JPS62501115A - Data compression computer system and method therefor - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

[Detailed description of the invention] data compression computer system and How to do that Skin - 10,000 yen per diem The present invention can code any numbers, alphabets and special characters, or any suitable number. data, compressed into memory systems, and currently stored in communications systems. A new and useful method for transmitting and receiving in a more economical manner than currently used methods. The present invention relates to a method and apparatus for compressing data to a higher level of compression. for compression into memory systems, communication systems, and their associated equipment. The present invention relates to a method and apparatus for iterative coding.

Five once, one technique, one hesitation Since the first computer was created, data has been represented in computer memory. The most advantageous way to do this is to make the base 3 more difficult to implement electronically. Although both are likely to result in some expansion of computer memory, It has become a self-evident fact that it is a binary system. Also, the amount of data is less important buffer for numeric data with better coding density, resulting in less improvement. Encrypted formats and methods of compressing data are also known.

1. U.S. Pat. No. 3,71 to John Cocke et al., dated February 20, 1973. No. 7,815 uses a single character (a 1 etter) and Rather than using letters, groups of letters of the alphabet (word prefixes) were used.

2. Joseph L., OoNell Jr., dated February 3, 1976. National Patent No. 3,936,808 utilized by video display monitor The data was compressed in order to Each coated repeat sequence is The display data character is a repeat operation and the display data character is the display - Specify a specific number of times that should be repeated within the line.

3. U.S. Patent No. 4,1 to Wan Y. Weng, dated April 17, 1979. No. 50,400 is coated to distinguish between the white background and the black elements of the text. and using a coarse scanning device for compressed, analog signals. This invention The data is compressed by removing some elements of the data.

4. U.S. Patent to John F., Paulgstat, dated November 4, 1980. No. 4,232,375 is compressed regarding the types of commodity transactions that are being processed. was producing data records that were

The present invention uses specified codes for limited data groups. Therefore, the data is compressed.

5. Glenn A., Reijmeier, U.S. Special Issue dated August 4, 1981. No. 4,282,546 uses interpolated pixel values and uses unimportant data. The size of the television image was compressed by removing the .

6. US Reference Manual-TBM 1410 Data ProcessingSystal (April 1982) Zero control operation was used for this purpose.

The prior art in all the above examples was based on prior theory. Therefore, data The development of the coating system in the sense of compressing data is based on mathematical information theory, Among others, it was done in that by C.5hannon. C,5hannon's According to theory, the transmission speed cannot exceed the technical speed of the transmission line. That is, , the data cannot be compressed and the code as shown in the example above The only thing that can be done is to improve the quality of the especially in the case of binary systems The maximum speed is 73.32...(numbers per second), i.e. in hexadecimal format. In this case, the speed is very close to speed/4.

According to the above passage of prior theory and the examples given therein, it is possible to greatly compress the data. There is no prior art that reduces this.

Regarding the implementation of the block diagram shown in the figure, if Even if there is something similar in appearance to the block diagram of It does not necessarily represent the process of Because it was invented by the creator of this invention This is because the compression process used was unknown.

The creators of this invention challenge the fundamentals of information theory to compress data beyond the speed of the transmission line. make a discovery that goes against the obvious with current methods, that it cannot be done, Based on this discovery, the novelty of the invention was established.

The new mathematical idea created by the inventor has never been published. , described and explained below. Below, inter alia, the section “Description of the Preferred Embodiments” The terms and symbols used in this are probably not computer terms but It uses terminology from Shiro Mathematics. However, the invention itself does not require a person skilled in the art to make or utilize it. The term is used so that it can be done.

Now, this new mathematical idea consists of the following.

Once the number of bits of data, or any other representation, is at least Find a way to reduce it by (1+a) times (where (a) is any number or fraction) If possible, by repeating the method (m) times, It would be possible to increase memory capacity or transmission line speed by a multiple of That is, n=(1+a,)'').

However, information theory proves that the maximum speed of transmission is log, n. certain types of mathematical structures that are clear, such as the sum of , *pX +-+kx++'P0 and other methods and combinations; If so, it is impossible to even increase the transmission speed by (1+a) times.

Some more powerful structure and method should be used. The creator of this invention of the data as described in the Section and the “Description of the Preferred Embodiment” section. Dimensional co-ordin of data position as value ate) is used. According to this method, the transmission speed The power can be increased by (1+a) times.

Let's take a group of bytes like this:

However, (b) is the number of bits in each byte. Therefore (b) is from (0) (n) numbers up to (2b), or log. (2b*n) represents digit be able to.

The first group of bytes in F can be transformed as follows.

However, each shaded area indicates the “ON” (or “OFF”) state. bits that are in the “off” state and each blank part determines the interval dk. (or "on") state of one or more bits.

At least for some types of dk, the interval is large enough that dk is the new bit can be encoded by expressing it as a group of It does not require log 2 (dk) bits.

As noted in the “Description of the Preferred Embodiment” section, such bits The possible savings is 75% of the total number of bits for data with a sufficiently large amount of bits. It is.

Therefore, in the case of data with a sufficiently large amount of bits, it is contrary to information theory and this restriction Breaking the convention is to encode (1+a) bits of information for each bit. can be done.

The actual algorithm for this coding method is described in the “Description of the Preferred Embodiment” section. is explained in.

λ”LΩ”L The present invention overcomes the limitations of memory capacity and increases the speed of computers and related devices. Increase. The present invention, together with the present invention, exceeds the current state of the art in computers. Take advantage of the fact that for most problems today you have virtually unlimited memory capacity. Making inexpensive computers, computer systems, and related equipment ( or bring about the necessary improvements), and the present invention also makes it possible to A compact alternative to main memory and auxiliary memory in a system. It creates possibilities for using the device and opens it up in many traditional and new areas. This allows you to use less expensive portable memory. The final purpose is Increase processing speed, or at least coding and The purpose is to compensate for the decoding time.

Another object of the invention is to develop new codes for transmission and reception in communication and information systems. the possibility of using coding and decoding methods, and this possibility is not currently used. This would increase the transmission speed currently available many times over.

Wildebeest Jtt Day The invention is thus described below with reference to non-limiting embodiments of the invention and the attached drawings: It will be easier to understand in more detail. In the attached drawing The drawings may be made in accordance with the principles of the invention or used in conjunction with the invention. A block diagram of systems and equipment that can be used to describe the invention in detail. The block diagram and the following diagram are block diagrams of one useful variation of this method. This is a diagram. That is, FIG. 1A is a block diagram of a coding system that can be used with the present invention. tire gram, Figure 1B is a block diagram of a decoding system that may be used with the present invention. diagram, FIG. 1C is for an iterative coding system that can be used with the present invention. Block diagram of the control unit of 'fJZA diagram is one example of a coding method that can be used with the present invention. A block diagram of a modified example, FIG. 2B, for use with the present invention is shown in FIG. Block diagram of one useful variant of the decoding method that can be used, Part 2C The figure shows counters for coping and decoding that can be used with the present invention. 1 is a block diagram of one useful variant of the decoding method;

10 beautiful 'Jino' balls For illustrative purposes, only one example of each type of system and method is described below. However, stiffness is in no way intended to limit the scope of the present invention to stiffness.

Figure 1A schematically shows the equipment of a coding system for encoding data. The device stores the coordinates of the “on” (or “off”) bit position in the data. By using a coding method that uses new values, data can be reduced. Compress to zero bits.

Figure 2A shows a compression coding method that can be used as follows. .

The first stage (25) is to apply the average data probability to the original bytes and perform compression. It consists of grouping into a set of numbers large enough to This stage For example, groups 1000 bits or more for one coding. can be converted into If this coding process is repeated, The bite from the larger group should be selected. If a different array If a process is selected, this step can be omitted.

The second stage (26) consists of expanding the original bytes. This stage is , the selection is made only when presented with an original byte of a different base than the selected pace. used to obtain the number of bits in the group as selected.

The third stage (27) can be used as follows. First, as the original part-time job, Select the desired number of bits. The selected group of bytes is changed into a set of bits.

The spacing between each two adjacent “on” (or “off”) state bits is are combined into a new set of numbers. For a sufficiently large amount of data, the interval (0) becomes The probability of being found is 25% of all bits.

Therefore, the probability for each of the following intervals (l, 2.3, ..., n) is 1 2.5%, 6.25%,..., (1/(2°+2))*too% and Nako The stage can consist of the following sub-stages:

1. During the last pair between the closest states of “on” (or “off”) bits A new group is selected from the interval. This new set shows only intervals (0) and (1) Therefore, the first set of bits consists of 37.5% bits and the separately coded l Some to identify the length of a new set that only has length og2L+e bits. However, (L) above is the new group leader. and (e) is some number of bits for additional coding. child A subset of the and this subset can be coded as Can be done.

one subset containing intervals greater than (0) and (1) from the original set of intervals; , a second subset is selected indicating their position within the interval of the original set. It will be done. The remaining number is a subset of the intervals (0) and (1) chosen from the original set. is encoded into memory, and this is followed by subprograms in memory as described above. The beginning and end of the set are identified.

2. The two derived subsets are the same as the spacing of the original set in the previous sub-stage. is encoded into memory in the same way that it is encoded into memory. this The process at least returns the last of the remaining subset to the directly coded interval We can continue until it becomes economical to represent it as a subset of . second sa There may be cases where the bus set is already sufficiently economical. In this case the following 5 sets of subsets and 5 specifications to identify the length of each subset. will result in a byte: L Coordinates of the interval of the second subset From the second subset L from the second subset 3. All subsets consisting of two or more different numbers (in the last example The subset (4) of directly encoded using

Original interval of next compression process based on subsets (2) and (5) or the same data is added to the set of

The fourth step (28) is the interval already recognized as the length of each of the above subsets. The beginning of the subset encoded as Bi + IBi and put into memory give the beginning and end each an address for this specified byte and subset. and encode them into memory in a known order by It is made up of

The fifth step (29) is to encode the data to be encoded and put into memory. is the last compression process based on the same data, the computer system It consists of updating the original byte group to use the system. . This step is omitted if another device coding method is chosen. Something happened.

Figure 1B shows the original code for using coded data in a computer system. 1 shows an apparatus of a decoding system for decoding into shapes.

Figure 2B shows the decode for data coded by the above coding method. It shows you how to do this. This method can be used as follows.

The first step (30) is to convert the data in the form of original bytes into the coding process. It consists of grouping into a set of bytes as encoded.

The second step (31) is to locate the beginning and end of each coded subset in the correct order. It consists of decoding in sequence.

The third step (32) involves performing arbitrary computer processing from the subset of intervals and their values. In order to generate the number of original combinations of coding surfaces using It consists of decoding the data. For example, at the end of coding Use the 5-tuple subset from the example to reconstruct the original bit combinations be able to.

The fourth step (33), if necessary in the equipment chosen for this process, For example, the combination of decoded bits can be used to utilize the original computer system. Convert to byte format.

FIG. 1C shows the control unit for the iterative process of coating and decoding. The layout of the parts is schematically shown.

Figure 2C shows the iterative process of coding and decoding; The process can be used as follows.

steps (34), (35), (36), (37), (38), and step (40), (41), (42), (43) are steps (25), (26), (27), (28) , (29), and steps (30), (31), (32), and (33), respectively. It is. The level of order of iterative compression of the data (39) is determined after each coding. Depending on the amount of memory occupied by the data, the Indicates the number of times of mapping and decoding. code for the same data The number of rings is written into the specified byte of memory. This process Iterations can be repeated as many times as necessary until the desired compression of the data is obtained. Cor Use the decoding process to avoid possible errors during decoding. and compare it with the original data.

The equipment shown in Figures 1A and 1B includes information and communication computers and related equipment. As transmitting units (1), (6) and receiving units in systems and equipment (5) and (10). It also looks like the above coding and independent decoding methods including software and hardware based on decoding methods. It can be constructed as a data processing and compression system and device. those systems and the device provides directed access to information of different priorities in the encoded data. Use different levels of data compression for different information depending on accessibility be able to. It requires more time due to higher data compression level. To the extent that it is necessary, it will be a "multi-level depth process."

Therefore, the deeper the compressed data is in memory, the easier it is to access it. It will probably take a lot of time. Data accessibility is determined by each code. Data entry containing newly coded blocks with level of coding It can be increased by coding the date tree. . When decoding this information, the minimum access time is You will need it for the ninth part of the book.

Brief description of the drawing Figure 1B Figure 1C @2A diagram 14410-PCT Figure 2B Figure 2C (1441,0-PCT) Submission of translation of written amendment 1! ) (Patent Law Article 184-7, Paragraph 1) July 1986! 10

Claims (4)

[Claims] 1. Any information can be stored by turning a bit in memory “ON” (or “OFF”). )), and then this combination can be represented as a combination of the states of the bits the distance between “on” (or “off”) states, or just “dots” (or “blank”). ”) and these intervals may be New groove information, software and hardware for data volume and processing equipment A bit “on” (or “off”) that can be coded into software. ) as the new value for that data. How to compress. 2. A subset that removes the intervals (0) and (1) from the original combination of intervals, and the original a second subset is selected indicating the position of said interval in the set of The remaining number is a subset of the intervals (0) and (1) chosen from the original set. into memory, identifying the beginning and end of the last subset in memory. The two sets of subsets coded and pulled out first are intervals (0) and (1 ) into memory in the same way that the original set for At least the last of the remaining subset was directly coded The above process continues until it becomes economical to express it as a subset interval. , from the combination of coded data, the first part of each coded subset is The bytes that identified the beginning and end are decoded in the exact order, and each decoded The same number of original sets as before coding are created according to each subset; The process can also be used to compare the encoded data with the original data. 2. The coding and decoding method according to claim 1, wherein: 3. Data is accessible throughout the process according to the desired data compression level new data and compress the coded numbers again according to the capacity and quantity. The method described in claim 1 can be repeated as many times as necessary by using the method as described in claim 1. Coding and decoding methods described. 4. Coding and decoding according to the amount of information and the priority of information accessibility. for iteratively compressing data using a control unit that directs the number of Systems and equipment, also within the central unit, have different priorities and accessibility A control unit that uses different levels of data compression for the information in the claim A data storage and processing device according to scope 1.