JPH0774652A - Data compressing device - Google Patents

Abstract

PURPOSE:To transfer and store the same amount of data from the data compressor to the other device. CONSTITUTION:A data compressor 11 is provided with the 1st-Nth compression processing parts 231-23N which perform compression processing in dissimilar ways and the 1st-Nth FIFO memories 241-24N which input the corresponding output. A CPU 22 monitors the point when the half bit of each FIFO memory changes to '1' and judges that the compression system of the FIFO memory 24 which is changed to '1' last is the most effecient compression system and stores the system and its storage data in a storage device 13 by each cycle. The Nth compression processing part 27N performs no compression and it can be selected as the most efficient one according to the contents of data.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a data compression apparatus used for compressing various types of data such as image data for processing and transmission, and more specifically, it is suitable for use among a plurality of processing methods. The present invention relates to a data compression device that selects a compression method.

[0002]

2. Description of the Related Art It is widely practiced to store data in a storage device such as a magnetic disk, a hard disk or a magnetic tape, and to read and process the data as needed. When it is necessary to store a large amount of data in such a storage device, compression processing is performed on the data before storage. Although various compression processing methods have been proposed, a method that always has a good compression rate for all data has not been developed yet.

For example, in the field of facsimile communication, a variable code length compression system is adopted. According to this method, a relatively long code is assigned to data that has a low appearance probability, while a relatively short code is assigned to data that has a high appearance probability. For this reason, it is possible to compress image data with a high compression rate for data in which data that is said to have a high appearance probability appears at such a high frequency, but for data that does not. In the case of, the effect of compression cannot be expected so much, and in an extreme case where data with a low occurrence probability frequently appears, the amount of data becomes larger than in the case of no compression.

In order to solve such a problem, in Japanese Unexamined Patent Publication No. Sho 62-209948, data to be compressed is divided into appropriate lengths (blocks), and each of these divided data is divided. A plurality of conversions based on different compression conversion rules are simultaneously performed in parallel. Then, as a result of these conversions, the conversion data according to the compression conversion rule having the shortest length is added to the data indicating the type of the conversion rule and transmitted to the receiving side. That is, according to the proposed data compression apparatus, the minimum compression method can be selected for each divided block to compress and transmit the data.

[0005]

However, according to the proposed data compression apparatus, it is necessary to perform a plurality of conversions in parallel at the same time to determine the amount of data of each block after compression. It cannot be judged unless the results are compared. Therefore, on the data compression device side, a counter is prepared to count the data compressed based on the respective compression conversion rules each time,
As a result, the minimum amount of data is transmitted each time, and the amount of data transmission is not constant.
Therefore, on the device side, a relatively large capacity buffer for temporarily storing the compressed data is required, and in addition, a circuit such as the counter as described above is required, which complicates the configuration of the device. was there.

Further, even in the receiving side device for receiving the data transmitted through the data compression device, the amount of data sent at one time is indefinite, so that a relatively large receiving buffer is required. In addition to this, there is a problem that the process for checking a data transmission error becomes complicated.

SUMMARY OF THE INVENTION An object of the present invention is to provide a data compression apparatus capable of sequentially executing data compression processing while arranging the data amount after compression into substantially constant amounts.

Another object of the present invention is to avoid the situation where the data amount is increased even for data that is unfavorable for compression, while the data amount after processing is adjusted to a substantially constant amount and the compression process is sequentially performed. It is to provide a data compression device that can be executed.

[0009]

According to a first aspect of the present invention, a plurality of compression processing means for performing parallel compression processing on the same data to be compressed by different compression processing methods, respectively.
A plurality of FIFO memories (first-in first-out memory) for storing the processing results of these compression processing means for each compression processing method, and when a certain amount of data less than the amount corresponding to the full capacity is stored in each of these FIFO memories, Constant storage amount detection means for sequentially detecting, and all these FIFOs
When the constant storage amount detection means detects that the predetermined amount of data is stored at the end of the memory, the last F
When there is data to be compressed at the stage when the data is output from the last FIFO memory, the compressed data output means for adding the data indicating the compression method to the data stored in the IFO memory and outputting the data is output. The data compression apparatus is provided with a compression repetition control means for clearing the contents of the memory and sequentially repeating the data processing by each means.

That is, according to the first aspect of the invention, a plurality of compression processing means are prepared and the same data to be compressed is compressed in parallel by different compression processing methods.
Then, these results are sequentially stored in the corresponding FIFO memory. The fixed storage amount detecting means is based on these FIFs.
It is detected that the compressed data stored in the O memory reaches a certain storage amount before reaching the full capacity. Then, assuming that the compression processing method corresponding to the FIFO memory in which the detection is performed last has the highest compression efficiency, data indicating the compression method is added to this and output. By repeating such a compression process until the end of the data, it is possible to repeatedly transfer the fixed storage amount or almost the same amount of compressed data to another place or store it in a storage means such as a magnetic disk. It will be possible.

According to the second aspect of the present invention, a plurality of compression processing means for performing parallel compression processing on the same data to be compressed by different compression processing methods, processing results of these compression processing means and compression processing are provided. A plurality of FIFO memories that store uncompressed data for each processing method, and a fixed storage amount that sequentially detects when a fixed amount of data less than the full amount is stored in each of these FIFO memories. The detection means and the last FI when the fixed storage amount detection means detects that the fixed amount of data is stored at the end of all the FIFO memories.
Data output means for adding data indicating the processing method to the data stored in the FO memory and outputting the data, and a final FI.
When there is data to be compressed when the data is output from the FO memory, the data compression apparatus is provided with a compression repetition control means for clearing the contents of these FIFO memories and sequentially repeating the data processing by each means. .

That is, according to the second aspect of the present invention, a plurality of compression processing means are prepared and the same data to be compressed is compressed in parallel by different compression processing methods.
Then, these results are sequentially stored in the corresponding FIFO memory. At this time, data that is not subjected to compression processing is also prepared, and this data is sequentially stored in a dedicated FIFO memory. The constant storage amount detecting means detects that the compressed or non-compressed data stored in these FIFO memories reaches a constant storage amount before reaching full. Then, assuming that the compression processing method or the non-compression processing corresponding to the FIFO memory in which this detection is performed last has the smallest data amount, data indicating the compression method or the non-compression is added to this. Output. By repeating such a compression process until the end of the data, it is possible to repeatedly transfer the fixed storage amount or almost the same amount of compressed data to another place or store it in a storage means such as a magnetic disk. It will be possible.

[0013]

EXAMPLES The present invention will be described in detail below with reference to examples.

FIG. 1 shows an outline of the configuration of a data compression apparatus according to an embodiment of the present invention. The data compression device 11 is a storage device 1 including a host computer 12 as a data supply source and a magnetic disk device for storing compressed data (hereinafter referred to as compressed data).
It is connected between 3 and. Then, a predetermined amount of block data 14 is obtained from the host computer 12, compressed, and stored as compressed data 15 in the storage device 13.

The data compression device 11 uses the block data 1
A host interface circuit 21 for inputting 4 and a CPU (central processing unit) 2 for controlling the entire device 11
2 and the host interface circuit 21 and compresses data by different compression methods.
Compression processing units 23 ₁ to 23 _N and _first to _Nth compression processing units 23 ₁ to 23 _N corresponding to the _first to _Nth compression processing units 23 ₁ to 23 _N and sequentially storing processed data.
_{1 to} 24 _N and these first to Nth FIFO memories 24
Compressed data 1 selected from _{1 to} 24 _N output
The storage device interface circuit 25 outputs the data to the storage device 13 as 5.

Here, the first to Nth compression processing units 23 ₁ to
Compression processing section 23 _N of the last N-th of 23 _N does not perform compression processing, and has a circuit which outputs the input data. Depending on the contents of the data, the amount of data may be smaller if compression is not performed. In such a case, the data itself that is not compressed can be stored in the storage device 13 via the storage device interface circuit 25. To do so.

The first to Nth FIFO memories 24 _{1 to} 24 _N used in the present embodiment are half (hal).
f) A bit output terminal is provided. These half-bit output terminals output the signal "1" instead of the signal "0" from the half-bit signals 27 ₁ to 27 ₁ when the amount of data corresponding to exactly half the memory capacity is written.
It is designed to output as 27 _N. These half bit signals 27 _{1 to} 27 _N are taken into the CPU 22. Then, the first to Nth FIFO memories 24 _{1 to} 24
Of the _N, the one fetched latest is determined to have the highest compression ratio, and the storage device interface circuit 25
Is instructed to select the compressed data.

At this time, the first to N-th FIFO memories are
Mori 24₁~ 24_NStorage of data for
FIFO corresponding to the compression method with the highest compression rate
Read the stored compressed data from the memory 24
Once done, all these FIFO memories 24₁~ 2
Four_NThe stored contents of are cleared and the same operation is performed again.
The data compression processing is performed. Also
In this embodiment, the block data 14 for one block is used.
First to Nth FIFO memory 24₁~ 24 _NNote
Set the amount of data equivalent to exactly half of the capacity
I am trying.

FIG. 2 is for explaining the control operation on the host computer side for sending data to the data compression apparatus having the above-mentioned configuration. Host computer 12
When the data is to be stored in the storage device 13, requests the data compression device 11 to start the data storage operation (step S101). When a data transmission request is sent from the data compression device 11 in response to this (step S102; Y), the host computer 1
2 sends block data 14 as one block of data to the data compression device 11 (step S1).
03). Then, it is checked whether or not there is still data to be sent after this (step S104). Then, if the data to be transmitted still exists (Y), the process returns to step S102 and the block data 14 is repeatedly transmitted every time a transmission request is received.

On the other hand, when the host computer 12 completes the transmission of all the data as a result of the repeated transmission of the block data 14 (step S104; N), a transmission end notification is issued to the data compression device. 11
(Step S105), and the processing for storing data in the storage device 13 is completed (end).

FIG. 3 shows a state of control on the side of the data compression apparatus in accordance with the above operation of the host computer. The data compression device 11 includes a host computer 12
It waits for the request to start the data storage operation from (step S201). Then, when this request arrives (Y), the compression methods to be adopted are designated to all the compression processing units 23 ₁ to 23 _N-1 excluding the _Nth compression processing unit 23 _N (step S202). The reason why the compression method is not designated for the N-th compression processing unit 23 _N is that the circuit configuration is such that the compression processing is not performed for this in the present embodiment. You can specify the method.

The compression method is designated by the CPU 2 shown in FIG.
2 assigns each logical content of the compression method written in its own ROM (read only memory) (not shown) to the respective compression processing units 23 ₁ to 23 _N-1 . What kind of compression method to select and allocate
The CPU 22 determines in accordance with the logic written in the ROM using the information indicating the content of the data sent by the host computer 12 prior to the start of sending the block data 14 as a key. Of course, the host computer 12 may instruct which compression method to select and allocate. When the first to (N-1) th compression processing units 23 ₁ to 23 _N-1 can provide a sufficient number of compression methods in parallel in relation to the numerical value N, the compression method is selected. It is also possible to omit the operation of allocating them and to keep these compression methods fixed. In this case, the circuit configuration can be simplified. It is also possible to write each logic of the compression method written in the ROM in the storage device 13 and read it appropriately.

Returning to FIG. 3, the description will be continued. 1st to (N
-1) When the compression method is set for the compression processing units 23 ₁ to 23 _{N- 1 of FIG} .
4 ₁ to 24 _N to reset the contents of (step S20
3). Then, the host computer 12 is requested to transmit the block data 14 (step S204).
After this, the block data 1 is sent from the host computer 12.
When 4 is sent (step S205; Y), the host interface circuit 21 stores this one block of data. Then, this is read by a predetermined unit amount and supplied to the _first to _Nth compression processing units 23 ₁ to 23 _N at the same time, and the data compression processing is performed in parallel by these (step S206). At this time, the Nth compression processing unit 23
_{In N} , the received data is output as it is without being processed.

The processed data respectively correspond to the first
˜Nth FIFO memories 24 _{1 to} 24 _N are supplied (step S207). At this time, it goes without saying that the data output from the compression processing unit 23 having a low degree of compression has a larger data amount than the data output from the compression processing unit 23 having a high degree of compression. At this stage, the CPU 22 outputs the half bit signals 2 of all the FIFO memories 24 _{1 to} 24 _N.
7 ₁ ~ 27 _N to check whether changes to "1" (step S208). All half-bit signals 27
_{In the state where 1 to} 27 _N remains “0” or only part of them changes to “1” (N), the compression processing unit 23 adopting the compression method with the best compression ratio is still supported. Half the amount of data is not stored in the FIFO memory 24.

In this state, the CPU 22 checks whether or not the processing of the block data 14 is completed from the remaining data amount of the host interface circuit 21 (step S209), and if there is still data, returns to step S206. Read the next predetermined unit amount of data,
The same processing as described above is performed. When the processing of the block data 14 sent from the host computer 12 is completed (step S209; Y), it is checked whether or not a transmission completion notification is received from the host computer 12 (step S210). If the notification has not been received (N), the process returns to step S204 to request the host computer 12 to transmit the block data 14 for the next one block.

In this way, the host computer 12
Result of the processing of the block data 14 sent from
As a result, at some point all half-bit signals 27₁~ 27
_NIs changed to "1" (step S208; Y), C
The PU 22 sends a control signal to the memory device interface circuit 25.
Signal, and finally the half-bit signal 27 changes to "1".
Select the contents of the converted FIFO memory 24 and write it down.
The compressed data 15 is stored in the storage device 13 (step
S211). Then, each FIFO memory 24 ₁~ 24
_NIs reset (step S212), the
Return to step S209.

On the other hand, if it is determined in step S210 that the transmission end notification is received from the host computer 12 (Y), whether or not the compressed data is stored in each of the FIFO memories 24 _{1 to} 24 _N. Is determined (step S213). Then, when one block is completed in the state where the compressed data is not stored (N), the whole process is ended as it is.

On the other hand, when it is determined that the host computer 12 side does not send data, these F
When the compressed data is stored in the IFO memories 24 _{1 to} 24 _N (step S213; Y),
It is necessary to select one of these as compressed data and store it in the storage device 13. In this embodiment,
F that uses the same compression method as the one used last time
The IFO memory 24 is selected and its contents are stored in the storage device 13 together with the data indicating the compression method (step S214). As a selection method other than this, a method of excluding the compression method in which the half-bit signal 27 has already changed to "1" at this point and performing selection is also effective.

In this way, the host computer 12
The processing of all block data 14 sent from is ended (END).

FIG. 4 illustrates the above processing. The data 31 desired to be stored in the storage device 13 by the host computer 12 is block data 1 composed of a plurality of blocks (a, b, c, ...) From the viewpoint of efficient transmission.
It is divided into 4. In the figure, the block data 14 from “a” to “j” are classified for convenience. In this case,
First, the block data 14 of “a” is the data compression device 11
And the predetermined compression processing is performed by the _first to _N-th compression processing units 23 ₁ to 23 _N , respectively. However, in this embodiment, the Nth compression processing unit 23 _N does not perform compression.

The amounts of the compressed data 32 _{1 to} 32 _N are not constant as shown in the figure. These are the corresponding F
The data is sequentially written into the IFO memories 24 _{1 to} 24 _N ,
Finally, the FIF in which the half bit signal 27 has changed to "1".
O memory 24 (in this example, the first FIFO memory 24
₁ ) is selected by the storage device interface circuit 25 and stored in the storage device 13. Similarly, the process is repeated up to the last "j" block data 14.

FIG. 5 shows how compressed data is transferred from the data compression device to the storage device. When compression of all data is completed, each FIFO memory 24 ₁
The compression method each time the compressed data 32 is output from the storage device interface circuit 25 from each of .about.24 _N is stored in the header file 41. Since these are stored in the storage device 13 together with the compressed data 42 _{1 to} 42 _M sent each time, the compressed data 42 ₁ to 4 4
It is possible to determine what kind of compression each of the 2 _M has been compressed.

[0033] In the embodiment reads the block data 14 by a predetermined unit amount in step S206, and the reading the results in each of the FIFO memories 24 to 24 _N. Therefore, a situation may occur in which all the half bits become "1" when the amount of data strictly exceeding the half bits is stored in them.
If it is necessary to transfer the compressed data that is exactly equal to half bits to the storage device 13 every time, step S20 is performed.
Instead of storing all the data processed in 6 in each of the FIFO memories 24 _{1 to} 24 _N in a lump, a predetermined amount is stored and the compressed data is stored when the last half bit of the FIFO memory 24 changes to “1”. Should be stopped. In this case, the remaining compressed data is first stored in each of the FIFO memories 24 _{1 to} 24 _N in the next cycle.

Further, in the embodiment, the transfer of the compressed data is started on the condition that the half bits of the _first to Nth FIFO memories 24 _{1 to} 24 _N are set. The FIFO that last reached these quantities, relative to bits that indicate other quantities, such as the ones shown
It is also possible to transfer the contents of the memory 24 to the storage device 13.

Furthermore, although the compressed data is stored in the storage device 13 in the embodiment, it goes without saying that these may be sequentially transferred to another device such as a receiving device.

[0036]

As described above, according to the first aspect of the present invention, the same data is processed by each of a plurality of compression methods and stored in the corresponding FIFO memory, and the FIFO memory which reaches the fixed storage amount most recently is reached. We decided to extract the compressed data from, and add the data indicating the compression method to the output. Therefore, the compressed data can be output by an amount equal to or substantially equal to the fixed storage amount, which is convenient when data is transferred or stored by repeating these processes. In particular, the present invention can be effectively utilized in a system in which the amount of data accessed at one time is limited. Further, in the present invention, FI
When the amount stored in the FO memory reaches a predetermined value, one unit of compression is terminated, so there is no need for a counter for measuring which compression method has achieved the most efficient compression. By simplifying the circuit configuration of the device, it is possible to reduce cost and improve reliability.

According to the second aspect of the present invention, the FIFO memory for storing the compressed data and the FIFO memory for storing the same data as the compressed data are also prepared. It is possible to obtain the same effect as the effect according to the invention described in claim 1, and it is possible to prevent the disadvantage that the data amount becomes larger than the unprocessed data depending on the state of the data.

[Brief description of drawings]

FIG. 1 is a block diagram showing a connection state of a data compression device, a host computer that transfers data to the data compression device, and a storage device that stores compressed data according to an embodiment of the present invention.

FIG. 2 is a block diagram showing the operation of the host computer for the data compression apparatus in this embodiment.

FIG. 3 is a flowchart showing the operation of the data compression apparatus corresponding to FIG.

FIG. 4 is an explanatory diagram showing a state of data compression processing by a data compression apparatus in the present embodiment.

FIG. 5 is an explanatory diagram showing a relationship between compressed data and a head file indicating the compression method in the present embodiment.

[Explanation of symbols]

11 ... Data compression device, 12 ... Host computer (data transmission source), 13 ... Storage device, 14 ... Block data, 22 ... CPU, 23 ... Compression processing unit, 24 ... FIFO
Memory, 25 ... Storage device interface circuit, 32 ...
Compressed data

Continuation of front page (51) Int.Cl. ⁶ Identification code Office reference number FI Technical display area H04N 7/24

Claims (2)

[Claims] 1. A plurality of compression processing means for parallelly processing the same data to be compressed by different compression processing methods, and a plurality of compression processing means for storing the processing results of these compression processing means for each compression processing method. FIFO memories, a fixed storage amount detecting means for sequentially detecting when a fixed amount of data less than a full amount is stored in each of the FIFO memories, and the fixed amount at the end of all the FIFO memories. When the constant storage amount detecting means detects that the data of (1) is stored, the compressed data output means for adding the data indicating the compression method to the data stored in the last FIFO memory and outputting the data from the last FIFO memory. When the data to be compressed exists at the stage when
A data compression apparatus, comprising: a compression repetition control means for clearing the contents of the memory and sequentially repeating the data processing by each means. 2. A plurality of compression processing means for performing parallel compression processing on the same data to be compressed by different compression processing methods, processing results of these compression processing means and pre-compression before compression processing. A plurality of FIFO memories for storing data according to processing methods, and a fixed storage amount detecting means for sequentially detecting a fixed amount of data less than a full amount in each of these FIFO memories, and a total storage amount detecting means Data output for adding the data indicating the processing method to the last data stored in the FIFO memory when the constant storage amount detecting means detects that the predetermined amount of data is stored at the end of the FIFO memory Means, and when the data to be compressed is present at the stage when the data is output from the last FIFO memory, these FIFO memories
A data compression apparatus, comprising: a compression repetition control means for clearing the contents of the memory and sequentially repeating the data processing by each means.