JPH01101740A - System for compressing data - Google Patents

Abstract

PURPOSE:To widely improve a compressing rate by converting a code for the unit of character codes and setting a converting code to correspond to an appearing frequency for the unit of the character codes. CONSTITUTION:When data which include the character code composed of plural bites are code-converted and compressed, data having a character code assembling part 1 and one bite unit are assembled to the data having the character code unit. By a code converting part 2, conversion is executed to a variable length code for the unit of the converted character codes. Thus, not only to one bite code but also to the character code of the plural bites, the variable length code of the optimum length code to correspond to the frequency can be used and the data compressing rate can be widely improved.

Description

[Detailed description of the invention] [Table of contents] Overview Industrial field of application Prior art (Figure 6) Problems to be solved by the invention Means for solving the problems (Figure 1) Working examples (a) ) Explanation of one embodiment (Figures 2, 3, 4, and 5) (b) Explanation of other embodiments Effects of the invention [Summary] Converting data in bytes into variable length codes Regarding the data compression method for data compression, the purpose is to compress data containing character codes consisting of multiple bytes with a high compression rate, and converting the data containing character codes consisting of multiple bytes to compress data. a character code assembling unit that assembles data in units of 1 byte into data in units of character codes;
and a code conversion unit that converts the converted character code into a variable length code in units of units.

[Industrial application field]

The present invention relates to a data compression method that converts data in units of bytes into variable length codes and compresses the data.

Data compression techniques are used to reduce the amount of storage and transmission when storing and transmitting data.

In particular, in data compression for text data, a method of converting character codes into variable length codes is often used.

In data compression using such code conversion,
A high data compression rate is desired.

[Conventional technology]

FIG. 6 is an explanatory diagram of the prior art.

As shown in FIG. 6(A), original data is input in 1-byte units, is latched by a latch circuit R, and then converted into a variable-length code by a code conversion circuit DC that converts variable-length codes in 1-byte units. Perform compression.

The code conversion circuit DC converts each 1-byte code into a variable-length code having a code length corresponding to the denomination of each 1-byte code, and stores a comparison table of each 1-byte code and the corresponding variable-length code in the ROM ( (read-only memory).

For this reason, as shown in FIG. 6(B), a Kanji code consisting of two bytes is also decomposed into bytes, and code conversion is performed on the decomposed one-byte code.

[Problem that the invention seeks to solve]

Generally, data whose content is a Kanji code, which is a 2-byte code, has a high degree of redundancy and is expected to have a high data compression rate.

However, in the conventional technology, the character code, which is originally composed of multiple bytes, is decomposed into each byte, which is the code conversion unit, and the code is converted, so it is impossible to increase the compression rate of the character code. There was a problem.

In other words, in the conventional technology, a variable-length code with a high compression rate can be used for a 1-byte code because code is converted in units of 1-byte, but a multi-byte code is a combination of multiple bytes of conversion code of a 1-byte code. This is because the code becomes a compression code, so it is not possible to use a conversion code that is optimal for multi-byte codes.

In particular, when a variable-length code with a small code length is assigned to a frequently occurring code, a variable-length code related to the multi-byte code itself is not assigned, and therefore the compression ratio does not improve even if the multi-byte code is used frequently. was occurring.

SUMMARY OF THE INVENTION An object of the present invention is to provide a data compression method that can compress data including a character code composed of multiple bytes at a high compression rate.

[Means for solving problems]

FIG. 1 is a diagram explaining the principle of the present invention.

In the figure, 1 is a character code assembly unit, which assembles data in units of 1 byte into data in units of character codes, and 2 is a code conversion unit, which converts converted character codes into variable length codes. It is something that converts.

[Effect]

In the present invention, data is assembled in character code units and code conversion is performed in character code units, so variable length codes can be set in character code units.

Therefore, a variable length code having an optimum code length depending on the frequency can be used not only for a 1-byte code but also for a multi-byte character code, and the data compression rate can be greatly increased.

〔Example〕

(a) Description of an Embodiment FIG. 2 is a block diagram of an embodiment of the present invention, and FIG. 3 is an explanatory diagram of a conversion code of an embodiment of the present invention.

In FIG. 2, the same components as those shown in FIG.
12 is a control circuit that latches each byte of original data, and 12 is a control circuit that latches the original data for each byte.
), determines whether it is a 1-byte code or a 2-byte code from the shift out code (SO), controls the latch circuit 10.11, and outputs a character code length (1 byte, 2 bytes) signal, 20 is a code 21 is a conversion ROM which stores variable length codes allocated to part of the 1-byte code and 2-byte code; 21 is a judgment conversion circuit;
If the given character code is every 2 bytes, determine whether it is a 2nd level kanji, and if it is a 2nd level kanji, use the code conversion ROM
20 to generate a conversion code.

Reference numeral 3 denotes a 1-byte matching circuit, which matches the variable length code from the code conversion circuit 2 to 1-byte length according to the code length signal, and outputs the matched code.

The conversion code will be explained using FIG. 3.

First, among the 1-byte codes (EBCDIC codes),
The 128 frequently occurring codes are assigned to an 8-bit code, and the first (most significant) bit is set to "0".

Next, 64 out of the remaining 1-byte codes (128 codes) and 192 2-byte codes (JEF codes are used here) that appear frequently, such as hiragana, katakana, numbers, etc.
56 bits are assigned to a 10-bit code, and the most significant 2 bits are set as "IO".

A total of 1536 characters, including the remaining 64 1-byte codes and 1472 relatively frequently used first-level Kanji characters, are assigned to 13-bit codes. The most significant 4 bits of this 13-bit code are "1100", "1101" or "1110".
”.

Furthermore, the remaining 2-byte codes, that is, the remainder of the first-level Kanji characters, the second-level Kanji characters, etc., are assigned to 19-bit codes.

This 19-bit code is a 16-bit (2-byte) code with 3 bits "111" added to the upper part, and the most significant 4 bits are 1111.

For the first level kanji of the 1-byte code and 2-byte code, the corresponding assignment conversion code is stored in the code conversion ROM 20 (.

However, among the first level kanji, those with 19-bit code are
The conversion code of the Kanji code is not stored in the code conversion ROM 20, but only the bit length is stored.

Therefore, part of the 1-byte code and the 2-byte code is converted into 8 bits 1 to 1 by using the code conversion ROM 20. 10 bits, 1
The remaining 2-byte codes, which are converted to 3-bit codes and are used less frequently, have the most significant bit of the JEF code set to “1”.
By using the fact that “111” is added to the original 2-byte code
This can be done by adding 3 bits and converting to 19 bits.

This makes it possible to limit the number of codes that are subject to code conversion using the code conversion table (ROM) 20, and eliminate the need to store conversion codes for level 2 kanji, etc. in the code conversion table. The size of the code conversion table for decoding (compression) and encoding (decompression) can be reduced.

FIG. 4 is an explanatory diagram of the operation of one embodiment of the present invention.

When a JEF code is used as a Kanji code, a shift-in (S) code and a shift-out (SO) code are used to identify the JEF code (2-byte code) and the EBCD IC code (1-byte code).

The SOC code indicates switching from a 2-byte code to a 1-byte code, and the Sl code indicates switching from a 1-byte code to a 2-byte code.

Therefore, when original data is input to the control circuit 12 along with the clock from the left in FIG.
If it is byte data, the input data is set in the latch circuit IO and a character code length signal (1 byte) is sent to the code conversion circuit 2.

On the other hand, if it is 2-byte data, the first byte of the input data is set to the latch circuit 10, and the second byte is set to the latch circuit 10.
After setting to 1, a character code length signal (2 bytes) is sent to the code conversion circuit 2.

The code conversion circuit 2 receives a 2-byte width character code,
If the character code length signal indicates 1 byte, refer to the code conversion ROM 20 with a 1-byte character code, use the 8-bit, IO-bit, or 13-bit code corresponding to that character code as the conversion code, and further change the code length to 1. Output to byte matching circuit 3.

Conversely, if the character code length signal indicates 2 bytes,
Conversion circuit 2 that determines whether the character code of the byte is a level 2 kanji
1, and if the character is other than the second level kanji, the code conversion ROM 20 is referred to using the 2-byte character code.

At this time, as mentioned above, for those to which a 10-bit code or 13-bit code has been assigned, a 10-bit or 13-bit conversion code is output from the code conversion ROM 20.
The code length is also output.

On the other hand, for unallocated characters and second-level kanji characters, the first 2-byte character code is 3 in the judgment conversion circuit.
A bit "111" is added, the code is converted into a 19-bit code, and a 19-bit long signal is also output.

The 1-byte matching circuit 3 calculates 1 from the conversion code and code length.
Matches compressed data in bytes and outputs compressed data in bytes in parallel.

In this way, 2-byte codes can be converted into 2-byte code units, and 1-byte codes can be converted into 1-byte code units, that is, character code units.

In the assignment of conversion codes shown in FIG. 3, the conversion code for each character code is unique in that only one exists.

Therefore, shift-in and shift codes as input mode transition codes can be omitted from the compressed data, and -plane compression efficiency is improved.

FIG. 5 is a detailed explanatory diagram of the present invention.

Figure 5 shows the number of converted code pits and the code conversion for each character code (2 bytes) when converting each character code 1 byte by byte for the 18 character character code "About data compression for kanji code" This shows the number of conversion code bits at the time.

Here, for example, the 2-byte character code for the character "Kan" is 1-byte code "B4" and 1-byte code "C".
1".

For example, for the character "su", if you convert the code for each byte, the first byte "A4" will be converted to a 5-bit code, the second byte "B9" will be converted to a 6-bit code, and a total of 11 Bits required.

On the other hand, if the code is converted for each character code, only 10 bits are required.

In this way, the total number of bits for 18 characters is 213 bits when code is converted for each byte, and the character code (2 bytes)
Each code conversion requires 195 bits, and in this example, the compression efficiency increases by about 10%.

For example, if the original data is compressed and stored in an external storage device, the amount of storage will be reduced by 10%, or conversely, 10% more data can be stored.

When restoring data, the code length is determined by checking the upper 1 to 4 bits of the allocation conversion code shown in FIG.

(b) Description of other embodiments In the above embodiments, a 2-byte code is used as an example, but the character code may be any number of bytes, and is not limited to a 2-byte code. Not limited to the one shown in the figure, various cord lengths and types can be adopted as necessary.

Furthermore, a 19-bit long code may be stored in the ROM 20, and various codes may be used depending on the required function.

Although the present invention has been described above using examples, the present invention can be modified in various ways according to the gist of the present invention, and these are not excluded from the present invention.

〔Effect of the invention〕

As described above, according to the present invention, since code conversion is performed in character code units, a conversion code can be set in accordance with the appearance frequency in character code units, and the data compression rate can be significantly improved.

In particular, for data compression that includes kanji codes as content,
High compression ratio can be achieved.

[Brief explanation of the drawing]

Fig. 1 is an explanatory diagram of the principle of the present invention, Fig. 2 is a configuration diagram of an embodiment of the invention, Fig. 3 is an explanatory diagram of the conversion code in Fig. 2, and Fig. 4 is an explanation of the operation of an embodiment of the invention. 5 is a detailed explanatory diagram of the present invention, and FIG. 6 is an explanatory diagram of the prior art. In the figure, 1--character code assembly section, 2-- code conversion section.

Claims (1)

[Claims] In a data compression method that converts and compresses data that includes a character code made up of multiple bytes, there is a character code assembly that assembles data in units of 1 byte into data in units of character codes. A data compression method comprising: a code conversion unit (1); and a code conversion unit (2) that converts the converted character code into a variable length code in units of units.