KR100448283B1 - Method for balanced encoding and decoding of blocks having the different constant weight - Google Patents

Abstract

PURPOSE: Balanced encoding and decoding method using a grouped weight block are provided to improve a code ratio by encoding source blocks corresponding to the number of bits in a weight block to weight blocks. CONSTITUTION: The balanced encoding method converts a source block of m bits into the first and second weight blocks of n(where, n is an odd number) bits(n>m). A source block is grouped by (n+1) blocks. One of the (n+1) source blocks is converted into the first weight block. The remaining blocks of the (n+1) source blocks are converted into the first weight block or the second weight block according to a bit logic of an encoding balanced block. One of the (n+1) grouped source blocks is the first source block of the source blocks in a group. A weight block according to one source block among the (n+1) grouped source blocks is eliminated. The number of 1 in the first weight block is set to a value which satisfies 2¬m<nC2. The number(t) of 1 in the second weight block is set to a value which satisfies t=(n-a). The source block is converted into the first weight block and the second weight block by turns.

Description

METHOD FOR BALANCED ENCODING AND DECODING OF BLOCKS HAVING THE DIFFERENT CONSTANT WEIGHT}

The present invention relates to a block encoding and decoding method, and more particularly, to a grouped weight block encoding and decoding method for improving encoding efficiency.

In the field of digital video cassette recorders or holograms, a weight block coding method is used as a method of encoding information. The weight block code and the decoding method are configured to equally configure the number of 0 bits and 1 bit in a block composed of n bits. In a digital video cassette recorder, the DC value of a recording signal can be set to 0. Can be kept constant.

The weight block code may be divided into a weight block code such as 6: 8 or 8:12. A 6: 8 weight block code converts a block of 6 bits (called the original block) into an 8-bit block (called a weight block). An 8-bit weight block does not contain all the bits represented by 8 bits. It consists of only bits ( ₈ C ₄ ) consisting of four 1s and four zeros. A 6-bit circle block can represent 64 pieces of information (2 ⁶ = 64), and an 8-bit piece of information consisting of four 1s and four zeros can represent 70 pieces of information ( ₈ C ₄ = 70). . Therefore, the 64 pieces of information of the 6-bit original block may be symmetrically encoded with arbitrary information in the 8-bit weight block.

In order to decode the coded 8-bit weight block codes, four bits having the highest intensity (or level) among the 8-bit information in the weight block are defined as 1 and the remaining bits as 0. The information is then symmetrical to 6-bit original block information and decoded into 6-bit original information.

The weight block coding method of 8:12 is also the same as the weight block coding method of 6: 8. That is, the 12-bit weight block corresponding to the original 8-bit block consists of 12 bits ₁₂ C ₆ composed of six 1s and six zeros. Here, the 8-bit original block can express 2 ⁸ = 256 information, and the 12-bit weight block can express ₁₂ C ₆ = 924 information, so 256 information of the 8-bit original block is 12-bit weight block. I symmetrically encode it with my random information.

In order to decode the coded 12-bit weight block codes, 6 bits having the highest intensity (or level) among 12 bits of information in the weight block are defined as 1 and the remaining bits as 0. The information is symmetrical with 8-bit original block information and decoded into 8-bit original information.

The weight block coding method described above has the advantage that the error rate is low, but has a problem that the code rate is low. For example, in the 4: 6 weighted block coding method, since the six bits constituting the block code include only 4 bits of information bits, only 67% of the codes are used. In the 6: 8 weighted block coding method, the block code can be configured. Eight bits contain only 6 bits of information, so only 75% of the code is used.In the 8:12 block code, the 12 bits that make up the block code contain only 8 bits of information, resulting in 67% of the code. Only code is used. Therefore, the weight block coding method has a problem that the code rate is considerably lowered.

In order to solve this problem, the present inventor has filed a 'balanced encoding and decoding method using a weighted block (application number No.)' with the same as the present application. In the present invention, the present inventor converts m-bit original block information into n-bit weight block information and balance-codes it using the same, and decodes the encoded n-bit weight block information back to m-bit original block information. The method is presented. The contents of the present invention are outlined as follows.

The present invention uses two weight blocks instead of one weight block in encoding the original block into a weight block. For example, two 7-bit weight blocks are used to encode a 5-bit original block with a 7-bit weight block code.

Specifically, a 5-bit original block can express 2 ⁵ = 32 pieces of information. In addition, the number of information ( ₇ C ₃ ) that can be represented by 7 bits (hereinafter referred to as A type) in which the number of 1s is 3 (the number of 0s is 4) is 35, and the number of 1s is 4 ( The number of information ( ₇ C ₄ ) that can be represented by 7 bits (hereinafter, referred to as B type) having 3 numbers of 0 is 35 pieces. Therefore, it can be seen that the original block of 5 bits can be represented by 7 bits of type A or B. That is, the weight block code can be formed by 7 bits of A type or B type. However, in the case of the A-type or B-type weight block code described above, the original purpose of the weight block code cannot be achieved because the number of 0s and the number of 1s are different with respect to the entire code. The inventors have solved this problem in the following way.

If the first 5-bit information is input after the A-type and the 5-bit information is input by the B-type weight block code without encoding 5-bit information into one of the A-type or B-weighted block codes, the entire weighted block code is encoded. The number of 0's and 1's is the same. Thus, for example, when the odd-numbered five-bit information is converted into an A-type weight block code, and the even-numbered five-bit information is converted into a B-type weight block code, the total number of 1's and 0's in the weight block code is the same. Do.

When decoding the weight block code of type A or B into 5 bits of original information, a process opposite to that described above may be performed. That is, since the first weight block is a type A weight block in the above example, three of the bits in the block are processed in logic 1 as logic 1 and the rest as 0, and the processed bits are decoded back into 5 bits of information. do. Since the second weight block is a type B weight block in the above example, four of the bits in the block are processed in logic 1 as logic 1 and the rest as 0, and the processed bits are decoded back into 5 bits of information. That is, even in the decoding process, the odd-numbered weight blocks are recognized as the A-type weight blocks, and the even-numbered weight blocks are recognized as the B-type weight blocks and processed.

The above-described example describes a case in which balanced information is encoded and decoded by encoding information of an original block into information of two weight blocks having odd bits. In this case, the weight blocks that can be selected include three and one. Only a combination of four is possible. In the present invention, a combination of selectable weight blocks may exist in two or more plurality. An example is described below in this case.

In this example, an 8-bit original block is encoded into an 11-bit weight block.

The number of information that an 8-bit original block can represent is 2 ⁸ = 256 pieces. In addition, the number of information that can be represented by 11 bits having 4 (or 5) numbers of ₁ is ₁₁ C ₇ = 330 (or ₁₁ C ₆ = 462) and 11 bits having 7 (or 6) numbers of 1s. The number of information that can be represented is ₁₁ C ₄ = 330 (or ₁₁ C ₅ = 462). Therefore, 11 bits having 4 (or 5) numbers of 1s are type A weight blocks and 7 of 1s ( (Or 6) 11 bit is set as a B-weighted block (regarding the operation speed, it is preferable to search 11 bits having 4 (or 5) numbers of 0). It can be seen that the block is encoded with arbitrary bits of the A-type weight block, and the even-numbered input original block can be encoded with arbitrary bits of the B-type weight block.

In order to decode an A-type or B-type weight block into an 8-bit original block, a process opposite to that described above may be performed. That is, since the first weight block is an A-type weight block in the above example, four of the bits in the block are processed in logic 1 as logic 1 and the rest as 0, and the processed bits are decoded into 8 bits of information. do. Since the second weight block is a type B weight block in the above example, seven of the bits in the block are processed in logic 1 as logic 1 in order of the rest, and the processed bits are decoded back into 8 bits of information. That is, even in the decoding process, the odd-numbered weight blocks are recognized as the A-type weight blocks, and the even-numbered weight blocks are recognized as the B-type weight blocks and processed.

In the above-described example, the present invention is encoded by converting an m-bit original block into an n-bit weight block. Here, in encoding an m-bit original block into an n-bit weight block, it is a question of how many n-bit blocks having 1 out of the n-bit blocks are set as weight blocks. For example, in converting an 8-bit original block into an 11-bit transformed block, only 11 bits having 4 or 7 numbers of 1 were selected as weight blocks. The present inventors have found that the number of 1s that can form a weight block among n-bit blocks can be selected using Equation 1 while trying to solve this problem.

Where m is the number of bits in the original block, n is the number of bits in the transform block, and a is the number of bits with logic 1 in the transform block. That is, when 8-bit original block is to be converted into 11-bit transformed block, 2 ⁸ = 256, so that a value having 256 or more values is searched for. In this case, the value a having ₁₁ C _a <256 may be 4 or more, that is, 4, 5, 6,... In the above example, the smallest of these values, i.e., 4, is set to the actual a value. The reason for setting the value a to 4 will be readily apparent to those skilled in the art from the following description.

Meanwhile, the value a set by Equation 1 relates to one weight block (for example, A type) of two weight blocks (type A or B). The number of bits ₁₁ C _t in the weight block of another type (B type) can be detected by the following equation (2).

t = n-a

Applying Equation 2 to the above example, it can be seen that the value t of the B-type weight block is 7 (or 6).

The weight block of the present invention can be used to obtain an effect of improving the code rate. However, the present invention and the inventor of the present invention have continuously devised a method of further improving the code rate by using the weight block.

The present invention is based on the results of the research, and an object of the present invention is to provide a method for encoding and decoding a weight block, which improves a code rate by encoding a plurality of original blocks into group weight blocks.

In order to achieve the above object, the present invention provides a weight block encoding method for converting an m-bit original block into odd first and second weight blocks of n (n> m) bits, wherein the original block is n + 1 each. Grouping and converting one predetermined circle block among the grouped n + 1 circle blocks into a first weight block; The remaining original blocks in the group are converted into first or second weight blocks according to the logic of the bits in the first weight block.

The present invention also provides a method of decoding n weight blocks into n + 1 original blocks, and adds bits in the weight block to determine which type of weight block is the first or second type weight block according to the addition value. Detecting; Forming a new weight block by configuring the n-bit logic corresponding to the type information of the weight blocks; Decoding the weight blocks including the new weight block into original blocks according to the type.

1 is a view for explaining grouped weight block encoding and decoding according to the present invention;

2 is a block diagram of a grouped weight block encoding apparatus according to the present invention;

3 is a block diagram of a grouped weight block decoding apparatus in accordance with the present invention.

<Description of the code | symbol about the principal part of drawing>

1: analog / digital conversion circuit 3: buffer circuit

5 switch 7 control unit

9 type A conversion circuit 11 type B conversion circuit

21: buffer circuit 23: control unit

24: switch 25: A type reverse conversion circuit

27: B type reverse conversion circuit

Hereinafter, with reference to the accompanying drawings will be described the principle of the present invention and the embodiments will be described in detail sequentially.

The present invention is characterized in that the information in the weight block is not directly provided to a recording medium or the like but is grouped and processed by the number of bits of the weight block. This feature will be described with reference to FIG. 1.

FIG. 1 illustrates a process of encoding 8-bit original blocks OB1-OB12 into 11-bit weight blocks BB1-BB11. The figure shows the weight blocks BB1-B12 corresponding to the original blocks OB1-OB12 using arrows.

In the prior art, in order to encode the original blocks OB1-OB12 into the weight blocks BB1-BB12, the odd-numbered one blocks OB1, OB3, OB5, OB7, OB9, and OB11 are, for example, A-type weight blocks (1 out of 11 bits). The four even blocks (OB2, OB4, OB6, OB8, OB10, OB12) are encoded as, for example, B type weight blocks (blocks having 1 of 7 out of 11 bits). Code like this:

That is, as shown in FIG. 1, the first circle block OB1 is encoded as, for example, an A-type weight block BB1. When encoding from the second circle block OB2 to the twelfth block OB12 into a weight block, the second circle block OB2 is encoded into a type A or B weight block according to a logic state of 11 bits in the type A weight block BB1. For example, if the 11-bit logic state in the type A weight block BB1 is shown, and logic 1 is set to type A and logic 0 is set to type B, the second circle block OB2 is the type A weight block BB2. The third original block OB3 is encoded into a B-type weight block BB3. In this way, a type (type A or type B) for weight block coding the second to twelfth circle blocks OB2-OB12 may be determined.

Although the encoded weight blocks BB1-BB12 may be obtained through the above-described process, in the present invention, only the information of the weight blocks BB2-BB12 except for the first weight block BB1 among the weight blocks BB1-BB12 may be obtained. Record on the recording medium or transmit to the outside. That is, the weight block BB1 is erased without using it. However, the information of the weight block BB1 may be recovered through the decoding process described below.

The decoding process is as follows.

First, the information of the weight blocks BB2-BB12 is divided into 11 bit units, that is, the weight blocks B2-BB12, and the values of the bits in the weight blocks B2-B12 are added for each block. In this case, since there are four 1s in the A-type weight block (BB2, BB6, BB9, BB12 in the above example), a value of 4 will appear, but the B-type weight block (BB3, BB4, BB5, BB7, BB8, BB10, BB11) will have a value of 7. In addition, the type (type A or B) of the weight blocks BB2-BB12 is determined by the bit logic in the weight block BB1, and since the weight block BB1 is A type, 1 is 4, so the weight block BB2 -BB12) means that there are 4 types A and 7 types B. Therefore, the internal bit values are added for each weight block BB2-B12, and the seven weight blocks having a relatively large added result value are type B, and the four weight blocks having a relatively small added result value are type A. Can be set.

Types of the weight blocks B2-B12 may be detected through the above-described process, and the detected types are arranged in the order of the weight blocks BB2-B12 as follows.

A (BB2), B (BB3), B (BB4), B (BB5), A (BB6), B (BB7), B (BB8), A (BB9), B (BB10), B (BB11), A (BB12)

Here, if type (A) is set to logic 1 and type (B) is set to logic 0, it is as follows.

1 (BB2), 0 (BB3), 0 (BB4), 0 (BB5), 1 (BB6), 0 (BB7), 0 (BB8), 1 (BB9), 0 (BB10), 0 (BB11), 1 (BB12)

This logic value is equal to the logic value of the bits in the weight block BB1 as can be seen in FIG. Therefore, the bit information of the weight block BB1 that has been erased in the encoding process may be restored.

When bit information and type information in the weight blocks BB1-BB12 are detected through the above-described process, the process of decoding the weight blocks BB1-BB12 into the original blocks OB1-OB12 can be performed in the same manner as in the related art. Since a person skilled in the art of the present invention will easily know, detailed description thereof has been omitted.

Hereinafter, embodiments of the present invention having the above-described principle will be described in detail.

2 shows a schematic block diagram of an apparatus for carrying out the invention.

2 is a schematic block diagram of an apparatus for weight block coding according to the present invention. As shown, the weight block encoding apparatus includes an analog / digital conversion circuit (hereinafter referred to as an A / D conversion circuit) 1, and the A / D conversion circuit 1 converts an input video signal into a digital signal and buffers the same. To the circuit 3. The buffer circuit 3 is composed of m (8 in the example of Fig. 1) outputs the input digital signal in m-bit units, and outputs a predetermined timing signal each time the digital signal is output in m-bit units. Here, the m-bit signal becomes a one block.

The original block of the buffer circuit 3 is provided to the switch 5, and the switch 5 provides the original block to the A-type conversion circuit 9 or the B-type conversion circuit 11 under the control of the control unit 7. do. The A-type conversion circuit 9 converts the m-bit original block into n-type weighted blocks of n (11 in the example of FIG. 1) and outputs it. The B-type conversion circuit 11 outputs the m-bit original block n. The result is converted into a B-type weight block of bits and output.

The control unit 7 has counting means therein to count the timing signals provided from the buffer circuit 3, and the counting means is configured to be reset when n + 1 timing signals are applied. Here, n + 1 is set to a value larger than 1 by the number of bits of the weight block when the original block of m bits is converted into a weight block.

The control unit 7 controls the switch 5 when the count value of the internal counting means is 1, that is, when the first timing signal is applied after the counting means is reset, and provides the original block to the A-type conversion circuit 9. do. The control unit 7 also controls the internal counting means until the count value becomes 2 to n + 1 (that is, until the n + 1 th timing signal is applied from the second timing signal after the internal count value is reset). The original block of the buffer circuit 1 is an A or B type conversion circuit 9, 11 according to the logic order of bits in the A type weight block provided from the A type conversion circuit 9 each time a timing signal is applied. It controls the switching of the switch 5 so that it can be provided. That is, as shown in FIG. 1, the first circle block OB1 is converted into the A-type weight block BB1 by the A-type conversion circuit 9, and the subsequent circle blocks OB2-OB12 are A-type conversions. The circuit 9 or the B type conversion circuit 11 converts the A or B type weight blocks BB2-BB12.

A switch 13 is connected to the output of the A-type conversion circuit 9, the switch 13 is turned on / off under the control of the control unit 7, and the control unit 7 has a coefficient value of the internal counting means 2. To n + 1 (i.e., until the n + 1 th timing signal is applied from the second timing signal after the internal count value is reset), but the first timing signal is turned on. If provided, switch 13 is turned off. That is, in the example of FIG. 1, the bits of the first weight block BB1 of type A are provided to the controller 7, but the balance of type A by the type A conversion circuit 9 of the second and subsequent weight blocks BB11 is provided. The blocks BB2, BB6, BB9 and BB12 are not provided to the control unit 7 but are output directly.

The weight block (BB2-BB12 in the example of FIG. 1) outputted through the device switch 13 or the B-type conversion circuit 11 of the present invention having the above-described configuration is a magnetic medium or a hologram medium by recording means (not shown). Is written on.

FIG. 3 shows an example of an apparatus for decoding the original blocks OB1-OB12 from the weight block (BB2-BB12 in the example of FIG. 1) recorded on the magnetic recording medium or the hologram medium by the apparatus of FIG. As shown, the decoding device includes a buffer circuit 21, and the buffer circuit 21 transmits bit signals of a weight block that is n bits (11 in the example of FIG. 1) to n buffers B2 to Bn + 1. Store sequentially.

The control unit 23 is connected to the buffers B2 to Bn + 1, and the control unit 23 adds internal bits of the weight blocks stored in the buffers B2 to Bn + 1 for each buffer (by weight block). In accordance with the addition value, it is determined whether the weight block in the buffers B2 to Bn + 1 is A type or B type. That is, according to the example of FIG. 1, four 1s exist in the A-type weight blocks BB2, BB6, BB9, and BB12, and seven in the B-type balance blocks BB3, BB4, BB5, BB7, BB8, BB10, and BB11. Since 1 is present, as a result of adding bits of the weight block, seven weight blocks having a relatively large addition value are determined as B type, and four weight blocks having a relatively small addition value are determined as A type.

The control unit 23 which ends this process writes a logic value (for example, type A is 1 and type B is 0) according to the type (type A or B) of the buffers B2 to Bn + 1 in the buffer B1. do. Therefore, the weight blocks (BB1-BB12 in the example of FIG. 1) are stored in the buffers B1 to Bn + 1.

The buffer circuit 21 is further connected with an A-type inverse conversion circuit 25 and a B-type inverse conversion circuit 27 via a switch 24, and the control unit 21 controls the switch 24 to buffer the buffers B1 ˜. The weight blocks of Bn + 1) are provided to the type A inverse conversion circuit 25 or the type B inverse conversion circuit 27 according to the type thereof. Since the type of the weight block in the buffers B1 to Bn + 1 has been detected through the above-described process, the provision of the weight block to the type A inverse conversion circuit 25 or the type B inverse conversion circuit 27 can be simply performed. Those skilled in the art will readily know. Meanwhile, the type of the weight block in the buffer B1 is not detected through the above-described process, and depends on what type the weighting device BB1 encodes in the encoding apparatus of FIG. 2. In the example of FIG. Since BB1 is coded as A type, the weight block BB1 should be provided to the A type inverse conversion circuit 25.

The type A inverse conversion circuit 25 and the type B inverse conversion circuit 27 convert the input type A and B weight blocks BB1-BBn into corresponding original blocks and output them.

As described above, in encoding the original block into the weight block, since only the original blocks corresponding to the number of bits in the weight block are encoded into the weight blocks, the code rate can be improved. For example, when encoding an 8-bit original block into an 11-bit weight block, the encoding efficiency is 8/11 = 72.2%. However, in the present invention, 96 data bits (96 × 8) and 121 data bits ( 11 × 11), so the code rate is 96/121 = 79.3%. Therefore, it can be seen that the code rate of the present invention is greatly improved.

Claims (6)

A weight block encoding method for converting an m block of original blocks into first and second weight blocks of odd n (n> m) bits, Grouping the original blocks by n + 1 and converting any one original circle block among the grouped n + 1 original blocks into a first weight block; And grouping the remaining original blocks in the group into first or second weight blocks according to logic of bits in the first weight block. The method of claim 1, And a predetermined one of the grouped n + 1 original blocks is the first one among the original blocks in the group. The method according to claim 1 or 2, The weighted block in which one predetermined original block is encoded among the grouped n + 1 original blocks is erased. The method of claim 3, wherein The number of 1s in the first weight block is Is set to a value (a) that satisfies The number t of 1s in the second weight block is set to a value satisfying t = n-a. A method of decoding n weight blocks into n + 1 original blocks, Adding bits in the weight block to detect which type of weight block is the first or second type according to the addition value; Forming a new weight block by configuring the n-bit logic corresponding to the type information of the weight blocks; And decoding the weight blocks including the new weight block into original blocks according to the type. The method of claim 5, wherein And the type of the new weight block is preset.