig. 18 is a diagram showing a configuration example of the oversampling circuit shown in Fig. 13. Fig. 19 is a diagram showing an example of digital data input into the compression device of the second embodiment. Fig. 20 is a diagram showing an example of data obtained by outputting the digital data shown in Fig. 19 through the oversampling circuit of Fig. 13. FIG. 21 is a diagram showing a configuration example of the differentiator shown in FIG. 13. FIG. 22 is a diagram showing an example of data obtained by passing the oversampled data shown in FIG. 20 through the differentiator of FIG. 13. Fig. 23 is a diagram for explaining the processing in the case where the same absolute differential is continuously generated. Fig. 24 is a block diagram showing an example of a configuration for detecting a specimen point by performing double differentiation. Printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs. Figure 25 shows an example of compressed data that is output from the oversampled data shown in Figure 20 through the results of the compressed data generation unit shown in Figure 13. Fig. 26 is a diagram for explaining other examples of data interpolation processing on the expansion side. Fig. 27 is a view showing a case where the compressed data shown in Fig. 25 is used as an object for expansion processing. The expansion processing section of Fig. 14 An example of the expanded data being output. This paper size applies the Chinese National Standard (CNS) A4 specification (210X297 mm) -13 · 484259 A7 B7 V. Description of the invention (n) Figure 2 8 is a block diagram showing a configuration example of a compression device according to the third embodiment . (Please read the precautions on the back before filling this page) Figure 2 9 is a diagram explaining the operation principle of the timing generator and amplitude generator in Figure 28. Fig. 30 is a block diagram showing a configuration example of the timing generator in Fig. 28. FIG. 31 is a diagram showing a detailed configuration example of a part that generates a timing pulse. Fig. 32 is a diagram for explaining an actual operation example of the compression process performed by the compression device of the third embodiment. FIG. 33 is a diagram showing an example of the structure of the serial compressed block data according to the third embodiment. Fig. 34 is a block diagram showing a configuration example of a deployment device according to the third embodiment. Fig. 35 is a diagram showing a detailed configuration example of the square interpolation data generating section in Fig. 34. Fig. 36 is a diagram for explaining the actual operation utilization of the unfolding processing performed by the unfolding device of the third embodiment. Printed by the Employees' Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs. Figure 37 shows an example of the original data and the expanded data before compression. Comparison table of main components 1 Oversampling circuit 2 PLL circuit 3 Differentiator 4 Compressed data generation department The paper size applies Chinese National Standard (CNS) A4 specification (21 × 297 mm) -14-484259 A7 B7 V. Description of the invention (12 ) 5 Error correction coding section 6 Data memory 1 1 Error correction circuit 1 2 Clock generator 1 3 Timing generator 1 5 Unrolled processing section 1 6 D / A converter 2 1 Normalized data record 2 2 Phase shift section 4 1 1st differential section 4 2 Rounding calculation section 4 3 2nd differential section 4 4 Polarity change check 4 6 Data generating section 1 1 1 Timing generator 1 1 2 Amplitude generator (please read the precautions on the back first) (Fill in this page) 41. Order the best form for printing and implementing the invention by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs. Below, one implementation form of the present invention will be described with reference to the drawings. (First embodiment type) Fig. 1 is a diagram for explaining a compression method according to the first embodiment, and shows an example of an input analog signal to be compressed. Fig. 2 is a diagram for explaining a developing method according to this embodiment. This paper size applies the Chinese National Standard (CNS) A4 specification (210X297mm) -15- 484259 A7 Printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs B7 V. Description of the invention (13) First, the compression processing will be described using FIG. In this embodiment, from the input analog signal 1 0 1, the differential absolute value 値 (the tilt of the signal) is detected to be a point below the specified value including A 0 〃 (hereinafter, referred to as the specimen point) 1 0 2 a ~ 1 0 2 f. Further, obtain digital data that converts the amplitude of each of the specimen points 1 0 2 a to 1 0 2 f by analogy to digital digits 値 and timing data indicating the time interval at which each specimen point 1 0 2 a to 1 〇 2 f appears. , The combination of this amplitude data 値 and timing data 传送 is transmitted or recorded as compressed data. In the example of FIG. 1, as the amplitude data 値 of the digits 1 0 2 a to 1 2 2 f for each specimen point, '' 7, 3, 9, 1, 6, 3 〃 is obtained as the specimen point 1 0 2 a ~ 1 0 2 f at the time of occurrence T 1-T 2, D 2-T 3, T 3-T 4, T 4-T 5, .T 5-T 6 Time interval data, find '' 5, 7, 3, 3, 3 〃. The numbers displayed as timing data here indicate the number of clocks based on a certain sampling frequency. At the time point of time T 1, the amplitude data of the specimen point 1 0 2 a 値 “7” and the timing data indicating the time interval from the time when the previous specimen point (not shown) is detected (Not shown), the combination of these data frames is transmitted or recorded as compressed data at time T 1. Then, at the time point T 2 when the specimen point 1 0 2 b is detected, timing data of the time interval from the time T 1 when the specimen point 1 0 2 a is detected before can be obtained 可以获得 '' 5 〃 And the amplitude data of the specimen point 1 0 2 b 値 3 3 故, the combination of these data 本 This paper size applies the Chinese National Standard (CNS) A4 specification (210X297 mm) (Please read the precautions on the back before (Fill in this page) -16- 484259 A7 B7 V. Description of the invention (14) (5, 3) is transmitted or recorded as compressed data at time T2. (Please read the precautions on the back before filling in this page.) Then, at the time point T 3 when the specimen point 1 〇 2 c is detected, you can get the time before the specimen point 1 0 2 b is detected. The timing data of the time interval from T 2 値 '' 7 〃 and the amplitude data of the specimen point 1 〇 2 c 値 '' 9 〃 For this reason, the combination (7, 9) of these data is used as the compressed data at time T 3 Transmit or record. The same applies below, showing the timing data of the time interval between time T3-T4, time T4-T5, time T 5-T 6 and the specimen points detected at time T 4, T 5, T 6 1 〇 2 The combination of amplitude data 値 of d, 10 2 e, 1 〇 2 f (3, 1), (3, 6), (3, 3) are transmitted as compressed data at time T 4, T 5, T 6 respectively. Or record. Next, the expansion processing of the data compressed as shown in FIG. 1 will be described with reference to FIG. 2. In the case of compressing the input analog signal 1 0 1 in the manner shown in FIG. 1, the obtained compressed data becomes (*, 7) (5, 3) (7, 9) (3, 1) (3, 6) (3, 3). Also, * is printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs. Enter the compressed data in the order shown here on the expansion side. On the unfolding side, first, the data of the amplitude data 値 a 7 〃 and the timing data 値 '' 5 输入 which are inputted first are used to generate the data of the waveform a 1 by interpolation. Next, from the above timing data 値 '5 上述 and the next input amplitude data 値 3 输入 2 data 値, the data of waveform a 2 is generated by interpolation. Then, from the above-mentioned amplitude data (A 3) and the entered standard paper size, the Chinese National Standard (CNS) A4 specification (210X297 mm) is applied. &Quot; -17- 484259 Printed by the Employee Consumption Cooperative of Intellectual Property Bureau, Ministry of Economic Affairs A7 B7 V. Description of the invention (15) The data 値 '' 7 〃 2 data 値, the data of waveform b 2 is generated by interpolation. Further, from the above-mentioned timing data 値, 7 〃, and then the input amplitude data 値 '' 9 〃, data of waveform b 1 is generated by interpolation. The following is the same. The waveforms c 1, c 2, d 2, d 1, e 1, e 2 are sequentially generated from the combination of the amplitude data 値 and the timing data 输入 which are sequentially input. With the above processing, the digital signals of waveform a1, b1, c1, d1, and e1 (the upper part of Fig. 2) and waveforms a2, b2, c2, d2, and e2 are continuous. A continuous digital signal (lower part of Figure 2) is generated. Then, the two digital signals thus generated are added to each other to perform digital-to-analog conversion, and the original analog signal shown in Fig. 1 is reproduced. Fig. 3 shows the interval between Ding 1 and Ding 2 at the time shown in Fig. 2 and is not displayed. Fig. 3 (a) shows the two waveforms a1, a2 before addition, and Fig. 3 (b) shows The synthesized waveform a 1 + a 2 reproduced by the addition. In FIG. 3 (a), D 1 is the amplitude data of time T 1 (in the example of FIG. 2, ', 7 〃), and D 2 is the time Amplitude data of T 2 (in the example in FIG. 2, 3, 3), T is timing data indicating the time interval between time τ 1 and τ 2 (in the example in FIG. 2, it is '' 5 〃) And t are arbitrary timings between time T1 and time T2. As shown in FIG. 3 (a), the amplitude data 値 D 1 at time T 1 and the timing data 値 T representing the time interval between time T 1 and T 2 are used, and any time between time T 1 and T 2 is used. Timing t is a variable, which means that according to the paper size, the Chinese National Standard (CNS) A4 specification (210X297 mm) is applied (please read the precautions on the back before filling this page), order-13- 484259 A7 B7 ____ V. Invention Explanation (彳 6) According to the clock of a certain sampling frequency, the data of waveform a 1 is generated by interpolation while increasing the timing t one by one. (Please read the precautions on the back before filling this page.) Also, use the amplitude data 値 D 2 at time T 2 and the timing data 値 T indicating the time interval between time T 1-T 2. Similarly, use time t as Data of the waveform a 2 by interpolation. Further, by adding the data of the waveforms a 1 and a 2 thus generated at the above-mentioned timing t as a variable, the waveform shown in FIG. 3 (b) is synthesized. In this way, the original analog signal before being compressed can be reproduced. In the following, the principle that the original analog signal can be reproduced by using the expansion processing by interpolation as described above will be described. In general, in order to obtain continuous analog signals from discrete digital data, interpolation is performed between the discretely input digital data to virtually increase the sampling frequency. Generally, such data interpolation is performed using a specified sampling function. Figure 4 shows an example of the sampling function. In the example in Figure 4, only the specimen point at t = 0 becomes 1 1 gt, all other specimen points at equal intervals (t = soil 1, soil 2, ± 3, soil 4, ...), 値All of them have been printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs. Fig. 5 is a diagram illustrating a general data interpolation operation using such a sampling function. In FIG. 5, it is considered that the discrete data 値 of equally spaced specimen points t 1, t 2, t 3, and t 4 are set to Y (t 1), Y (t 2), Y (t 3), and Y ( t 4), for example, to find the case where the specified position 丨 0 (distance 3 from 12) is interpolated 値 v corresponding to the specimen points t 2 and 丨 3. This paper size applies the Chinese National Standard (CNS) A4 specification (210X297 mm) ~ -19- 484259 A7 B7 V. Description of the invention (17) (Please read the precautions on the back before filling this page) Generally use the sampling function to find When the interpolation 値 y is obtained, about each of the discrete data given, find 値 of the sampling function at the interpolation position t 0, and use this to perform the convolution operation. Specifically: each sample point of t 1 to t 4 makes the peak height of the center position of the sampling function consistent, and obtains the 値 of the sampling function of the individual interpolation position t 〇 at this time (indicated by X mark) And add them all up. This interpolation process is performed by sequentially shifting the interpolation position t 0 with the passage of time (in parallel with the increase in the sampling clock), and sequentially obtain the interpolation 値 y (t 0) that changes continuously. By this, a continuous analog signal can be obtained which is smoothly connected between discrete data. This embodiment uses such a data interpolation processor. That is, as shown in FIG. 3 (a), at time T2, the first specimen is obtained from the amplitude data 値 D1 (= 7) and the timing data 値 T (= 5) that have been input. At the point (time T 1), at the same time as the waveform a 1 of the sampling function of 値 other than 〃, the amplitude data 値 D 2 (= 3) is printed and timed by the consumer cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs. Data 値 T (= 5), find the waveform a 2 that forms part of the sampling function using 値 other than '' 0 中 in the second specimen point (time T 2). Moreover, by adding each of these waveforms a 1 and a 2 sequentially with each interpolation position t that moves sequentially over time, an analog signal between the continuous discrete data D 1 and D 2 is obtained. Here, after obtaining the data of the waveforms a 1 and a 2 respectively, they will be added together to explain, but at the time point of time T 2, all the data for generating the synthesized waveforms a 1 and a 2 have been For this reason, you can also apply the Chinese National Standard (CNS) A4 specification (210X297 mm) to this paper size. -20-484259 Printed by the Consumers ’Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs A7 _ B7 __ V. Description of the invention (18) The specified calculation formula, etc., once obtained the synthesized waveform. However, in this embodiment, as shown in FIG. 1, on the compression side, the input analog signal 1 0 1 which is smoothly changed is sampled at a time interval whose differential absolute 値 becomes a point below the specified 値 to obtain Discrete data as compressed data. Therefore, the interval between individual specimen points for which discrete data can be obtained is not limited to equal intervals. In many cases, it is an indefinite interval. (In the example of FIG. 1, the interval between each specimen point becomes `` 5, 7, 3, 3, 3 indefinite intervals). Therefore, in the unfolding side shown in FIG. 2, for example, when the interpolation 値 between the times T 1 and T 2 is obtained, as shown in FIG. 3, the time between the specimen points at the times T 1 and T 2 The interval uses only the sampling functions a 1 and a 2 and performs the convolution operation as described above. The time interval between the specimen points is different. The other sampling functions bl, b2, cl, c2, dl, d 2, e 1, e 2 are not taken into account in this convolution operation. For example, when interpolating between time T 2-T 3 is obtained, the time interval (= 7) between the specimen points at time T 2 and T 3 is rolled using only the sampling functions b 1 and b 2 For product operations, other sampling functions a 1, a 2, c 1, c 2, dl, d2, el, and e2 with different time intervals between specimen points are not considered during the convolution operation. The same applies to the interpolation between other specimen points. That is, in the compression-expansion system of this embodiment, on the compression side, the input analog signal 1 0 1 which is smoothly changed is sampled at an indefinite time interval below which the absolute value 値 becomes the specified value, and discrete amplitudes are obtained as compressed data The timing of the data frame and the variable time intervals that represent these. And this paper size applies Chinese National Standard (CNS) A4 specification (21〇 > < 297 mm) (Please read the notes on the back before filling in this search) -21-484259 A7 _ B7 V. Invention Description (19) (Please read the notes on the back before filling out this page) In the side, following the amplitude data 定时 and the timing data 被 included in the compressed data, as in the compressed side, a continuous analog signal is generated between discrete data that is connected to indefinite time intervals by using the interpolation function of the sampling function described above. . Next, a specific processing example of the above-mentioned data interpolation will be described below. As described above, for example, when the interpolation 値 between the times T 1-T 2 is obtained, only the time interval between the amplitude data 値 of the times T 1 and T 2 and the time between the times T 1-T 2 is used. Timing data: the obtained sampling functions a 1 and a 2. That is, all the data necessary to obtain the interpolation position t at each of the interpolation positions t between time T 1 and T 2 are obtained at the time point T 2. At this time point, it can be reproduced as shown in FIG. 3 (b). The original analog signal. Therefore, in this embodiment, two amplitude data 値 D1, D2 and timing data 値 T indicating the time interval can be obtained at each discrete time of T1 ~ T6, using these data 値, By following the interpolation expressions described below to calculate the interpolation 値, the original analog signals are sequentially reproduced. FIG. 6 is a diagram for explaining the interpolation expression. Printed by the Consumer Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs as shown in Fig. 6, it has amplitude data (D 1 and D 2 interpolation between the two specimen points). The numbers X 1 and X 2 are expressed as a function that is continuous at intermediate time points. That is, in this embodiment, the two specimen points are divided into the first half and the second half, and the interpolation 値 is calculated using the quadratic functions X 1 and X 2 respectively. Here, the timing data T of the time interval between the specimen points may be odd or even. In the case of odd numbers, the paper size applies the Chinese National Standard (CNS) A4 specification (210X 297 mm) just in the middle. ) -22-484259 A7 B7 _ ___ V. Description of the invention (20) (Please read the notes on the back before filling this page) There is no state where the position t is interpolated. Therefore, in the present embodiment, by performing twice the oversampling at the time of compression, the timing of the obtained lean material is suspended to become an even number. That is, the five timings shown in FIG. 1 値 5, 7, 3, 3, 3, / r are actually oversampled by 2 times, and 、, 10, 1, 4, 6, 6, 6 〃値 is transmitted or stored. In Fig. 6, the time interval between specimen points is represented by 2T after being oversampled. In Figure 6, the two second-order functions X 1 and X 2 are respectively represented by: X 1 = D 1 + at 2… (1) X 2 = D 2-a (t-2 T) 2… (2) . In addition, these functions X 1 and X 2 are continuous at exactly the intermediate time point T of the continuous specimen point, so X 1 = X 2 (t = T) ". (3) Here, in formula (3) Substituting into the formula (1) (2), becomes ...

Dl + aT2 = D2-aT2 ". (4), solve a for it, and become a =-(D1-D2) / 2T2-(5). Therefore, by substituting this formula (5) into formula (1) (2), it is printed by the Consumer Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs: xl = Dl— {(Dl-D2) / 2T2} t2 ... ( 6) x2 = D2 + {(D1-D2) / 2T2} (2 T- t) 2… ⑺ That is, the interpolation is performed by using the variable t at the interpolation position t in the order of the original 2 times the sampling frequency. The operation of the above equations (6) and (7) can reproduce the original analog signal. In this embodiment, the signals formed by the amplitude data 値 and the timing data 列 are listed at each discrete time T 1 ~ T 6. The paper size is applied to the Chinese National Standard (CNS) A4 (210X297 mm) -23- 484259 A7 _ B7 ___ V. Description of the Invention (21) Enter and perform such interpolation operation processing in order. (Please read the precautions on the back before filling this page.) In the example in Figure 2, at the time point T 1 and T 2 the amplitude data of the specimen point 値 and the timing data between them are input. Interpolation between specimen points instantly reproduces the original analog signal. Furthermore, at the time point T 3, the amplitude data of the specimen point and the timing data between the specimen points T 2-T 3 are inputted, and an interpolation operation is performed therebetween to instantly reproduce the original analog signal. The following processes are also performed in the same order. As described above, in this embodiment, instead of performing time / frequency conversion on an analog signal to be compressed, compression and expansion are performed on the time axis, processing is not complicated, and the configuration can be simplified. In addition, in the case where compressed data is transmitted from the compression side and reproduced on the expansion side, the input compressed data can be sequentially processed by simple interpolation operations on the time axis, and real-time actions can be realized for reproduction. In addition, the interpolation calculation processing shown in the above formulas (6) and (7) may be implemented by hardware such as a logic point circuit, or may be implemented by a DSP (Digital Signal Processor) or software ( Programs stored in ROM or RAM). Printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs Next, other processing examples of the above-mentioned data interpolation will be described below. Here, it is explained that the amplitude data 値 D1, D2 and the timing data 値 T between the continuous sample points are used to obtain the interpolation 値 by performing oversampling and convolution operations twice. FIG. 7 is a diagram showing the processing results for the cases (d 1 = 7, D2 = 3, and T = 5) where the oversampling and convolution operation is applied between the times T 1 and T 2 shown in FIG. 2. In Figure 7, the leftmost column r 1 to 4 is a figure of 2 amplitudes. The capital paper size is applicable to the Chinese National Standard (CNS) A4 specification (210X 297 mm) ~ -24- 484259 A 7 B7 __ V. Description of the invention (22) The difference between materials D1 and D2 (= D1-D2). The numbers of ', 4' are arranged in the vertical direction. 10 shows the amplitude data. D1, (please read the precautions on the back before filling this page) D2 difference 値 1 clock 1 clock Sequentially delayed, it is maintained in a state where only the timing data 値 T = 5 is twice the number of D-type flip-flops (not shown). The number 値 of the second row r 2 from the left shows the result of shifting each specimen 第 of the first row r 1 by 1 clock. Furthermore, each of the third to fifth columns r 3, r 4, and r 5 shows the result of shifting the nature of each of the targets in the second column r 2 and shifting each clock in order. In addition, the 6th row r 6 number row is the result of adding the 1st to 5th row r 1 to r 5 number row between the corresponding rows, that is, for the 1st to 5th column r 1 to The sequence of numbers of r 5 is the result of a 5-stage convolution operation. The 7th to 10th columns r 7, r 8, r 9, and r 1 0 are each a series of samples showing the results of the convolution operation in the 6th column r 6 and further 1 clock 1 The results are sequentially shifted. Printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs. The first 11th column r 1 1 is the result of adding the 6th to 10th columns r 6 to r 1 0 That is, for the 6th to 10th columns r 6 to r 10, the result of the convolution operation of 5 stages is performed. In addition, the number 12 of the 12th column r 1 2 shows the results of convolution operations performed on the 11th column r 1 1 and the results are shifted by 1 clock. Further, the number row of the 13th row r 1 3 is the result of adding the number row of the 11th? Fj r 1 1 and the 12th row r 1 2 between the corresponding rows. For example, M t represents the number of rows 13 and r 1 2 obtained by addition. The paper size is applicable. National Standard (CNS) A4 (210 × 297 mm) -25- 484259 A7 _ B7 ____ 5 The invention description (23) column, the interpolation of the two amplitude data 値 D1, D2 interpolation Sout is expressed by the following formula:

Sout = Dl—Mt (Dl— D2) / (8TxT) = 7-M t / 5 0 If this interpolation 値 S ο ut is depicted, it becomes a curve as shown in Fig. 7, which can be reproduced at the time T 1 -T as shown in Fig. 1. The two analog signals are the same. That is, in the case where the two amplitude data 値 D 1 and D 2 have opposite magnitude relationships, for example, in the case of D 1 = 3 and D 2 = 7, the result of the oversampling and convolution operation at 2 times becomes as shown in FIG. 8 As shown. By sequentially performing such data interpolation processing among all the specimen points, the original analog signal shown in FIG. 1 can be reproduced. In addition, the operations shown in FIG. 7 and FIG. 8 described above can be implemented by appropriately combining the hardware configuration of a D-type flip-flop, an adder, and a multiplier for holding the data 値 1 clock 1 clock while delaying and holding the data achieve. Next, a configuration for implementing the compression processing and expansion processing described above will be described. Fig. 9 is a block diagram showing a configuration example of a compression device according to this embodiment. In FIG. 9, the input analog signal 1 〇 1 is easy to detect the specimen point. After removing noise by LP F 〇 3, it is converted into digital data by an A / D converter 104. At this time, the A / D converter 1 〇4 follows the input clock CK of a specified frequency (for example, in the case of a sound signal, 4 4. 1 Κ Η z) through a PLL (Phase Locked Loop) circuit. The clock C κ 1 of the twice frequency (8 8 2 8 Η z) generated by 1 0 is subjected to a / D conversion processing. This paper size applies to China National Standard (CNS) A4 (210 X297 mm) " " -26- (Please read the precautions on the back before filling this page) Order f System 484259 A7 B7 V. Description of the invention (24) (Please read the notes on the back before filling in this page) Average 値 Interpolation data generating section 10 6 Oversampling by 2 times the frequency by A / D converter 1 0. 4 Digital data to be output. Furthermore, for the plurality of specimens 藉 obtained by this, an average 値 is calculated between successive specimens 値 to generate interpolated data. That is, if over-sampling of the digital data is input, it is possible to obtain the same 値 2 2 consecutive series. For such a number 如, if the continuous 连续 is used to calculate the average 値, and 値 is maintained as it is, if the continuous 値 is used to calculate the average 値, the intermediate 値 between the different 値 s can be obtained. Fig. 10 is a block diagram showing an example of the structure of the average 値 interpolation data generating section 106. In Fig. 10, the D-type flip-flop 201 follows the input clock C K which becomes the reference, and holds the digital data outputted by the A / D converter 104 of Fig. 9. In addition, the D-type flip-flop 2 0 2 connected to the subsequent stage keeps the digital data output from the D-type flip-flop 2 01 according to the clock C K 1 which is twice the frequency of the input clock C K 0. Printed by the Consumer Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs. The adder 2 0 3 is added to the digital data of the 2 D-type flip-flops 2 0 1 and 2 2 and the result is output to 1/2. The multiplier 2 0 4. The 1/2 multiplier 2 0 4 makes the addition result of the adder 2 0 3 1/2 times, and the result is kept at the D-type flip-flop 2 according to the clock CK 1 of the double frequency. 0 5. Furthermore, the digital data held in this D-type flip-flop 205 is output as interpolation data accompanied by double oversampling. By setting such an average 値 interpolation data generation unit 1 06, it is possible to make the absolute data of the differential 値 become the timing data of the time interval between the specimen points below the specified 値, which often becomes even, even if it is not performed because it is even or This paper size applies the Chinese National Standard (CNS) A4 specification (210X 297 mm) 27- 484259 A7 B7 __ V. Description of the invention (25) (Please read the precautions on the back before filling this page) Odd numbers and complicated processing separately The processing is also OK. In addition, instead of oversampling, calculating the average value of consecutive specimens and outputting them, the phased data waveform can become a smoother waveform closer to the original analog waveform. Therefore, it is possible to improve the reproducibility of the original analog signal when performing the expansion processing by the expansion device described later. By this means, the digital data which is oversampled by the interpolation data generation unit 106 is input to the timing synthesizer 107 and the compression processing unit 108. The timing synthesizer 107 differentiates the digital data supplied by the average 値 interpolation data generation unit 106 to detect the specimen point. In addition, the sampling clock indicating the timing of this detection point and the timing data (the number of double frequency clocks C K 1) indicating the time interval between each specimen point are obtained and output. Fig. 11 is a block diagram showing a configuration example of the timing synthesizer 107. In FIG. 11, the differentiator 3 0 1 differentiates the digital data inputted by the average 値 interpolation data generating unit 1 0 6. In addition, the specimen point detection unit 3 02 detects that the differential of the digital data is absolutely not more than the designated specimen point based on the result of differentiation by the differentiator 3 01. The Ministry of Economic Affairs, Intellectual Property Bureau, Employees' Cooperatives, and the Producer's Department produced 3 3 counts the number of clocks CK 1 at twice the frequency that were supplied until the next specimen point was detected. When outputting timing data, the sampling clock indicating the timing of the detection points of each specimen point is output. In addition, this timing generating section 3 03 also generates a read clock which will be described later. In addition, the compression processing unit 108 follows the sampling clock output by the timing synthesizer 107 and extracts only the digital data of the corresponding specimen point position and outputs it as amplitude data. FIF 0 memory 1 〇9 according to this paper size applies Chinese National Standards (CNS) A4 specifications (210 father 297 mm Ding '' '-23- 484259 A7 B7 __ 5. Description of the invention (26) (Please read the back (Please fill in this page again for the matters needing attention.) The sampling timing is used to take in the amplitude data of each specimen point output by the compression processing unit 108 and the timing indicating the time interval between each specimen point output by the timing synthesizer 107. The combination of data is sequentially read out in accordance with the readout clock. The readout of the combination of amplitude data and timing data is transmitted or recorded as compressed data. Figure 12 shows the structure of the expansion device according to this embodiment. The block diagram of the example. In Figure 12, the clock generator 4 0 1 generates a clock CK 1 with a frequency twice as high as the reference input clock CK 〇. Moreover, the timing generator 4 0 2 accepts to be included in the compression The timing data in the data is generated by the above-mentioned 2 times frequency clock CK 1 and the readout clock representing an indefinite time interval is the same as that between the specimen points detected on the compression side. The type D type is printed by the Consumer Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs Flip-flop 4 0 3 will be included The amplitude data in the compressed data is sequentially taken in and held in accordance with the readout clock generated by the timing generator 4 02, and output to the unrolling processing section 4 0. The D-type flip-flop 4 0 3 The amplitude data of the input and output sections, that is, the amplitude data of the D-type flip-flop 4 0 3 at the timing of a readout clock, and the D-type flip-flop 4 0 3 should be held at the next read-out clock. The amplitude data (amplitude data of two consecutive sample points) is input to the expansion processing unit 4 0 4 〇 The expansion processing unit 4 0 4 uses the two amplitude data thus input and the timing generator 4 0 2 The input timing data is generated by the interpolation operation according to the above formula (6) (7), or the convolution operation shown in FIG. 7 and FIG. 8 to generate digital interpolation data between each specimen point. In addition, the digitally interpolated data thus generated is converted into a paper-like standard by a D / a converter 4 0 5 and the paper standard is applicable to the Chinese National Standard (CNS) A4 (21〇X 297 mm) -29- 484259 A7 B7 5 Explanation of the invention (27) After the comparison signal, it is output as a reproduction analog signal through LPF 4 06. (Please first Read the notes on the reverse side and fill in this page) As mentioned above in detail, in this embodiment, on the compression side, the input analog signal is smoothly changed by sampling at variable time intervals that are different from absolute 値 to specified 値. As the compressed data, discrete amplitude data 値 and timing data indicating their indefinite time intervals can be obtained. Furthermore, in the expansion side, the amplitude data 値 and timing data 包含 included in the compressed data are followed in order to Read out discrete data at the same indefinite time interval on the compression side and output continuous analog signals between them by interpolation. Therefore, when compressing and expanding analog signals on the time axis, frequency conversion can be performed on the time axis. For processing. Therefore, the processing of compression and expansion is not complicated, and the structure for this can be simplified. In addition, in the case where compressed data is transmitted from the compression side and reproduced on the expansion side, the compressed data inputted on the expansion side can be sequentially processed and reproduced by simple interpolation operations on the time axis, and real-time operation can be realized. Printed by the Consumer Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs. In this embodiment, the differential of the digital data is absolutely designated as the designated point and detected as the specimen point. The amplitude data of each detected specimen point is displayed and displayed. Because the timing data of the time interval at which each specimen point appears does not generate compressed data for transmission or recording, it is possible to obtain only the specimen point data as compressed data, which can achieve a high compression rate. In addition, according to this embodiment, the turning points existing in the signal of the compression target are detected as the specimen points, and the minimum necessary points for reproducing the original data are included in the compression by the interpolation operation on the expansion side. Information. Therefore, the reproducibility of the original materials can be improved, and the high paper size can be obtained. The Chinese National Standard (CNS) A4 specification (210X297 mm) can be obtained. 2. Description of invention (28) Reproducible data of quality. In the first embodiment described above, the case where the input signal to be compressed is an analog signal is described, but the input signal may be a digital signal. In this case, the L P F 103 and A / D converter 104 shown in Fig. 9 and the D / a converter 400 and LPF 406 shown in Fig. 12 are unnecessary. Further, in the first embodiment described above, although the oversampling is performed twice in the average 値 interpolation data generating section 106, it is not limited to two times as long as it is an even number. (Second Embodiment) A second embodiment of the present invention will be described below with reference to the drawings. In the second embodiment, digital data is input and compressed. In the compression device of this embodiment, first, η-fold oversampling and moving average operation or convolution operation (hereinafter, referred to as convolution operation) are performed on the digital data input as a compression target to obtain the internal data. Interpolate smoother data between successive data. Then, from a series of data obtained in this way, the position where the absolute differential 値 becomes smaller than the position before and after, that is, the point where the absolute absolute 値 becomes the smallest is detected as the specimen point. Furthermore, the amplitude data 値 of each detected sample point and the timing data 表示 indicating the time interval at which each sample point appears are obtained, and the combination of this amplitude data 値 and the timing data 压缩 is transmitted or recorded as compressed data. Figure 1 3 shows this embodiment according to the above-mentioned compression method. The paper size applies the Chinese National Standard (CNS) A4 specification (210X297 mm). -31-(Please read the precautions on the back before filling this page) 484259 A7 B7 __ V. Block diagram of an overall configuration example of the compression device of the invention description (29). (Please read the precautions on the back before filling this page) As shown in Figure 1 3, the compression device of this embodiment includes: oversampling circuit 1, PLL (Phase Locked Loop) circuit 2, and differentiator 3 , And a compressed data generating section 4, an error correction coding section 5, and a data memory 6. The above-mentioned oversampling circuit 1 performs an oversampling and convolution operation of n times the digital data that is input as a compression target, to obtain an interpolator of the digits between the complemented discrete data. In the example shown in Figure 13 below, as the data of the compression target, the sound data sampled at a frequency of 4 4. 1 κ Η ζ is input, and this is passed at a frequency of 8 times (3 5 2 · 8 Κ Η ζ). At the same time as sampling, a convolution operation is performed. Then, a series of oversampling data obtained thereby is output to the differentiator 3 and the compressed data generating section 4. The PLL circuit 2 generates a clock 8 CLK which is 8 times the frequency (352 · 8KHz) from the input clock CLK of the reference frequency (44. ΙΚΗζ). It is supplied to the compressed data generating unit 4 in addition to the oversampling circuit 1 described above. Error correction coding unit 5 and data memory 6. In the oversampling circuit 1, the compressed data generation unit 4, the error correction coding unit 5, and the printed data memory 6 of the employee property cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs, this clock 8 clocks 8 clocks in synchronization with 0 differentiator 3 A series of oversampling data generated in the oversampling circuit 1 is differentiated at each specimen point, and the absolute chirp is obtained and output to the compressed data generating section 4. The compressed data generating unit 4 is provided by one of the oversampling data provided by the above-mentioned oversampling circuit 1, and will be compared with the position before and after. -32- 484259 Α7 Β7 V. Description of the invention (30) (Please read the precautions on the back before filling this page) The small position is detected as the specimen point. Furthermore, a combination of the amplitude data 値 of each detected sample point and the timing data 表示 indicating the time interval at which each sample point appears is output to the error correction coding section 5. The compressed data generating unit 4 obtains a timing clock indicating the timing of the detection point of each specimen point, and outputs it to the data memory 6. The error correction coding unit 5 detects the changed bit and corrects the digital data on the transmission path or the memory even if an error occurs due to a change in noise or the like. The data is supplied by the compressed data generation unit 4 described above. Information is incorrectly corrected. The data thus obtained is output as compressed data on the transmission path or the data memory 6. The data memory 6 is a recording medium storing compressed data, and the compressed data generated by the error correction encoding section 5 is recorded in accordance with the timing clock from the compressed data generating section 4. In addition, in response to a read request signal R E Q given from the outside, the stored compressed data is read and output. Fig. 14 is a block diagram showing an example of the overall configuration of a spreading device according to the present embodiment that realizes the expansion method corresponding to the compression method described above. The consumer cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs printed as shown in FIG. 14. The deployment device of this embodiment includes: an error correction circuit 11, a clock generator 12, and a timing generator 13, and A D-type flip-flop 14, a development processing unit 15, a D / A converter 16, and a low-pass filter (LPF) 17 are configured. The above error correction circuit 11 inputs the compressed data generated in the compression device of FIG. 13 and uses the error correction symbol attached thereto to detect the bits that have changed on the transmission path or in the memory to correct the error. Paper size applies to China National Standard (CNS) A4 (210 × 297 mm) -33-484259 A7 B7 Printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs 5. Description of the invention (31) 〇The clock generator 1 2 by the reference frequency The input clock 0 L κ generates an 8-times clock 8 CL κ, which is supplied to the timing generator 1 3, the expansion processing unit 15, and the D / Α converter 16. In addition, the timing generator 13 receives the timing data included in the compressed data by the error correction circuit 11, and generates the same indefinite time as that between the specimen point detected on the compression side by the clock 8CLK of the above-mentioned 8-frequency. The interval readout clock is supplied to the error correction circuit 11 and the D-type flip-flop 14. The D-type flip-flop 14 will read the amplitude data contained in the compressed data in accordance with the timing of the readout clock generated by the above-mentioned timing generator 13 by the error correction circuit 11 1 and sequentially read in to hold and output.于 展 处理 部 15. The amplitude data of the input and output sections of the D-type flip-flop 14 described above, that is, the timing of the read-out clock is kept at the amplitude data of the D-type flip-flop 14 and the timing of the read-out clock should be The amplitude data (amplitude data of two consecutive sample points of two consecutive sample points) held in the D-type flip-flop 14 is input to the expansion processing unit 15. The expansion processing unit 15 uses the two amplitude data inputted in this way and the timing data inputted from the timing generator 13 to generate digitally interpolated data between each specimen point by interpolation operations described later. The D / A converter 16 converts the digitally interpolated data thus generated into an analog signal. In addition, 丄 P F 1 7 processes the analog signal output from the D / A converter 16 described above with a low-pass filter, removes noise, and outputs it as a reproduced analog signal. Next, the following detailed description is shown in the compression device of Fig. 13 above. The paper size applies the Chinese National Standard (CNS) A4 specification (210X297 mm) " " " -34-

f Please read the precautions on the back before filling in this page J 484259 A7 B7 V. The structure and operation of the oversampling circuit 1 of the invention description (32). (Please read the precautions on the back before filling in this page.) Before explaining the oversampling method of this embodiment to obtain the interpolation between discrete digital data, first explain the general content of the conventional practice. Insert method. The data interpolation processing for interpolating between discrete digital data to obtain more continuous data is performed as illustrated in FIG. 5 using a sampling function called a s i n C function shown in FIG. 4, for example. This sinc function is defined by S 1 Π (7Γ ft) / (7Γ ft) when the sampling frequency is set to f. Only at the specimen point at t = 0, 1 becomes 1, which is equal to all others at equal intervals. Specimen points (t = ± 1, ± 2, ± 3, ± 4, ...), all of them are zero. In the oversampling circuit 1 of this embodiment, different from the general data interpolation method using such a sinc function, when obtaining the interpolation between two discrete data, each of the oversampling methods should include this 2 The discrete data, the basic waveform of the amplitude of the n discrete data, and the digital data of the basic waveforms are obtained by synthesizing the n data obtained by the convolution operation to digitally complement the two discrete data. Interpolate between. Printed by the Consumer Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs. Figure 15 is an explanatory diagram of the basic digital waveforms used in this embodiment. The digital basic waveform shown in Fig. 15 becomes a basic function of a sampling function used when performing data interpolation by oversampling. This digital basic waveform is created by changing the data frame to -1, 1, 8, 8, 1, 1, 1 every 1 clock (CLK) of the reference frequency. In the following, using FIG. 16 to discrete data corresponding to the digital basic waveform normalized as shown in FIG. 15 (—1, 1, 8, 8, 1, -1) This paper standard is applicable to the China National Standard (CNS) Α4 specifications (210 × 297 mm) -35- 484259 Α7 Β7 V. Description of the invention (33) (Please read the precautions on the back before filling in this page}) / 8, generated by oversampling and convolution operation of η times The case of interpolation is used as an example to explain the principle of data interpolation according to this embodiment. Also, in FIG. 16, for convenience of illustration, although an example of 4 times oversampling is shown, in FIG. 1 In the oversampling circuit 1 of 3, an oversampling of 8 times is actually performed. In FIG. 16, one of the series of numbers shown in the leftmost column is for the original discrete data (1, 1, 8, 8). , 8, 1, 1, 1) / 8 for 4 times oversampling. Also, from left to right, the number in the fourth column will be displayed in the leftmost column. Figure 16 shows the time axis in the direction of the column. Moving the number column to the bottom corresponds to the number of columns displayed in the leftmost column. Delayed. Printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs. That is, the second column from the left is displayed so that the number displayed on the leftmost column is shifted by only 4 times the frequency of the clock. 4 CL 1 / 4 The number of columns of the phase. The number of columns in the third column from the left is displayed so that the number of columns displayed in the second column from the left is shifted by only 4 times the frequency of the clock 4 CL Κ 1 / 4 The number of columns of the phase, the fourth column from the left is displayed so that the number of columns displayed in the third column from the left is shifted by 4 times the frequency of 4 CL κ. The number of phases is the number of columns. In addition, the number of columns in the fifth column from the left is a division of the number of columns in the first to fourth columns by 4 in the corresponding row. The processing up to 5 columns is performed digitally with 4 times oversampling with the 4-phase convolution operation. From the 5th column to the right, the number of columns is 4 (from the left 5 to 8 columns) The number displayed in the fifth column is one, one is shifted to the lower paper scale. The Chinese National Standard (CNS) A4 specification (210 × 297 mm) -36- 484259 A7 B7 V. Description of the invention (34) ( First read the notes on the back before filling out this page). Also, the 9th column from the left is the division of the 5th to 8th columns by 4 in the corresponding row. By this processing from the left to the ninth column, four times of oversampling digits with four-phase convolution operation are performed twice. Furthermore, the tenth column from the left is displayed. The number 9 in the 9th column is shifted downward. In addition, the number 11 in the 11th column (the rightmost column) from the left is the row corresponding to the number 9 in the 9th column and The number in column 10 is the sum of 2 divided by 2. This rightmost sequence becomes the interpolation of the destination. FIG. 17 is a graph showing the finally obtained number of curves shown in the rightmost column of FIG. 16. The function with the waveform shown in Fig. 17 can be differentiated only once in the whole area. When the specimen position t along the horizontal axis is from 1 to 3 3, it has a finite 値 other than 0. In other regions, 値 becomes a function of 0. In addition, a case where the function 具有 has a finite value other than 0 in a local area, and a value other than 0 is called "finite". Printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs, the function in Figure 17 has a maximum value obtained only at the specimen point at t = 17, and at four specimen points at t = 1, 9, 2, 5, 3,値 becomes a sampling function with the characteristics of 0, and all the necessary sample points are passed to obtain the data of the smooth waveform. In this way, the function shown in Fig. 17 is a sampling function, differentiable once in the whole area, and at the specimen position t = 1, 3 3, it is a function with limited convergence. Therefore, by using the sampling function of Fig. 17 to superimpose the discrete data, the function can be interpolated from the paper size using the 1-time differential function. The Chinese National Standard (CNS) A4 specification (210X297 mm) is applicable. -37- 484259 A7 B7 Printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs. It is known that the sinc function used is a function that converges to 0 at the sample point of t = ± 〇〇. To correctly obtain the interpolation 値, it is necessary to calculate the interpolation position corresponding to each discrete data up to t = 00. One of the sinc functions, use this to perform convolution operations. In contrast, the sampling function of FIG. 17 used in this embodiment converges to 0 at the sample points of t 2 1 and 3 3, so only the discrete data within the range of t = 1 to 3 3 can be considered. . Therefore, in the case of obtaining a certain interpolation, it is only necessary to consider only a limited number of n discrete data, and the processing amount can be greatly reduced. Moreover, the discrete data outside the range of t = 1 to 3 3 should be considered. The consideration of processing volume or accuracy is not ignored, but it is necessary for theoretical considerations, and no resection error occurs. Therefore, if the data interpolation method of this embodiment is used, the correct interpolation can be obtained. When the compression processing is performed using this correct interpolation, the original data before compression can be improved on the data that is reproduced on the unfolding side. Reproducibility of the data. Fig. 18 is a block diagram showing a configuration example of the oversampling circuit 1 shown in Fig. 13 above. As shown in FIG. 18, the oversampling circuit 1 of this embodiment includes a normalized data storage unit 21, a phase shift unit 22, and a complex digital multiplier 2 3a to 2 3d. And a complex digital adder 2 4 a to 2 4 c. The P L L circuit 2 shown in FIG. 18 is the same as that shown in FIG. 13. The normalized data storage unit 21 described above stores the normalized data rows 4 in a staggered manner as shown in the rightmost column of FIG. 16. Right, although Figure 16 shows the basic digital waveforms shown in Figure 15 and performs 4 (Please read the precautions on the back before filling this page). The size of the paper is applicable to the Chinese National Standard (CNS) A4. Specifications (210X297 public holidays) -38- 484259 A7 B7 __ V. Description of the invention (36) (Please read the precautions on the back before filling this page) Example of oversampling, but the oversampling circuit in Figure 1 3 In order to perform oversampling by 8 times, the digital basic waveform is oversampled by 8 times, and the data row normalized by volume calculation is stored in this normalized data storage unit 21. The four-phase normalized data stored in this normalized data storage unit 2 is read out in accordance with the clocks CLK and 8 CLK supplied from the PLL circuit 2 and supplied to four digital multipliers 2 3 a ~ 2, respectively. 3 d to ~ square input terminals. Further, the phase shifting unit 22 performs a phase shifting process of shifting the phase of the scattered data input as a compression target to four phases. The four-phase discrete data generated by the phase shifting section 22 is outputted in accordance with the clocks CLK and 8 CLK supplied through the PLL circuit 2 and supplied to four digital multipliers 2 3 a ~ 2, respectively. 3 d The other input terminal. The Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs printed the above-mentioned four digital multipliers 2 3 a to 2 3 d. The four-phase normalized data output from the above-mentioned normalized data storage section 21 and the phase shifting by the above-mentioned The part 22 is multiplied by the discrete data of 4 phases. The three digital adders 2 4 a to 2 4 c connected to the subsequent stages add all the multiplication results of the four digital multipliers 2 3 a to 2 3 d, and output the addition result in The differentiator 3 and the compressed data generating section 4 in FIG. 13. In the configuration of the oversampling circuit 1 shown in Fig. 18 described above, the memory means of the present invention is constituted by the normalized data storage section 21. The phase shifting unit 2 2, the digital multipliers 2 3 a to 2 3 d, and the digital adders 2 4 a to 2 4 c constitute the synthesis means of the present invention. It can be clearly understood from the structure shown in FIG. 18 that the paper size in this embodiment applies the Chinese National Standard (CNS) A4 specification (210X297 mm) 39-484259 A7 _ B7 V. Description of the invention (37) (please first Read the notes on the back and fill in this page), and store the normalized data in the rightmost column obtained by the convolution operation shown in Figure 16 in the normalized data storage unit such as ROM in advance 2 1 . In addition, this normalized data is modulated into discrete data that is input as a compression target, and the data obtained from this is synthesized and output by a 4-phase convolution operation. Although it is also possible to multiply the amplitude of the discrete data input as the object of compression by the basic digital waveforms shown in Figure 15 and the data obtained by this, the volume shown in Figure 16 is compressed during compression. Product operation, but in the case where the oversampling circuit 1 is configured as shown in FIG. 18, the convolution operation of FIG. 16 does not need to be performed during actual compression, and it has the advantage that the compression processing can be speeded up. Fig. 19 is a diagram showing an example of digital data inputted into the above-mentioned oversampling circuit 1, and Fig. 20 is a diagram showing the output data after the data is processed by the oversampling circuit 1 for this digital data. From this, it is clear that by using the oversampling circuit 1 of this embodiment, it is possible to obtain continuous oversampling data that smoothly changes from the original discrete digital data. Printed by the Consumer Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs

Next, the following describes in detail the processing of generating the compressed data by the differentiator 13 and the compressed data generating section 4 of FIG. 13 for the digital data thus oversampled, and by the timing generator 1 of FIG. 14 , D-type flip-flop 14 and expansion processing unit 15 to expand the compressed data processing. First, the compression processing will be explained. As described above, the differentiator 13 in FIG. 13 differentiates at each specimen point a series of oversampling data generated in the oversampling circuit 1 to obtain its absolute value, and outputs it to the compressed data generation department. The paper size is applicable to China. Standard (CNS) A4 specification (210 X 297 mm) -40- 484259 A7 B7 ___ 5. Description of the invention (38) 4. (Please read the precautions on the back before filling out this page.) Figure 2 shows a configuration example of the differentiator 3. As shown in FIG. 21, the differentiator 3 of this embodiment type is constituted by calculating a differential absolute 値 difference absolute 差分 circuit between data of two consecutive sample points. In FIG. 21, the differentiators 3 1 and 3 2 respectively calculate the difference of the data of two consecutive sample points inputted by the nodes a and b. That is, the differentiator 31 calculates the difference a-b, and the differentiator 3 2 calculates the difference b-a, and outputs the results to the OR circuits 3 3 and 34, respectively. When these differential units 3 1 and 3 2 become negative, the difference 値 is output as a borrowed value of ~ 1 値 out of the difference 値. Printed by the Consumer Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs, the OR circuit 3 3 obtains the logical sum of the difference between the difference calculated by the differentiator 31 and the borrow output, and outputs the result to the A N D circuit .35. The other OR circuit 34 obtains a logical sum of the difference 运算 calculated by the differentiator 32 and the borrow output, and outputs the result to the AND circuit 35. The AND circuit 35 obtains a logical product of the outputs from the two OR circuits 3 3 and 34 and outputs the result to the node c. The borrow output of the differentiator 31 is output to a node d, and the difference 値 calculated by the differentiator 32 is output to a node e. As a result, the absolute difference aI a-b | of the data of two consecutive specimen points is output at node c, and when the data at node b is larger than the data at node a, '' 1 〃 之 値 is output at The difference 値 b-a between the data of node d and nodes a and b is output to node e. In addition, in FIG. 21, for the convenience of explanation, although only the data lines of nodes a, b, c, and e are displayed in one bit, in fact, only the data bits I of the data are applicable to the Chinese National Standard (CNS) A4. Specifications (210X297 mm) -41-484259 A7 B7 __ 5. Description of the invention (39) Number. (Please read the precautions on the back before filling in this page.) In addition, the compressed data generating unit 4 in FIG. 13 uses a series of oversampling data supplied through the oversampling circuit 1 to The position where the differential absolute value becomes smaller is detected as the specimen point. Then, a combination of the amplitude data 各 of each detected sample point and the timing data 表示 indicating the time interval at which each sample point appears is output to the error correction coding section 5. In the present embodiment, the differential absolute value 在 calculated by the differentiator 3 is used to remove the difference calculated in the differentiator 3 in order to provide a margin for the determination of whether or not the sample point is detected when detecting the sample point. Determined by the absolute lower order bit. For example, if the lower 1 bit is removed, the absolute difference 实际 actually calculated is in the range of 0 to 1, and all the absolute differences 全部 can be judged as 0. In addition, if the lower 2 bits are removed, the absolute difference actually calculated is in the range of 0 to 3, and all the absolute differences can be judged as 0. By doing so, it is possible to avoid the influence of small changes such as noise, and not to detect unnecessary points as specimen points, thereby improving the compression ratio. Printed by the Employees' Cooperative of Intellectual Property Bureau of the Ministry of Economic Affairs. Figure 22 shows the result of calculating the absolute value of the differential value using the differentiator 3 for the output of the oversampling circuit 1 shown in Figure 20 above. As described above, in the compressed data generating unit 4, according to the output data of the differential absolute value shown in FIG. 22, the position where the differential absolute value becomes smaller than the position before and after, that is, the absolute absolute value of the differential absolute value is extremely small. The position where 値 appears (the position shown by the arrow in Figure 2 2) is detected as the specimen point. In addition, the first and last occurrences of the minima can also be considered as data, which is not used as the specimen point. This paper size is applicable to Chinese National Standard (CNS) A4 specification (210X297 mm) -42- 484259 A7 B7 -------- 5. Description of invention (40) (Please read the precautions on the back before filling this page) However, in order to provide a margin for the detection of the specimen point, the case of removing the lower digits of the absolute absolute value may be the same as the minimum value. At this time, according to the borrowed 値 of the differentiator 31 which is output to the node d of the differentiator 3 shown in FIG. 21, the polarity of the positive and negative of the differential 値 is judged, and the polarity change side of the differential 値 is detected as the specimen point. In addition, when the polarity of the differential 値 does not change, as shown in FIG. 23, the relationship between the differential absolute 値 of the sample points A and D before and after the two consecutive sample points B and C of the same 値 is shown.点 The point near the small side is detected as the specimen point. In the example in Fig. 23, the differential absolute value 标 of specimen point D is smaller than the absolute absolute difference 点 of specimen point A, so the specimen point C close to the specimen point D is detected as the specimen point. In addition, as a process for detecting the specimen point, once the data supplied by the oversampling circuit 1 is differentiated, the differential absolute value obtained by the differentiation is further double-differentiated, and the polarity of the double-differential value is changed from negative or zero. It is also possible to extract the point just before as the specimen point. Furthermore, among the points extracted based on the polarity of the double differential ridge in this way, only the point where the absolute differential 値 is smaller than a constant 1 once is regarded as a regular specimen point detection process. Printed by the Consumers' Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs, that is, the polarity of the absolute absolute value of the differential absolute value obtained by the first differential, and the differential of the double differential absolute value of the differential absolute value of the first differential absolute value must change from negative to positive. Therefore, the double differential 値 of the oversampled data is obtained, and the point where the polarity changes from negative to positive (including the point where the double differential 値 is zero) can be detected correctly. At the same time, the minimum point of the absolute 値 of the differential can be detected at the same time. In the case of two consecutive minimal points, one of them can be detected as a specimen point. This paper size applies to China National Standard (CNS) A4 (210X297 mm) -43- 484259 A7 B7 V. Description of the invention (41) (Please read the precautions on the back before filling this page) Also, if only The point where the differential absolute value is smaller than a certain value is regarded as a regular specimen point detection, and unnecessary points are not detected as specimen points, which can improve the compression rate. Fig. 24 is a block diagram showing an example of the configuration for detecting the specimen point by double differential detection as described above, and shows an example of the configuration of the differentiator 3 and the compressed data generating section 4 in Fig. 13. As shown in Fig. 24, the differentiator 3 includes a first differentiator 41, a rounding arithmetic unit 42, and a second differentiator 43. The compressed data generating unit 4 includes a polarity change point detecting unit 44, a critical threshold processing J4 5, and a data generating unit 46. Printed by the Consumer Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs The above-mentioned first differential section 41 differentiates the oversampling data supplied by the oversampling circuit 1 of FIG. 13 at each specimen point to obtain its absolute chirp output. The rounding operation unit 42 performs a process of removing the number of digits below the absolute differential value of the first differentiation performed by the first differentiation unit 41. For example, by dividing the above-mentioned first differential absolute value by 8 and removing the lower 3 bits, it is possible to remove the image due to small changes such as noise. The data outputted by the rounding calculation unit 42 is supplied to the second differentiation unit 43 and the critical unit processing unit 45 in the compressed data generation unit 4. The second differentiation section 4 3 further differentiates the absolute absolute value of the first differentiation of the rounding operation by the rounding operation section 4 2 at each sample point. The double differential 値 obtained by the second differential section 43 and the borrowing 表示 indicating its polarity are supplied to the polarity change point detecting section 44 in the compressed data generating section 4. The above-mentioned polarity change point detection section 44 will change the polarity of the double differential supplied by the second derivative section 4 3 in the differentiator 3 from negative to positive. Before the paper size, the Chinese national standard (CNS) A4 specification (210X297) was applied. Mm) -44-484259 A7 B7 __ V. The point of the invention description (42) (please read the precautions on the back before filling this page), for example, the double differential with negative polarity and the value obtained continuously becomes negative The point, or the point where the double differentiation becomes zero, is extracted as a candidate for the specimen point. It is also possible to use the negative differential double differential 値 as a candidate for the sample point and extract it if it is not continuously obtained. The threshold value processing unit 45 compares the first differential absolute value supplied by the rounding calculation unit 4 2 among the candidates of the sample points extracted by the above-mentioned polarity change point detection unit 4 4 and the threshold value set in advance. Alas, only the points where the differential absolute value 値 is smaller than the critical value once are regarded as regular specimen point detections and transmitted to the data generation unit 46. The data generating unit 46 obtains the amplitude data of each sample point thus detected from the oversampled data supplied by the oversampling circuit 1, and generates it using the clock 8 CLK which is 8 times the frequency supplied by the PLL circuit 2. . Shows the timing information of the time interval at which each specimen point appears. The combination of these amplitude data and timing data is output as compressed data. Printed by the Consumer Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs. In the example shown in FIG. 24, although critical differentiation processing is performed using the absolute differentiation of the 1st rounding operation performed by the rounding operation unit 4 2, it is also possible to perform critical processing. The critical unitary processing is performed by using the absolute differential unitary 1 time before the first differential unit 41 is obtained and subjected to the rounding operation. In the above-mentioned FIG. 24, although the point before the polarity of the double differential 値 changes from negative to positive is extracted as the specimen point, the point after the negative change to positive may be extracted as the specimen point. In addition, the critical 値 processing shown in Fig. 24 is not limited to the case of performing double differentiation to detect the specimen point, and it can also be applied to the case of detecting the specimen point by only one differentiation. Figure 2 5 shows the size of the compressed paper output according to the example shown in Figure 2 2 above. Applicable to China National Standard (CNS) A4 (210X297 mm) '" " -45-484259 A7 B7 V. Description of the invention (43) (Please read the notes on the back before filling this page) Data (a combination of amplitude data 値 and timing data 各 of each specimen point). The compressed data output in this way, for example, determination data (2 bits) for determining the type of signal is added, and an error correction symbol is also added. After that, compression processing such as variable length encoding may be performed. The principle of compression and expansion according to the second embodiment are also shown in Figs. 1 and 2. However, in the case of this embodiment, the waveform data shown in FIG. 1 corresponds to the oversampling data that is continuously processed by the data interpolation processing in the above-mentioned oversampling circuit 1. In addition, the data of the waveform shown in FIG. 2 is equivalent to the data obtained in the process of regenerating the original oversampling data by expansion processing. First, the compression processing will be described using FIG. 1. In the present embodiment, from the inputted oversampling data 101, the absolute value of the detection differential becomes extremely small, and the specimen points 1 02a to 102f are extremely small. Then, obtain the digital data 振幅 of the amplitude of each of the specimen points 1 0 2 a to 1 〇 2 f and the timing data indicating the time interval at which each specimen point 1 0 2 a to 1 〇 2 f appears, and this amplitude 値The combination with the timing data 値 is output as compressed data in the error correction coding section 5 of FIG. 13. Printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs Next, using FIG. 2, the unfolding processing of the data compressed as shown in FIG. 1 will be described. When the input data 1 0 1 is compressed by the compression method in FIG. 1, the obtained compressed data becomes (*, 7) (5, 3) (7, 9) (3, 1) (3, 6) (3, 3 ). In addition, * in FIG. 1 shows those who are not shown. In accordance with the sequence shown here, compressed data is input to the timing generator 13 and the D-type flip-flop 14 on the expansion side. Jade paper scale is applicable to Chinese National Standard (CNS) A4 specification (210X297 mm) -46- 484259 A7 B7 V. Description of the invention (44) (Please read the precautions on the back before filling this page) Expand processing shown in Figure 1 4 In Section 15, first, the amplitude data 値, 7 〃, and timing data 値 '' 5 最初 which are initially output by the error correction circuit 1 1 are used to generate the waveform a 1 through interpolation. Information. Next, from the above-mentioned timing data 値, 5 〃, and two pieces of data 接着 '' 3 接着 which are then inputted, the data of waveform a 2 is generated by interpolation. Next, from the above-mentioned amplitude data 値, 3 〃, and two pieces of data 値 7 个 which are then inputted, the data of waveform b 2 is generated by interpolation. Further, from the above-mentioned timing data 7, '7 〃, and the amplitude data 値 9 输入 which is input next, the data of waveform b 1 is generated by interpolation. The following is the same. The waveform c1, c2, d2, d1, d1, e2 are sequentially generated from the combination of the amplitude data 値 and the timing data 输入 which are sequentially input. Printed by the Consumer Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs through the above process, digital data (a upper part of Fig. 2) and waveforms a 2 and b where waveforms a 1, b 1, c 1, d 1 and e 1 are continuously generated are generated. 2, c 2, d 2, e 2 is the continuous digital data (lower part of Figure 2). Furthermore, the expansion processing unit 15 adds the two digital data generated in this manner to each other, and outputs the digital data to the D / A converter 16 to perform digital-analog conversion to reproduce the oversampling data 1 shown in FIG. 1. 1 analog signal. As the synthesis, the digital data displayed in the upper stage of FIG. 2 and the digital data displayed in the lower stage are processed for data interpolation. For example, the process described in FIG. 6 may be applied. Also, in the second embodiment, 8 times oversampling is performed during compression. The paper size applies the Chinese National Standard (CNS) A4 specification (210X297 mm) 47-484259 A7 B7 V. Description of the invention (45) (please first (Please read the notes on the back and fill in this page). Therefore, the timing information of the time interval between specimen points is not odd and it often becomes even (corresponding to 2 T in the horizontal axis of Figure 6). That is, the timing data of the five pieces shown in FIG. 1 '' 5, 7, 3, 3, 3 〃 are actually oversampled by 8 times, and 4 0, 5 6, 2 4, 2 4, 2 4 6値 is transmitted or stored. In the case where the timing data 奇 T is an odd number, there is a state where there is no interpolation position t at exactly the intermediate time point between the specimen points, and the timing data 値 is even or odd and needs to be processed separately according to the situation. However, in the case of the present application form, the timing data indicating the time interval between specimen points is often an even number. Therefore, it is not necessary to perform complicated processing that requires separate processing due to even or odd numbers. Next, other processing examples of the above-mentioned data interpolation will be described below. Here, the method of using the amplitude data 。D1, D2 of the continuous sample points and the timing data 値 T between them to calculate the interpolation 値 by performing double oversampling and convolution operation is explained. Figure 2-6 shows the case where this oversampling and convolution operation is applicable for the first two specimen points of the compressed data shown in Figure 25 (Dl = 29 · 5, D2 = 24 · 4, T = 6 Printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs. In Fig. 26 (a), the number of 2 9 · 5 左 in the leftmost column is the 値 of the amplitude data D 1 in the first column. The numbers `` 2 9 · 5 〃 and 、, 〇 倂 are arranged in the vertical direction. 1 2 shows the above-mentioned amplitude data — D 1 —plane 1 clock 1 clock is delayed in order, and one side is maintained The state of d-type flip-flops and the like not shown in the timing data 値 T = 6 is twice the number of parts. In addition, the numbers in the second to sixth columns from the left are in accordance with the Chinese standard (CNS) A4 specification (210X297 mm). -48- 484259 Printed by A7, B7, Employee Cooperatives, Intellectual Property Bureau, Ministry of Economic Affairs Explanation of the invention (46) is the result of sequentially shifting the numbers 1 clock 1 clock displayed on the first column. In addition, the number column of the seventh column shows the result of dividing the number columns of the first to sixth columns by 6 in the corresponding row. That is, the number columns of the first to sixth columns are divided into six phases. The result of the convolution operation. Each of the eighth to twelfth columns is a number that is a result of a convolution operation performed on the seventh column, and the clock is further shifted by one clock and one clock. The numbers and columns of the 13th and 12th columns show the result of dividing the number and columns of the 7th and 12th columns by 6 in the corresponding rows. That is, the number and columns of the 7th and 12th columns are divided. The result of a 6-phase convolution operation. The number 13 of the rightmost column 13 is the interpolation curve to be obtained (in the example of FIG. 2, the sampling function a 1). The sequence is the same as that shown in Fig. 26 (a), using the same timing data 値 T (two 6) and the other amplitude data 値 D 2 (= 2 4 · 4) to calculate the same as the above sampling function a 1 Interval sampling function a 2. Furthermore, by obtaining the sum of the respective calculation results, an interpolation curve finally obtained in this interval is obtained as shown in FIG. 26 (b). This kind of data interpolation processing is performed sequentially between the specimen points across the whole frame to reproduce the original oversampled data. In addition, the operation shown in FIG. 2 (a) can be implemented by appropriately combining the hardware configuration of a complex D-type flip-flop, an adder, and a multiplier while holding the data 値 1 clock 1 clock delayed. achieve. As another processing example of the data interpolation operation, the method described in Fig. 7 may be used. Figure 2 7 shows the oversampled capital paper size reproduced from the compressed data shown in Figure 25. The Chinese paper standard (CNS) A4 specification (210X297 mm) is applicable to '-49-(Please read the precautions on the back before filling in this Page) 1 ^^.

、 1T 484259 A7 _ _ B7 _ V. Description of the invention (47) Material drawing. (Please read the precautions on the back before filling this page.) If you compare the oversampling data obtained by the expansion shown in Figure 27 and the oversampling data before compression shown in Figure 20, you can borrow it clearly. The unfolding processing of this embodiment can reproduce data almost equal to the original oversampling data. In the expansion device shown in FIG. 14, although the oversampled data thus reproduced is input to the D / A converter 16 for digital-analog conversion, the digital data before the D / A conversion is shown in FIG. 27. Generally, a smooth signal that has been continuous. Therefore, it is not necessary to use a digital filter to virtually increase the sampling frequency like a conventional D / A converter. Even if only D / A conversion is performed, the quality of the analog signal being output can be greatly improved. Printed by the Consumer Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs as detailed above. In the second embodiment, the input discrete digital data is oversampled and convolved on the compression side to produce smooth continuous data. The oversampling data obtained by sampling with its differential absolute 値 becoming the smallest indefinite time interval, as the compressed data can obtain discrete amplitude data 値 and timing data indicating their indefinite time interval 値. Further, in the unfolding side, the amplitude data 定时 and the timing data 包含 included in the compressed data are read, and the discrete data is read out at the same indefinite time interval as the compression side, and the output is continued by interpolation between them. data of. Therefore, when compressing and expanding analog signals on the time axis, processing can be performed on the time axis without frequency conversion. Therefore, the processing of compression and expansion is not complicated, and the structure for this can be simplified. This paper size applies Chinese National Standard (CNS) A4 specification (210X297 mm). Sun exposure 50_ 484259 A7 B7 __ V. Description of the invention (48) (Please read the precautions on the back before filling this page). It is also sent by the compression side. When compressed data is reproduced on the expansion side, it is possible to process the compressed data input on the expansion side sequentially and reproduce it by simple interpolation operation on the time axis, so that real-time action can be realized. Also, in this embodiment, the absolute absolute value of the digital data becomes extremely small as the specimen point detection, and the amplitude data of each detected specimen point and the timing data indicating the time interval at which each specimen point appears are compressed. The reason why the data is transmitted or recorded can mean that the data of the specimen point is obtained as compressed data, which can achieve a high compression rate. In addition, according to this embodiment, the turning points existing in the signal of the compression target are detected as the specimen points, and the minimum points necessary to reproduce the original data by interpolation operations on the expansion side are all included in the compressed data. Therefore, the reproducibility of the original data can be improved, and high-quality reproduced data can be obtained. Printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economy Because the sampling function can only be differentiated once, it is only necessary to consider only a limited number of discrete data in the case of obtaining a certain interpolation, which can greatly reduce the amount of processing. In addition, there is no rounding error, and correct interpolation can be obtained. When this interpolation is used for compression processing, the data that is reproduced on the expansion side can improve the original data before compression. Reproducibility. The convolution operation shown in the second embodiment is only an example, and the present invention is not limited to this. Moreover, in the second embodiment, although the digital basic waveform is set to -1, the paper size applies the Chinese National Standard (CNS) A4 specification (210X 297 mm) _ 51-484259 A7 __ B7 __ V. Description of the invention (49) The basic digital waveforms are not limited to this example (please read the precautions on the back before filling this page). In other words, the obtained interpolation function can be differentiated only once in the entire domain, and in a limited sample position, any waveform can be used as long as it is a finite function that converges to zero. For example, the weight of the parts corresponding to both sides may be not 1 but 1 or 0. It is also possible to make the weight equivalent to the middle part to a weight other than 8. Either way, good curve interpolation can be achieved. The interpolation operation performed by the expansion processing unit 15 in Fig. 14 may also perform a convolution operation as shown in Fig. 16 based on the digital basic waveform shown in Fig. 15. In this case, the continuous interpolation can be obtained only by the digital processing of the convolution operation, and the result of the D / A conversion becomes a smooth analog signal. Thereby, it is possible to omit L P F 1 7, and there is an advantage that deterioration in phase characteristics due to the filter can be suppressed. (Third Embodiment) A third embodiment of the present invention will be described below with reference to the drawings. Printed by the Consumers' Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs. The above-mentioned first and second embodiments are processed on the time axis and the variable clock length interpolation method is used. On the other hand, in the third embodiment described below, compression and expansion processing can be performed more easily without using a table. In the third embodiment, first, in the case where an analog signal is input as a signal to be compressed, the analog signal inputted by the A / D conversion is digital data. Furthermore, the digital data converted by A / D is processed using the first and second frames. The first frame and the second frame may be the same frame, but the second frame is preferably larger than the first frame. This paper size is applicable to China National Standard (CNS) A4 (210X297 mm) " -52- 484259 Α7 Β7 V. Description of the invention (50) (Please read the precautions on the back before filling this page) The first rounded digital data is differentiated once at each specimen point, and the point at which the polarity of the differentiated ridge is changed is detected as the specimen point. In addition, as the compressed amplitude data of each specimen point detected, while obtaining the second rounded digital data, the timing data indicating the time interval at which each specimen point appears is obtained. Furthermore, the difference data between the obtained compressed amplitude data is obtained, and the combination of the compressed amplitude difference data and the timing data is transmitted or recorded as compressed data. On the other hand, in the unfolding side of the compressed data generated as described above, the compressed amplitude difference data in the compressed data (combination of the compressed amplitude difference data and the timing data) is oversampled by a clock of an even multiple frequency. Moreover, the sign of this oversampled compressed amplitude difference data is reversed at the middle position of each interval of each specimen point represented by the timing data, and the data obtained by this is listed at each specimen point. After performing multiple integrals in the interval, a moving average operation or a convolution operation is performed. Printed by the Employees' Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs. With this, smooth waveform compression amplitude data can be obtained in each section of each specimen point. Then, using the thus obtained compressed amplitude data and the above-mentioned timing data, on the compression side, smoothing is performed by performing an interpolation operation including a process of multiplying the multiples of the number of bits that are rounded by the second unit to be rounded. The amplitude data of each consecutive interval is interpolated with each other. Furthermore, in response to the need to generate interpolation data for D / A conversion, analog signals are supplemented. Fig. 28 is a block diagram showing an example of the overall configuration of a compression device that implements the third embodiment of the compression method described above. In this figure 2-8, for example, it is input as the compression target to apply the Chinese National Standard (CNS) A4 specification (210 × 297 mm) at this paper scale -53- 484259 A7 B7 V. Description of the invention (51) (Please read the back Please fill in this page for the matters needing attention) 4 4 · 1 Κ Η z The sampling frequency (set this as the reference frequency) of the digital data sampled. The digital data input here is, for example, 16-bit signed digital data. In the following, as an example of digital data, a case where a sound signal is compressed will be described. Here, although the digital data is directly input as the data to be compressed, an analog signal may also be input. In this case, the input section of the compression device is provided with an L P F or A / D converter, for example. That is, the input analog signal is easily converted to digital data by the A / D converter after the noise is removed by the L P F in order to facilitate the detection of the specimen point. As shown in FIG. 28, the compression device of this embodiment includes: a timing generator 1 1 1 and an amplitude generator 1 1 2; a rounding operation unit 1 1 3; and a difference operation unit 1 1 4; And encoder 1 1 5 and data memory 1 1 6 (optional). Printed by the Consumer Cooperative of the Intellectual Property Bureau of the Ministry of Economics. The timing generator 1 1 1 differentiates the input digital data at each specimen point, and detects the specimen point according to the change in the polarity of the differential 値. Furthermore, a timing pulse T P indicating the timing of the detection point and timing data (the number of clocks C K of the reference frequency) indicating the time interval of each specimen point are obtained and output respectively. The timing generator 1 1 1 also generates and outputs various clocks including the read clock of the data memory 1 1 6. In addition, the amplitude generator 1 1 2 follows the clock of the reference frequency (: 1 ^, from the digital data of each sample point that is sampled and input, only taking out is equivalent to the above by the timing generator 1 1 1 The output digital data of the timing sample point position displayed by the output timing pulse TP will be output as the amplitude data of each sample point. This paper size applies the Chinese National Standard (CNS) A4 specification (210X297 mm) -54-484259 A7 _ B7 __ V. Description of the invention (52) Figure 2 9 is a diagram illustrating the operation principle of the above-mentioned timing generator 1 1 1 and amplitude generator (please read the precautions on the back before filling this page). In addition, although the data input to the timing generator 1 1 1 and the amplitude generator 1 1 2 are coefficient-bit data, the waveforms of the digital data are analogously displayed in FIG. 29 for the sake of explanation. In this embodiment, the Among the digital data 501 that is input as the compression target, the point at which the polarity of the differential 变化 changes and the point at which the differential 零 is zero 50 2 a to 5 0 2 f are detected as specimen points. Further, these specimens are obtained. Point 5 0 2 a ~ 5 0 2 f The frame data and timing data indicating the time interval between the occurrence of each specimen point 5 0 2 a ~ 5 0 2 f are output in the next paragraph. In the example of Figure 29, as the specimen point 5 0 2 3 ~ 5021 Digital amplitude data 値, find "D 〇, D 1, D 2, D 3, D 4, D 5, as the time t 0-t 1 and t 1 between the appearance of the specimen points 502 a ~ 502 f T 2-t 2-t 3-t 3-t 4-t 4-t 5 at different time intervals, find "T 1, T 2, D 3, T 4, T 5 ". The consumer cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs printed the time point at time t 〇, and can obtain the amplitude data of the specimen point 5 〇 2 a 値" D 〇〃 and the specimen point before it (not shown) Therefore, the timing data 时间 (not shown) of the time interval from the time when the test is performed, the combination of these data 値 is output as the data at time t 0. Then, at the time t 1 when the specimen point 5 0 2 b is detected At the time point, timing data 可以获得 τ 1 表示 representing the time interval from the time t0 at which the previous specimen point 5 0 2 a was detected can be obtained, Specimen point 5 0 2 b amplitude data 値, D 1 故, the paper size of these data 适用 applies to the Chinese National Standard (CNS) A4 specification (210X297 mm)--55-484259 Ministry of Economic Affairs intellectual property A7 B7___ printed by the bureau's consumer cooperative. V. Invention Description (53) The combination (T1, d1) is output as the data at time t1. Further, at the time point following the time point t 2 at which the specimen point 5 0 2 c is detected, timing data (T 2) indicating the time interval from the time point 1 before the specimen point 5 0 2 b is detected can be obtained. As well as the amplitude data 値 ′ and D 2 标 of the specimen point 5 0 2 c, the combination of these data 値 (T 2, D 2) is output as the data at time t 2. In the same manner, the timing data indicating the time interval between time t 2-t 3, t 3-t 4, t 4-t 5 and the specimens detected at time t 3, t 4, and t 5 The combination of the amplitude data 値 (T 3, D 3), (T 4, D 4), and (T 5, D 5) of the amplitude data 値 at points 502 d, 50 2 e, and 502 f are regarded as time t 3, respectively. Data output at t 4, t 5. Fig. 30 is a block diagram showing a configuration example of the timing generator 1 1 1 described above. In FIG. 30, the rounding operation unit 1 1 7 performs rounding operation on the digital data that is input as the compression target data by the first 1N1 (using the division of the 1N1). For example, the rounding operation unit 1 7 performs rounding processing of the input digital data by 8 or 16. The differentiator 1 1 1 differentiates the digital data which has been rounded by the rounding operation section 1 1 7. At this time, the differentiator 1 1 8 is given the clock C 4 at the input of 4 4 · 1 κ Η ζ, that is, the digital data is differentiated at each specimen point according to the reference frequency. The differential 値 is obtained, for example, by subtracting the current data acquired at a timing of a certain input clock CK from the data acquired at a timing of a clock before one in time. In addition, the specimen point detection section 1 1 9 calculates the paper size based on the differentiator 1 1 8 and applies the Chinese National Standard (CNS) A4 specification (210 × 297 mm) (Please read the precautions on the back before filling out this page) -56 -484259 A7 _ B7 V. Differential 値 of the description of the invention (54) (please read the precautions on the back before filling this page), and use the differential point of the differential 値 of digital data as the specimen point detection. For example, the specimen point detection section 1 1 9 detects a point at which the polarity of the differential 値 changes from positive to negative or from negative to positive, and a point at which the differential 値 becomes 0. Moreover, regarding the point where the polarity of the differential 値 changes from positive to negative or from negative to positive, the point before the change of its polarity is detected as the specimen point. In addition, the case where the point at which the differential 在 becomes 0 occurs continuously in two or more, for example, the positions of the two ends are detected as the specimen points. The timing generator 1 2 0 counts the number of clocks CK supplied from the time when one specimen point is detected until the next specimen point is detected, and outputs this as timing data T, and outputs the detection points representing each specimen point. The timing pulse TP. The timing generating unit 120 generates various clocks including the read clocks and outputs them. Printed by the Consumer Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs As described above, in this embodiment, before performing differentiation on the digital data to be detected at the specimen point, the digital data is rounded. Although this rounding operation is not necessarily performed, it is preferably performed. That is, if the differentiation is not performed under the condition of the original data without the rounding operation, there may be cases where small noise components or unwanted signal components contained in the original data are also detected as specimen points, and the compression rate is reduced. . Therefore, it is better to perform differentiation after performing the rounding operation. However, if the first 値 N 1 for the rounding operation is too large, the turning point (peak point) of the original data that originally changed the polarity of the differential 値 is smoothed, and it is considered that the necessary pivot point cannot be detected as the specimen point. In this case, there is a case where the correct data cannot be reproduced on the development side. Therefore, the first 値 N 1 needs to be selected not too small and not too large (with the first 値 N 1 = 8 this paper size applies the Chinese National Standard (CNS) A4 specification (210X297 mm) -57- 484259 A7 B7 5. Invention description (55) or 16 is preferred). (Please read the precautions on the back before filling this page.) Figure 3 1 shows a detailed configuration example of the part that generates the above-mentioned timing pulse T P. In FIG. 31, the 1D-type flip-flop 1 2 1 samples and holds the digital data input as the compression target according to the clock C K of the reference frequency. The multiplier (or divider) 1 2 2 doubles the digital data 1 / n 1/1 that is held at the first D-type flip-flop 1 2 1. As a result, the digital data multiplied by 1 / N by the multiplier 1 2 2 is supplied to the negative side of the subtractor 1 2 4, and the clock D is delayed by 1 2 with the second D-type flip-flop 1 2 3. After serving, it is supplied to the positive side of the subtractor 1 2 4. As a result, in the subtractor 1 2 4, the differential chirp can be obtained by subtracting the current data acquired at a timing of a certain input clock C K from the data acquired at a timing before one in time. Printed by the Consumer Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs. In this embodiment, when detecting the specimen point, the differential 値 itself is not necessary, as long as you know its polarity. Therefore, only the sign bits of the differential data are input by the subtractor 1 2 4. The sign bit of the differential data output by the subtractor 1 2 4 is supplied to one of the input terminals of the EXNOR circuit 1 2 6, and the 3D flip-flop 1 2 5 is delayed by only 1 clock CK. After that, it is supplied to the other input terminal of the EXNOR circuit 1 26. With this, by the EXN OR circuit 1 2 6, the specimen point at which the polarity of the differential chirp changes is detected, and as the data indicating the detection point thereof, the timing pulse T P is output. Returning to Figure 2 8 to continue the description. The rounding calculation unit 1 1 3 performs rounding operation on the amplitude data outputted by the amplitude generator 1 12 with the second 値 N 2 larger than the first 値 N 1, and outputs the compressed amplitude data. For example, the paper size applies the Chinese National Standard (CNS) A4 specification (210X297 mm Γ-58- 484259 A7 B7 __ V. Description of the invention (56) (Please read the precautions on the back before filling this page) The arithmetic unit 1 1 3 processes the amplitude data of each sample point outputted by the amplitude generator 1 2 4 and the amplitude generator 1 2 4. By the amplitude data of the amplitude of 104, each character can be adjusted. The length of data is reduced by 10 bits. Here, the amount of data can be greatly reduced. Also, the difference calculation unit 1 4 obtains the difference between the compressed amplitude data obtained by the rounding calculation unit 1 13. For example, by borrowing The differential amplitude data is successively obtained by subtracting the compressed amplitude data of a specimen point before the compressed amplitude data of a specimen point in time. As described in the example shown in FIG. 29, the difference calculation section 114 , Becomes the operation D1 / 1024 — D〇 / l〇24, D2 / l〇24 — D1 / 10 24, D 3/1 0 2 4 — D 2/1 0 2 4, ... (divide by 0 2 4 It uses the rounding operation by the rounding calculation unit 11 3). In this way, the difference amplitude and the compression amplitude data before the difference are obtained by calculating the difference. It can make individual data smaller, and reduce the length of data. The Intellectual Property Bureau, Ministry of Economic Affairs, Employee Consumption Cooperative, prints the encoder 1 1 5 block, and the timing data obtained by the timing generator 1 1 1 , And the combination of the compressed amplitude difference data obtained by the difference calculation unit 1 1 4 is used as the serial compressed block data and output to a transmission path or data memory 1 1 not shown. That is, The encoder 1 1 5 performs parallel / serial conversion to combine the compressed amplitude difference data and timing data to block, and outputs a header or various flags before the data block. The header includes, for example: Information such as identification marks, 値 N 1, etc. rounded by the rounding calculation unit 1 1 7. After this header, it is formed by a combination of the initial stage of the compressed amplitude data and the compressed amplitude difference data and timing data. The data blocks are connected to this paper in ascending order. The size of the paper applies to the Chinese National Standard (CNS) A4 (210X297 mm) 59- Printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economy 484259 A7 B7 __ V. Description of the Invention (57) Continue In addition, the rounded 値 N 1 is included in the head in response to the signal to be compressed, so that the rounded 値 N 1 can be changed to a signal suitable for the signal. The data memory 1 1 6 is for storing compressed data. The recording medium is recorded according to the serial compressed block data generated by the encoder 1 15 according to the clock sent by the timing generator 1 1 1 through the encoder 1 15. It should also be recorded by an external source. The given readout clock reads out the stored compressed data and outputs it. Fig. 32 is a diagram illustrating an actual operation example of the compression processing performed by the compression device shown in Fig. 28. In Fig. 32, the elapsed time is displayed from the top to the bottom in the vertical direction. Among the various data columns A to I shown in FIG. 3, the leftmost data column A is the raw data before the compression process is performed. The data of this life are sampled according to the sampling frequency of 4 4 1 κ Η z. The second data column B from the left is the data obtained by rounding the raw data by 16 through the rounding calculation unit 1 1 7 (the multiplier 1 2 2 shown in FIG. 31). . The third data row C is the data obtained by rounding the raw data by the rounding calculation unit 1 13 shown in FIG. The fourth data row D is the data obtained by differentiating the second data row B (the data of 16 whole life data) by the differentiator 1 18 shown in FIG. 30. For example, from the second differential 値, '2 4 〃 is the data using data row B, and it is obtained from 6 9 6 — 6 7 2 ，. The next information can be obtained from ', 7 0 7 — 6 9 6 〃. : Paper size applies to Zhongguan Family Standard (21〇 × 1mm) ^ — _ 60 turns (Please read the precautions on the back before filling this page)

484259 A7 ___B7 __ V. Description of the invention (58) The fifth column E indicates the flag of the point where the polarity of the differential 値 changes from positive to negative or from negative to positive. That is, at the point before the polarity of the differential ， changes, set 、, 1 1, and at other points, set 、, 0 〃. For example, among the differential 値 shown in the data column D of the fourth, in the part where the differential 値 changes from ,, 1 1 为 to -4 7 〃, before the polarity of the differential 値 changes, 1, 1 Set a flag at 1 point, 1 point. In addition, in the part where the differential 値 changes from -1 1 5 为 to 2 0 〃, before the extreme change of the differential 値, the point of '1 1 5 ，, set the flag, 1 〃. This flag, the point at which '1 〃 stands, becomes the specimen point. The sixth data line F is the compressed amplitude data generated by the rounding operation unit 1 1 3 of FIG. 28. Here, in order to make it easier to understand in comparison with the subsequent expansion processing, although the points of the case where the amplitude data is oversampled and compressed at twice the frequency are shown, in fact, only the flag “1” in the data column "There is compression amplitude data for standing specimen points. The same applies to the following data columns G to I. The seventh data row G is the compressed amplitude difference data generated by the difference calculation section 1 1 4 of FIG. 28. For example, the top data 値, 1 〃 are data using continuous specimen points in data row F, and are obtained from 、, 1 〇 1 1 〃, and the next data 値 0 0 〃 is the use of data. The data 连续 of the continuous specimen points in F are obtained from > 1 0-1 0 〃. The eighth data row is the timing data generated by the timing generator 1 in FIG. Here, the number of clocks C K supplied from the time when one specimen point is inspected to the next specimen point is detected is displayed. Here, although the timing data is displayed at each specimen point at 2 times the frequency, in fact, the Chinese National Standard (CNS) A4 specification (210X297 mm) is only applicable to this paper size ^ (Please read the precautions on the back before filling (This page)

Printed by the Consumer Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs of the 1T. -61-484259 A7 B7 V. Description of the invention (59) There is timing information on the specimen point of flag 1. (Please read the precautions on the back before filling this page) The 9th column I is the flag for switching the displayed data. That is, in the case where the compression amplitude data of consecutive specimen points are the same, in order to show the segmentation of the specimen points, the flags 0 and 1, and 1 are set. For example, the compressed amplitude data of the second specimen point and the third specimen point displayed in the data column F are both 、, 10, and although they are the same compressed amplitude data, they are shown as different specimen points. The data flags are set here. Among the above main data rows, the initial 値 1 1 压缩 of the compressed amplitude data displayed in data row F and the combination of the compressed amplitude difference and timing data of each specimen point displayed in data rows G and （(-1 , 2), (0, 1), (—3, 3) ... are block-coded by the encoder 1 1 5 and output as serial compressed block data. It can be clearly understood that, according to the compression device according to this embodiment, the raw data of the compression target displayed in the data row A can be almost compressed into the data displayed at the specimen points of the data rows G and Η Compared with the raw data, the size of the data at each specimen point can be reduced to a very small size. Printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs Figure 3 3 shows an example of the composition of the serial compressed block data in this embodiment. In this embodiment, as described below, the block data is assumed to be variable-length data. Figure 33 (a) shows the block structure of the compressed amplitude difference data. In Figure 3 3 (a), the first bit is the data sign bit (sign bit), which shows the polarity of the compressed amplitude difference data. For example, when the data sign bit is '' 1 〃, it means a negative number, and 0 〃 means a positive number. This paper size applies Chinese National Standard (CNS) A4 specification (210X297 mm) -62- 484259 A7 B7 ___ V. Description of invention (60) (Please read the precautions on the back before filling this page) Also, the second digit During the segmentation period, the number of bits in the compressed amplitude difference data is displayed. For example, when the segmentation flags 、, 1, ，, the compressed amplitude difference data is 2 consecutive bits (3rd to 4th bits), and 0, the compressed amplitude difference data is expected to be 5 consecutive bits (number 3 ~ 7 bits). The meaning is the segmentation flag which shows the segmentation from the next data block. As shown in the data row G of the mouse 32, the compressed amplitude difference data can be represented by almost 2 bits except for the sign bit. Therefore, for most compressed amplitude difference data, the segmentation flag is set to 1 、 and 2 bits are allocated. For compressed amplitude difference data that cannot be expressed with 2 bits, the segmentation flag is set to '' 0. Alas, allocated 5 bits long. Allocating 5 bits at the maximum can represent all the compressed amplitude difference data. On the other hand, Figure 33 (b) shows the block structure of timing data. The block of timing data is the latter following the block of compressed amplitude difference data. In Figure 3 3 (b), the first bit is the segmentation flag, which shows the number of bits in the timing data. For example, when the segmentation flag is `` 1 '', the timing data is 3 consecutive bits (2 to 4 bits), and when `` 0 '', the timing data is 8 consecutive bits (2 to 9 bits) yuan). Printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs As shown in the data column of Fig. 3, the timing data are all positive numbers, and most of them can be represented by 3 bits. Therefore, for most timing data, the segmentation flag is set to `` 1 '', and the length is allocated to 3 bits. For timing data that cannot be expressed in 3 bits, the segmentation flag is set to `` 0, '' and allocated. 8 bits long. If the maximum is 8 bits, all timing data can be expressed. In this way, in the compression device of this embodiment, the generated compressed data is further transmitted or recorded as variable-length block data. By applying the I paper size to the Chinese National Standard (CNS) A4 specification (210 × 297 mm) ~ -63- 484259, printed by the Consumers ’Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs, A7 and B7, the invention description (61) can make the compression ratio more It can be lifted 1.5 times to achieve a more compact compression ratio. For example, in a part of CD music data, a compression ratio of 12 or more can be achieved. Next, an expansion device corresponding to the compression device described above will be described. Fig. 34 is a block diagram showing a configuration example of a deployment device according to this embodiment. As shown in FIG. 34, the deployment device of this embodiment is provided with a PLL (Phase Locked Loop) circuit 1 3 1 and a data memory (option) 1 3 2 and a decoder 1 3 3, and The timing generator 1 3 4 and the square interpolation data generating section 1 3 5 are configured. The PLL circuit 131 generates a clock 2 C κ that is twice the frequency (8 8 · 2 κ Η z) from the input clock CK of the reference frequency (44 · ΙΚΗζ), and supplies it to the timing generator 1 3 4 and the square interpolation data.产. 生 部 1 3 5． The data memory 1 2 2 is a recording medium that stores serial compressed block data transmitted by the compression device. The decoder 1 3 3 decodes the serial compressed block data read from the data memory 1 2 2 by various clocks synchronized with the clock 2 C 2 of the frequency, and extracts the compressed amplitude difference data and timing data. Of combination. Furthermore, while outputting the extracted compressed amplitude difference data to the square interpolation data generating section 1 35, the extracted timing data is output to the timing generator 1 34 and the square interpolation data generating section 1 35. The compressed amplitude difference data is sampled in the square-interpolated data generating unit 135 by following the timing pulse TP to become data synchronized with the period between each specimen point. The timing generator 1 3 4 accepts the timing data supplied by the decoder 1 3 3 and generates the input clock 2 C KK to indicate the paper size of the specimen detected and detected on the compression side. The Chinese national standard (CNS) Α4 specification (210X 297) Mm) (Please read the notes on the back before filling this page) Order-64- 484259 A7 B7 V. Description of the invention (62) (Please read the notes on the back before filling this page) Time interval timing pulse TP. In addition, the timing generator 1 3 4 also generates and outputs various clocks including read clocks to the data memory 1 2 2. The square interpolation data generating unit 1 3 5 uses the compressed amplitude difference data and timing data input by the decoder 1 3 3 to perform a specified square interpolation operation to generate digital interpolation data to complement the sample points. Regarding the details of this square interpolation operation, although described later, the interpolation data generated here is a series of amplitude data which is subject to 2 times oversampling for the original data before being compressed. The digitally interpolated data generated in this way is output as expanded data. In the example shown in FIG. 34, although the digital data is displayed as being expanded, the digital data obtained may be converted into an analog signal and output as required. In this case, for example, the output section of the square-interpolated data generating section 135 includes a D / A converter and L P F. That is, the digitally-interpolated data outputted by the square-interpolated data generating section 135 is converted into an analog signal by a D / A converter, and then is output as a reproduced analog signal through L P F. Printed by the Employees' Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs. Figure 3 5 shows an example of the detailed structure of the above-mentioned square interpolation data generating unit 135. In FIG. 3, the timing data (T) input as part of the compressed data is sequentially given timing pulses τ P, which are respectively maintained at three D-type flip-flops 143, 146, and 149. In addition, a clock 2 C K of twice the frequency is input to the first counter 1 4 1, and the number thereof is sequentially counted. The first comparator 1 4 2 compares the paper size counted by the first counter 1 4 1 to the Chinese National Standard (CNS) A4 specification (210X29? Mm) -65-484259 A7 B7 Employees ’Cooperatives of Intellectual Property Bureau of the Ministry of Economic Affairs Printed 5. Description of the invention (63) The number of clocks 2 CK and timing data kept in D-type flip-flops 1 4 3. When the number of counted clocks 2 C K exceeds 値 of the timing data, a signal (A > B) indicating its purpose is output. The 20R circuit 1 4 8 takes the logical sum of the signal output from the first comparator 1 4 2 and the external start signal, and outputs the result as a timing pulse TP °. In this way, the 2 R0 circuit 1 4 8 The generation of the timing pulse τ p is necessary to reproduce the timing pulse TP only in the expansion device. In this case, if the logical sum of the output signal of the first comparator 142 and the external start signal is obtained by the 20R circuit, the timing pulse TP ° can be obtained, and the second counter 1 4 4 sequentially counts the reference frequency. The number of clocks CK. The second comparator 1 4 5 compares the number of clocks C K counted by the second counter 1 4 4 and the timing data held in the D flip-flop 1 4 6. Moreover, each time the number of counted clocks C K exceeds the timing data, a signal indicating its intention is output. The 10th circuit 1 4 7 obtains the logical sum of the signal output from the second comparator 1 4 5 and the external start signal, and outputs the result to the E X0R circuit 15 1. As described above, the number of clocks CK based on the reference frequency and the number of clocks 2 CK based on 2 times the frequency are compared with the timing data respectively. The exact middle position of the time interval between the two specimen points shown (representing half the number of clocks 2 CK at the time interval) is determined by the 10th circuit 1 4 7 and the coincidence signal (A > B) is output. On the other hand, the compressed amplitude difference data that is input as part of the compressed data is sequentially given timing pulses T P that are kept in the D-size original paper size application. National Standard (CNS) A4 specification (210X297 mm) ) (Please read the precautions on the back before filling this page) -66-484259 A7 _ _ B7 V. Invention Description (64) (Please read the precautions on the back before filling out this page) Inverter 1 5 2. The compressed amplitude difference data held in this D-type flip-flop 1 5 2 follows the clock 2 CK of twice the frequency. After being oversampled by the D-type flip-flop 1 5 3, it is supplied to the first adder 1 5 4. The first adder 1 5 4 performs the compressed amplitude by adding the compressed amplitude difference data supplied from the D-type flip-flop 1 5 3 and the accumulated data held so far in the D-type flip-flop 1 5 5. Points for difference data. At the time of integration, the first adder 15 4 inverts the sign of the compressed amplitude difference data inputted from the D-type flip-flop 15 3 appropriately according to the output signal from the EX0R circuit 151 described above. In addition to the coincidence signal (A > B) outputted by the 10th circuit above, according to the timing pulse TP, it is held at the data sign bit of the D-type flip-flop 1 5 2 (Fig. 3 3 The sign bit) is input to the above-mentioned EX0R circuit 151. Printed by the Consumer Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs. As a result, the data bit of the first adder 1 5 4 in the block data included in the compressed amplitude difference data may be derived from the 10th circuit 1 4 7 When the coincidence signal of the coincidence signal (A > B) is reversed, that is, the sign of the compressed amplitude difference data is reversed at the iu position of each specimen point and the exact middle position between the two specimen points. The first integration 値 of the compressed amplitude difference data obtained by the first adder 1 5 4 is supplied to the second adder 1 5 6. The second adder 1 5 6 adds up the 1st integral of the compressed amplitude difference data supplied from the first adder 1 5 4 and the accumulation so far held in the D-type flip-flop 1 5 7 so far. Data, for the above 1 point, and then points. ¥ The paper size applies the Chinese National Standard (CNS) A4 specification (210X: Z97 mm) -67- 484259 A7 B7 V. Description of the invention (65) The compressed amplitude difference data obtained by the second adder 1 5 6 The second integral is directly input to one of the input terminals of the third adder 1 5 9 and is temporarily held in the D-type flip-flop 1 5 8 and then input to the other of the third adder 1 5 9 Input. The third adder 1 5 9 performs a moving average operation (convolution operation) by adding the second integral 値 of the compressed amplitude difference data and the clockwise 2 CK parts by shifting it by one. Output to multiplier 1 6 0.

The D-type flip-flops 1 5 5 and 1 5 7 constituting the integrator of the first paragraph and the integrator of the second paragraph, and the D-type forward and reverse of the 2 CK components by a clock that shifts the integral 2 only once. Each of the three D-type flip-flops of the inverter 1 8 is reset to zero every time the timing pulse TP is given. As a result, the second integration and moving average calculation of the compressed amplitude difference data is performed individually in each period of the timing pulse τ p (each interval between each specimen point). The Z multiplier 1 6 0 is obtained by the third addition The moving average data 値 obtained by the device 1 5 9 is multiplied by 5 1 2 / T 2 (T is timing data), and the result is output to the fourth adder 16 1. The fourth adder 16 1 adds the data 由此 from the multiplier 16 0 and the initial stage 使 of the compressed amplitude data by the other multiplier 16 2 (in the example of FIG. 3, it is, 1, 1). 1 〃) 1 0 2 4 times the data 値. Furthermore, the data output by the fourth adder 16 i follows the clock 2 C K of 2 times the frequency, and the D flip-flop 1 6 2 is temporarily held, and is output as expanded data. With the above structure, the amplitude difference data is integrated and compressed twice, and then a moving average operation is performed in the first stage of fj, and the operation result data M is tried as shown in the following formula (8). This paper size applies Chinese National Standard (CNS) A4 specification (210X297 mm) (Please read the precautions on the back before filling out this page), τ Printed by the Employees' Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs -68- 484259 A7 B7 V. Description of the invention (66) (M / 2T2 + F) X1024 = (M / T2) x512 + Fxl024 ...... (8) (Please read the precautions on the back before filling this page) However, F is Compression amplitude data is now rampant. By this, without using a table, a circuit for performing oversampling square interpolation of sample points (variable clocks) corresponding to indefinite intervals can be realized. Fig. 36 is a diagram illustrating an actual operation example of the unfolding process performed by the unfolding device shown in Fig. 34. In Fig. 36, the passage of time is shown from the top to the bottom in the vertical direction. Among the various data rows A to G shown in Fig. 3, the leftmost data row A is a 2 CK oversampling compression amplitude difference data, and a half of the period of the timing pulse TP Week / period information that reverses the sign. For example, the four data rows {— 1, 1, 1, 1, 1} from the top are for the four data rows {— 1, 1, 1, — 1, — 1} to reverse the sign at its middle position. The second data row B from the left is data obtained by integrating the compressed amplitude difference data once by the first adder 1 54 shown in FIG. 35. For example, in the calculation of the top paragraph, the initial stage of D-type flip-flops 155, 〇 printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs " data from D-type flip-flops 153 ,,, One 1 〃 is added, and the accumulated 値 of the D-type flip-flop 15 5 becomes ,, 一 〃. In the operation of the second paragraph, the cumulative 値, 〃, and 〃 of the D-type flip-flop 1 5 5 and the lean material 値, ′-1 来 from the d-type flip-flop 1 5 3 are added, and the d-type The cumulative 値 of the flip-flops 1 5 5 becomes 一 12. In the same manner, by performing the calculations in the third and fourth paragraphs, the cumulative 値 of the D-type flip-flops 1 5 5 becomes,, 1 1 " ,,, 0 " 0 ^ The paper standard is applicable to China Standard (CNS) A4 Regulations for Public Holiday) ----69- 484259 A7 B7 V. Description of Invention (67) (Please read the precautions on the back before filling this page) The third column C is shown in Figure 3 5 The data shown in the second adder 1 5 6 is the result of integrating the compressed amplitude difference data twice. For example, in the calculation of the uppermost stage, the initial stage 値, 〇 正 of the D-type flip-flop 15 7 and the data 値, -1 〃 from the first adder 154 are added, and the D-type forward and reverse The accumulation of the device 1 5 7 becomes a 1 ". In addition, in the operation of the second paragraph, the accumulation 値, 一, 〃 of the D-type flip-flop 1 5 7 and the data 値 2〃 from the first adder 154 are added, and the D-type flip-flop is added. The accumulation of 1 5 7 becomes `` a 3 3 ''. In the same manner, by performing the calculations in the third and fourth paragraphs, the cumulative 値 of the D-type flip-flops 1 5 and 7 sequentially becomes,, -4 ", vv_4. The fourth data column D is obtained by The D-type flip-flop 1 5 8 shown in FIG. 3 is a result of shifting the third data row C by only one clock 2 CK. Moreover, the fifth data row E is obtained by The third adder 159 is the data obtained by adding the third data column C and the fourth data column D. The employee consumer cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs printed as above to obtain the fifth data column. The two integrals and moving average calculations up to E are performed individually in each cycle of each timing pulse TP (each interval between each specimen point). That is, each of them is separated by a dotted line in Figure 36 Each segment performs the above operation. As a result, an independent digital waveform is generated at each segment between individual specimen points. The sixth data column F is represented by the multiplier 1 6 0 shown in FIG. 3 5. The expanded data generated by subsequent processing. That is, the digital waveform obtained in each interval of each specimen point as described above is interpolated within formula (8). This paper is applicable to all paper scales that obtain smooth digital waveforms in successive intervals. National Standards (CNS) A4 Specification (210X297 mm) -70-484259 A7 _ B7 _ V. Description of the Invention (68) Digital Waveform (interpolated data that has been oversampled). (Please read the precautions on the back before filling this page.) In this embodiment, as shown in Figure 29, the turning points (peak points) of the digital waveform ) It is detected as the specimen point, and the data for processing this specimen point is regarded as compressed data. Therefore, the minimum data required to reproduce the original data by expansion processing is included in the compressed data. Therefore, by performing square The interpolation processing can smoothly interpolate data other than the return point with 16-bit precision. In this embodiment, when performing double integration of the compressed amplitude difference data, it will be input to the integrator (first addition) in the initial stage. The sign of the compressed amplitude difference data of the device 1 5 4) is switched at a period of half of the period between each specimen point. Thus, by the integration and moving average calculation of the second paragraph following this Digital waveforms with smoother amplitudes can be obtained. Printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economy The D-type flip-flop is reset at every timing pulse TP. With this, while the overall algorithm is correctly implemented, the cumulative error of the integrator can be eliminated, and a more accurate digital waveform can be reproduced. With the above, according to the basis The unfolding device of the present invention can reproduce the original data almost faithfully. The seventh data column G shown in Fig. 36 shows the data of the original data before compression is interpolated at twice the frequency. If it is compared and displayed in this data column The original data before the compression of G and the expansion data displayed in the data column F can clearly understand that the expansion data generated by the expansion device of this embodiment and the paper size are applicable to the Chinese National Standard (CNS) A4 specification (210X297 mm) ) -71-484259 Printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs A7 B7__V. Description of the invention (69) The original information before the contraction is almost the same値. Figure 37 shows the data lines F and G of the 7 series. It is also clear from this that, according to the expansion device according to this embodiment, it is possible to reproduce almost the same data as the original data before compression. As explained in detail above, according to this embodiment, it is not necessary to perform time / frequency conversion on the digital data to be compressed, and the compression and expansion are performed on the time axis, so the processing will not become complicated, and it can also make The composition is simplified. In addition, by simple interpolation operation on the time axis, the input compressed data can be sequentially processed and reproduced without using table data, and real-time actions can be realized. Also, in this embodiment, the point where the polarity of the differential data of the digital data changes is detected as the specimen point, and the amplitude data of each detected specimen point and the timing data indicating the time interval at which each specimen point appears are generated The compressed data is transmitted or recorded, and only the data of the specimen point can be obtained as compressed data, and a high compression rate can be achieved. In addition, in this embodiment, the amplitude data of each specimen point is not originally regarded as compressed data, and it is rounded to a value of 102. As a result, the data length can be reduced by a few digits per character. Reduce the amount of data. Furthermore, instead of using the rounded amplitude data as compressed data, the differential data can be obtained as compressed data, which can reduce the number of bits necessary for compressed data and reduce the amount of data. Furthermore, in this embodiment, the obtained compressed amplitude difference data and timing data are coded into variable-length block data to be used as the final compressed data. Therefore, the compression ratio can be increased by 1.5 times here. The paper size is applicable to the Chinese National Standard (CNS) A4 (210X 297 mm) (Please read the precautions on the back before filling this page)

-72- 484259 A7 B7 V. Description of the invention (70) The result becomes a compression ratio that can realize non-suspended rounds. In addition, in this embodiment, in order to detect the specimen point, before obtaining the differential value, the integer data is appropriately converted, so that the position of noise components or unwanted signal components may not be detected as the specimen point. It can detect only the correct position as the specimen point. In addition, during the unfolding process, in the early stage of the double integrator, the addition and subtraction can be switched between the first half and the second half of the period of the timing pulse TP, so that the rounding error on the compression side can be offset, and the amplitude 値 can be reproduced more smoothly. Changing digital waveform. In addition, in the present embodiment, when performing calculations such as double integration with oversampling, the accumulation of the integrator of each stage is reset at each timing pulse TP, so that the accumulation error of the integrator can be eliminated and can be reproduced More accurate digital waveform. With this, it is possible to obtain high-quality expanded data close to the original data before compression. From the above, according to the compression and expansion method according to this embodiment, it is possible to provide a new compression and expansion method that realizes both a very high compression ratio and the quality improvement of the reproduced data. In the third embodiment described above, the rounding calculation unit 1 13 shows an example of rounding the amplitude data 値 by 1,024, but it is not limited to this. In the third embodiment described above, although the oversampling is performed twice, it is not limited to two times as long as it is an even number. In the third embodiment described above, during the compression process, the amplitude data of the sample point is extracted from the digital data input as the compression target, and the extracted amplitude data is rounded by the second 値 N 2 However, this paper size is applicable. China National Standard (CNS) A4 specification (210X297 mm) (Please read the precautions on the back before filling out this page) Ordered by the Intellectual Property Bureau of the Ministry of Economic Affairs, Consumer Cooperatives -73- 484259 Α7 Β7 V. Description of the invention (71) After digitalizing the input digital data, it is also possible to extract the compressed amplitude data of the specimen point. In the third embodiment described above, during the unrolling process, the integration is performed at the same time as the one-step moving average calculation. However, the integration is not limited to double, and multiple integrations may be performed. Furthermore, the moving average operation is not limited to one stage, and a moving average operation or a convolution operation may be performed more than that. Further, the methods of compression and expansion according to the first to third embodiments described above may be appropriately combined, or the applicable element technology may be appropriately replaced. For example, in the first to third embodiments described above, although three methods are used as the method for detecting the specimen point, any method may be used in any embodiment. In addition, as described above, the compression and expansion methods according to the first to third embodiments may be implemented by any of a hardware configuration, DS, and software. For example, in the case of software implementation, the compression device and expansion device of this embodiment are actually constituted by a computer's CPU or MPU, RA M, ROM, etc., and can be stored in a program of RA M or R 0 M To achieve. Therefore, a program that enables a computer to function as described in the above embodiment can be recorded on a recording medium such as CD-ROM, and can be implemented by reading it into a computer. As a recording medium for recording the above programs, in addition to CD-ROM, you can also use: floppy disks, hard disks, magnetic tapes, optical discs, magneto-optical discs, DVD, non-volatile memory cards, etc. In addition, the above program can also be downloaded to a computer through the Internet and other networks to achieve this paper size. Applicable to the Chinese National Standard (CNS) A4 specification (210X 297 mm) (Please read the precautions on the back before filling out this page. ) Clothing ·-, 1T Printed by the Consumer Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs -74- Printed by the Consumer Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs 484259 A7 _ B7 _ _ V. Description of the Invention (72) The supplied program not only implements the functions of the above-mentioned implementation form, but also the program and the computer's operating system (OS) or other application software to realize the functions of the above-mentioned implementation form, or the supplied program. In the case where all or a part of the processing is performed by a function expansion board or a function expansion unit of a computer to realize the functions of the above-mentioned embodiment, such programs are also included in the embodiment of the present invention. It should be noted that each of the embodiments described above is merely an example of a specific embodiment when the present invention is implemented, and the technical scope of the present invention is not limitedly interpreted by them. That is, the present invention can be implemented in various forms without departing from its spirit or its main characteristics. Industrial Applicability The present invention is useful in providing a new compression and expansion method with a simple structure, short processing time for compression and expansion, and achieving both a very high compression rate and improved quality of reproduced data . This paper size applies Chinese National Standard (CNS) A4 specification (21〇 > < 297mm) (Please read the precautions on the back before filling this page)

-75-

Claims (1)

484259 A8 B8 C8 D8 6. Scope of patent application 1 · A compression method, which is characterized by: (Please read the precautions on the back before filling in this page) The differential absolute sampling does not become the time interval sampling compression target below the specified point The signal is used as the compressed data to obtain a combination of discrete amplitude data of each specimen point and timing data indicating the time interval between each specimen point. 0 A compression method, which is characterized by: oversampling the signal of the compression object, The differential absolute value becomes the combination of sampling the oversampled data at time intervals below the specified value, and obtaining the discrete amplitude data of each specimen point as compressed data, and the timing data representing the time interval between each specimen point. 3. The compression method as described in item 2 of the scope of the patent application, in which the above-sampled data is processed, and then a continuous sample 値 average 値 data is processed. 4. A compression device, comprising: a differential means for differentially compressing digital data of an object to be compressed; and a specimen point detection means for detecting the absolute value of the differential obtained by the above-mentioned differential means to become a designated specimen point below the designated point; and The Intellectual Property Bureau employee consumer cooperative prints the combination of the amplitude data of the specimen points detected by the above-mentioned specimen point detection method and the timing data indicating the time interval between each specimen point as a compression means for compressing the data output. 5. The compression device as described in item 4 of the scope of patent application, which includes: A / D conversion analog input signal to generate A / D conversion means of the digital data of the compression target. 6. The compression device described in item 4 of the scope of patent application, in which the paper size applies the Chinese National Standard (CNS) A4 specification (210 X 297 mm) 484259 A8 B8 C8 D8 The digital data of the object is an oversampling method that uses an oversampling clock. (Please read the precautions on the back before filling this page.) The above differentiation means differentiates the digital data generated by the above oversampling method. Processing, the above-mentioned compression means outputs the measured timing data based on the clock of the even-numbered frequency. 7 · The compression device as described in item 6 of the scope of patent application, which further includes: means for generating average data from digital data generated by the above-mentioned oversampling means, and processing for generating continuous specimens, averages, and data, The above-mentioned differentiation means performs differentiation processing on the digital data generated by the above-mentioned average data generation means. 8. A kind of expansion method, characterized by: Compressed data composed of a combination of amplitude data of a specified specimen point from which a signal of a compression target is extracted, and timing data indicating a time interval between each specimen point. Using the amplitude data of consecutive specimen points and the timing data therebetween, interpolating data between amplitude data having a time interval displayed by the timing data is obtained to obtain expanded data. Printed by the Employees' Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs9. The unfolding method described in item 8 of the scope of patent application, which uses a sampling function obtained from the amplitude data of two consecutive specimen points and the timing data between To find the interpolated data between the two amplitude data. 1 0 · —A kind of expansion device, which has the following characteristics: Timing data indicating the time interval between the specimen points according to the signal of the compression object contained in the compressed data is extracted, and the paper size of the above specimens is applicable to the Chinese country Standard (CNS) A4 specification (210X297 mm) -77- 484259 A8 B8 C8 D8 VI. Each time interval between the patent application points is sequentially taken from the amplitude data of each specimen point included in the above compressed data , Timing control means to control the timing; and (please read the precautions on the back before filling this page) use the amplitude data of 2 consecutive 2 specimen points taken in accordance with the control of the timing control described above and other The timing data between the two is obtained by interpolating the interpolation data between the two amplitude data to obtain the expansion means of the expansion data. 1 1. The expansion device described in item 10 of the scope of patent application, wherein the expansion means uses a sampling function obtained from the amplitude data of the two consecutive sample points and the timing data between them, Find the data between the two amplitude data interpolated. 1 2. A compression and expansion system, which is characterized in that: on the compression side, the signal of the compression object is sampled at a time interval where the differential absolute 値 becomes the specified point, and the discrete amplitude data of each specimen point is obtained as the compression data; and The combination of timing data indicating the time interval between each specimen point is printed on the expansion side by the consumer property cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs, and the amplitude data and timing data included in the compressed data are used to use the above timing. At the same time as the amplitude data is reproduced at the time when the data is displayed, the interpolation data between the two amplitude data is obtained based on the amplitude data of two consecutive sample points and the timing data between them. To get expanded information. 1 3 · —A computer-readable recording medium, which is characterized by recording a program for implementing the processing order of the compression method described in item 1 of the scope of patent application on a computer. 1 4 · A recording medium that can be read by a computer, which is characterized by: This paper size applies to China National Standard (CNS) 8 4 specifications (21〇 > < 297 mm) -78- 484259 A8 B8 C8 ____ D8 6 2. The scope of application for patent application records the program used to implement the processing method described in item 8 of the scope of application for patent on a computer. (Please read the notes on the back before filling this page) 1 5 · — A compression method, characterized by: Oversampling and moving average operation or convolution operation to synthesize one of the η discrete data that should be input After the digital data of the basic waveform is obtained, the interpolation of the digits of the discrete data is obtained, and the interpolation of the digits obtained above is sampled at the time interval at which the absolute value of the differential becomes a minimum point, and each specimen is obtained as compressed data A combination of discrete amplitude data of points and timing data representing the time interval between each specimen point. 1 6 · A compression device comprising: digital data of basic waveforms of η discrete data corresponding to η of input data synthesized by oversampling and moving average operation or convolution operation, and obtaining the above-mentioned discrete data Over-sampling means of digital interpolation; and differential means of derivation of digital interpolation obtained by the above-mentioned over-sampling method; and printed detection by the employee's consumer cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs through the above-mentioned differential means. The obtained differential absolute 値 becomes a very small specimen point, which is a means of generating compressed data as a combination of the amplitude data of the specimen point detected by the compressed data output detection and the timing data indicating the time interval between each specimen point. 17. The compression device as described in item 16 of the scope of the patent application, which includes: a rounding operation means for performing rounding operations on the lower-order bits of the differential absolute value obtained by the above-mentioned differentiation means, and the above-mentioned compressed data is generated The method will be applied to the paper size by the above-mentioned rounding calculation method. The paper size applies the Chinese National Standard (CNS) A4 specification (210 × 297 mm) 484259 A8 B8 C8 D8. 6. The range of patent application for the differentiation of rounding operation is definitely a very small point. Detected as a specimen point (please read the precautions on the back before filling in this page) 1 8 · If the compression device described in item 16 of the patent application scope, the above-mentioned compression data generation means are obtained by the above-mentioned differential means The absolute absolute value of the differential, or the differential absolute value applied to the rounding operation by the above-mentioned rounding operation means, becomes the smallest point, and only the point where the above-mentioned differential absolute value becomes smaller than a certain value is detected as the specimen point. 1 9 · A compression device comprising: digital data of basic waveforms corresponding to η discrete data to be input by oversampling and moving average operation or convolution operation, and obtaining the above-mentioned discrete data. An oversampling means for digital interpolation; and a first differentiation means for differentiating the interpolation of the digits obtained by the above oversampling means to obtain a differential absolute value; and further the differentiation is obtained by the above first differentiation means The obtained differential absolute value is the second differential means to obtain the double differential value; and the polarity of the double differential value printed by the employee consumer cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs will be changed from negative or zero to The positive point is regarded as the absolute value of the differential obtained by the above-mentioned first differentiation means, and becomes the specimen point detection means for detecting the extremely small specimen point; and the amplitude data of each specimen point detected by the aforementioned specimen point detection means, and The combination of timing data representing the time interval between each specimen point is used as the compressed data generating means for the compressed data output. 2 0. The compression device described in item 19 of the scope of the patent application, in which the above-mentioned sample point detection means will be obtained by the above-mentioned second differential means. The paper size is applicable to the Chinese National Standard (CNS) A4 specification (210 X 297). (Mm) 484259 A8 B8 C8 D8 6. Scope of patent application (please read the precautions on the back before filling this page) The polarity of the double differential 由 changes from negative or zero to positive, and is obtained by the first differential means above The point where the obtained differential is absolutely smaller than a certain value is detected as the above-mentioned specimen point. 2 1 · The compression device as described in item 19 of the scope of the patent application, which includes: a rounding operation means for rounding the lower-order digits of the differential absolute value obtained by the first differential means, 2 Differential means Absolutely perform differentiation on the differentiation that has been subjected to the rounding operation by the above-mentioned rounding operation means, and further differentiate. 2 2 · The compression device as described in item 21 of the scope of patent application, wherein the specimen point detection means is among points where the polarity of the double differential 値 obtained from the second differential means is changed from negative or zero to positive The point where the differential absolute 値 that is applied to the rounding operation by the above-mentioned rounding operation means is less than a certain value is detected as the above-mentioned specimen point. 2 3 · The compression device as described in item 16 of the scope of patent application, wherein the oversampling means performs digital data on the basic waveform corresponding to one of the η discrete data inputted above using an integer multiple frequency clock. Sampling, printed by the Consumer Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs, the above-mentioned compressed data generating means outputs the measured timing data according to the above-mentioned integer multiple frequency clock. 2 4 · The compression device described in item 16 of the scope of the patent application, wherein the oversampling means includes: oversampling the digital data of the basic waveform, pre-memorizing the data of the oversampled basic waveform, and performing moving average calculation Or the means of memorizing the data obtained by the convolution operation; and modulating the data memorized in the above memorizing means to apply the Chinese National Standard (CNS) A4 specification (210 × 297 mm) in accordance with the above paper size -81-484259 A8 B8 C8 D8 6. The amplitude of the 値 discrete data of the inputted patent range is synthesized by moving average calculation or convolution operation. The η data obtained by this are combined with each other (please read the back first) (Please fill in this page). 2 5 · — A compression and expansion system, characterized in that: on the compression side, digital data of basic waveforms corresponding to η discrete data inputted by oversampling and moving average operation or convolution operation are obtained and obtained After the digital interpolation of the discrete data, the obtained digital interpolation is sampled at a time interval where the absolute absolute value of the differential data becomes a minimum point, and the discrete amplitude data and representation of each specimen point are obtained as compressed data. At the same time as the combination of the timing data of the time interval between each specimen point, on the unfolding side, the amplitude data and the timing data included in the compressed data are used, based on the amplitude data of two consecutive specimen points. With the timing data in between, the interpolation data obtained by interpolating the two amplitude data is obtained to obtain the expanded data. Printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs 26. The compression and expansion system described in item 25 of the scope of patent application, where the digital data of the basic waveform corresponding to one of the η discrete data entered above is used as an integer Overclocked clocks are oversampled, and the measured timing data is output based on the clocks of the integer frequency. 2 7 · The compression and expansion system described in item 25 of the scope of patent application, wherein the limited sampling function obtained from the amplitude data of the two consecutive sample points and the timing data between them is used to obtain Interpolate data between the two amplitude data. 2 8 · —A computer-readable recording medium, characterized by: Recording the processing of the compression method described in item 15 of the scope of patent application. The paper size applies the Chinese National Standard (CNS) M specification (21〇 > <; 297 mm 484259 8 8 8 8 ABCD κ, a program implemented in the order of patent applications on a computer. 2 9 · —A computer-readable recording medium, characterized by: (Please read the notes on the back before filling in (This page) Records programs that make the computer function as the means described in item 16 of the scope of patent application. 30 0 —A kind of recording medium that can be read by a computer, which features: The function of the recorded compression and expansion system is implemented in a computer. 3 1. — A compression method, which is characterized by: Differentiating the input digital data, detecting the point where the polarity of the differential 値 changes as the specimen point, as Discrete compressed amplitude data at each specimen point is obtained as digital data rounded to a specified 値, and is used as compressed data to obtain the difference between the compressed amplitude data and each other. The combination of the compressed amplitude difference data and the timing data indicating the time interval between the specimen points. Printed by the Consumer Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs 3 2. The compression method described in item 31 of the scope of patent application, where The first rounding operation is performed on the input digital data, and the differential should be based on the first rounding digital data. The point of the polarity change of the differential round is detected as the specimen point, and it is used as a discrete compression of each specimen point. The amplitude data is obtained by the second rounded digital data larger than the first one. 3 3 · The compression method described in item 31 of the patent application scope, wherein the compressed amplitude difference data and the timing data are converted. It is a variable-length block data. The compression method described in item 3 of the scope of the patent application, wherein one of the variable-length blocks of the compressed amplitude difference data indicates the polarity of the compressed amplitude difference data. The size of the bit, and the size of the above paper is applicable to the Chinese National Standard (CNS) A4 specification (210 × 297) 484259 A8 B8 C8 D8 夂, The patent scope describes the segmentation flag of the bit length of the compressed amplitude difference data and the above-mentioned compressed amplitude difference data with the bit length expressed by the above-mentioned segmentation flag (please read the precautions on the back before filling in (This page). 3 5 · The compression method described in item 33 of the scope of patent application, wherein one variable-length block of the above-mentioned timing data is a flag indicating the bit length of the above-mentioned timing data. And the above-mentioned timing data having the bit length indicated by the above-mentioned segmentation flag. 3 6 · A compression device, which is characterized by: having: Differentiating the input digital data, and dividing the differential data into The point where the polarity changes is detected as the specimen point, and a timing data generating means is obtained for the timing data representing the time interval between the specimen points; and the discrete compressed amplitude data of each specimen point detected by the timing data generating means To generate a compressed amplitude data generating means for rounding the inputted digital data by a specified unit; and to generate a hand from the compressed amplitude data. The difference calculation method for the difference between the obtained compressed amplitude data is printed by the consumer cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs as compressed data to obtain the compressed amplitude difference data obtained by the above-mentioned difference calculation method and generated by the above-mentioned timing data. A combination of timing data obtained by means. 37. The compression device described in item 36 of the scope of patent application, which includes: a first rounding operation means for rounding the inputted digital data by the first round, and the timing data generating means is differentiated by the first 1 The rounded digital means is rounded digital data, and the point of change in the polarity of its differential 値 is used as the standard of this paper. Chinese National Standard (CNS) A4 specification (210X297 mm) -84- 484259 8 8 8 8 ABCD Economy The Ministry of Intellectual Property Bureau employee consumer cooperative printed the patent application, and the patent application scope was detected as the specimen point. The compressed amplitude data generating means generates the compressed amplitude by rounding the digital data with a second value larger than the first value. data. 38. The compression device described in item 36 of the scope of patent application, wherein the compression amplitude data generating means includes: extracting the amplitude data of the digital data of each specimen point detected by the timing data generating means as discrete amplitude data. A means for extracting; and a second means for calculating the amplitude data that has been extracted by the amplitude data extracting means in the second rounding. 39. The compression device according to item 36 of the scope of patent application, which includes encoding means for converting the compressed amplitude difference data and the timing data into variable-length block data. 40. The compression device as described in item 39 of the scope of the patent application, wherein one variable-length block of the compressed amplitude difference data is a size bit indicating the polarity of the compressed amplitude difference data, and the above A segmentation flag having a bit length of the compressed amplitude difference data and the above-mentioned compressed amplitude difference data having a bit length represented by the above-mentioned segmentation flag are configured. 4 1 · The compression device described in item 39 of the scope of patent application, wherein one of the variable-length blocks of the above-mentioned timing data is represented by a segmentation flag indicating the bit length of the above-mentioned timing data, The above-mentioned timing data of the bit length indicated by the above-mentioned segmentation flag is constituted. 4 2 · —A method of expansion, characterized by: About the designated specimen points extracted from the digital data of the compression target The paper size applies the Chinese National Standard (CNS) A4 specification (210 X 297 mm) (Please read first Note on the back, please fill out this page again) 484259 A8 B8 C8 D8 々 Compressed data composed of a combination of patented compressed amplitude difference data and timing data indicating the time interval between specimen points, (Please read the back Note that please fill in this page again) At the same time that the multiple integrals are oversampled by the above-mentioned compressed amplitude difference data, a moving average operation is performed on the integral 値, and the moving average operation 由此 obtained by this and the above timing data are used to obtain The interpolated data having the square interpolation between the amplitude data of each specimen point in the time interval displayed by the above-mentioned timing data is obtained as expanded data. 4 3 · The unfolding method described in item 42 of the scope of patent application, wherein the above-mentioned compressed amplitude difference data which is oversampled by even multiples at an intermediate position of each segment of each specimen point represented by the above-mentioned timing data The sign is reversed, and multiple integrals are obtained from the data rows obtained thereby. 4 4 · The unfolding method described in item 42 of the scope of the patent application, wherein the multiplexing of the above-mentioned compressed amplitude difference data of the even-numbered oversampling is performed at each segment of each specimen point represented by the above-mentioned timing data. Integral and moving average calculations. 4 5 · —A spreading device, characterized in that: the consumer cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs prints the compressed amplitude difference data of the designated specimen points extracted from the digital data of the compression target and indicates the time between each specimen point Compressed data composed of a combination of interval timing data includes: an integration means for the above-mentioned compressed amplitude difference data in which multiple integrals are oversampled by an even number of times; and a moving average calculation of the integrals obtained by the above integration means. Mean calculation method; and using the moving average calculation obtained by the above-mentioned moving average calculation method and the above-mentioned timing data, the interpolation is obtained by applying the Chinese National Standard (CNS) A4 specification to the paper standard by the above-mentioned timing data ( 210X297 mm) -86- 484259 A8 B8 C8 __ D8 々. The square interpolation data between the amplitude data of each specimen point at the time interval displayed in the patent application range is used as a means of generating the interpolation data of the expanded data. (Please read the precautions on the back before filling in this page) 4 6 · If the expansion device described in item 45 of the scope of patent application, the above-mentioned integration means has: In the initial stage of the above-mentioned multiple integrations, the The sign inversion means that the middle position of each segment of each specimen point inverts the sign of the above-mentioned compressed amplitude difference data that is oversampled by an even number. 4 7 · The unfolding device described in item 45 of the scope of patent application, wherein the above-mentioned integration means is provided with: resetting each of the above-mentioned even-multiplied oversampling at each segment of each specimen point represented by the above-mentioned timing data Reset method for integral repair of compressed amplitude difference data. 4 8 · The expansion device described in item 45 of the scope of patent application, wherein the moving average calculation means is reset at each segment of each specimen point indicated by the timing data, and is obtained by the integration means. Integral 値 Moving average calculation 重置 Reset method. 4 9 · A computer-readable recording medium, which is characterized by: Printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs, which records the processing method of the compression method described in item 31 of the scope of patent application. 50. A recording medium readable by a computer, which is characterized by recording a program for processing the compression method described in item 42 of the scope of patent application on a computer. 5 1 · —A kind of recording medium that can be read by a computer, _ Zheng Meng: Recording ^ described in item 36 of the scope of patent application, to make the computer function. This paper size applies Chinese National Standard (CNS) A4 specification (210X297 mm) -87- 484259 8 8 8 8 ABCD application patent scope 5 2. — A kind of recording medium that electric wax can read, The ft sign is: Record the program described in Item 45 of the scope of patent application ^ Make electricity \ 1 A · ,,., /,. ,, human \ brain function. Printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs (Please read the precautions on the back before filling this page) This paper size applies to China National Standard (CNS) A4 (210X297 mm) -88-