Error correcting apparatus

Abstract

An error correcting apparatus including a syndrome calculator for calculating syndromes from input data added with an error correction code, a data memory which is connected in parallel with the syndrome calculator for storing the input data, a syndrome memory to store syndromes calculated by the syndrome calculator, and an error correction processor for executing error correction of the data in the data memory by using the syndromes stored in the syndrome memory. The error correction code is provided in both first and second directions when data is two-dimensionally arranged, and the error correction processor obtains values in which the result of an arithmetic operation for error correction in at least one of the first and second directions is added to syndromes in the other direction. An error position within the input data is detected and corrected again on the basis of the values.

Description

BACKGROUND OF THE INVENTION

1. Field of the Invention

The present invention relates to an error correcting apparatus and, more particularly, to a method and apparatus for correcting errors a data to which an error correction code was added.

2. Related Background Art

Hitherto, a product code has been known as a form of an error correction code having a two-dimensional arrangement. FIG. 3 is an arrangement diagram of a product code. In the diagram, the portion of k.sub.1 .times.k.sub.2 corresponds to the original digital information having a two-dimensional arrangement. A predetermined check code m.sub.1 is added to every row in a k.sub.1 direction shown by an arrow A and a code c.sub.1 is constructed. On the other hand, a predetermined check code m.sub.2 is added to every column in a k.sub.2 direction shown by an arrow B and a code c.sub.2 is constructed. The code c.sub.1 is the (n.sub.1, k.sub.1) code and the code c.sub.2 is the (n.sub.2, k.sub.2) code.

A sole decoding method in the lateral or vertical direction will be first described as a decoding method of such a code. The error correction decoding method generally includes the following four steps.

step 1) Syndrome calculation step 2) Calculation of the error position polynomial and error evaluation polynomial step 3) Estimation of the error position and error value (size) in accordance with the results of the calculation in step 2 step 4) Execution of the error correction of the estimated position and size

In step 1), since syndromes are obtained as values of the polynomial of the (n-1) order in which all of the symbols of the number corresponding to a code length n of the code are used as coefficients, the arithmetic operation to sequentially input all of the symbols is executed. Therefore, n/S clocks are necessary when the calculation is executed S times in parallel for at least one syndrome. On the other hand, in many cases, S is generally set to 1 to reduce the circuit scale. In this case, the period of time corresponding to n clocks is necessary. With respect to steps 2) to 4) as well, in many cases, the processes are generally executed within the period of time of n clocks.

On the other hand, in the case of the product code construction, a series of error corrections are once executed with respect to the code words of, for instance, n.sub.2 (=k.sub.2 +m.sub.2) rows in the A direction and a series of error corrections are once performed with regard to the code words of k.sub.1 columns in the B direction. Due to this, the error correcting capability can be raised as compared with the case of, for instance, the construction in only the A direction.

Further, there has been known a repetitive decoding method whereby the error corrections in both of the A and B directions as mentioned above are further repetitively executed to the codes of the product code construction. According to such a method, the error correcting capability can be further raised as compared with the case where the error corrections in the A and B directions are executed once at a time as mentioned above.

However, in the conventional repetitive method, the above steps 1) to 4) are repeated with respect to each of the code words (there are n.sub.2 code words having the arrangement in the A direction and there are n.sub.1 code words having the arrangement in the B direction).

Therefore, assuming that the processes are repetitively executed t times for the period of time of n clocks in step 1) and for the period of time of n clocks in steps 2) to 4), the following processing time is needed.

{(n.sub.1 +n.sub.1)n.sub.2 +(n.sub.2 +n.sub.2)k.sub.1 }.times.t=2t(n.sub.1 n.sub.2 +n.sub.2 k.sub.1) (1)

As mentioned above, an extremely long processing time is necessary in the case of executing the repetitive method as mentioned above in order to raise the error correcting capability in the conventional manner.

On the other hand, in any of the A and B directions shown in FIG. 3, in the case of calculating the syndromes, in addition to the problem in the above repetitive method, there is also a problem such that the data shown in FIG. 3 needs to be read out a plurality of times and it is difficult to realize the high speed error correcting operation.

The above problems will now be described with reference to FIGS. 4 and 5.

FIG. 4 is a diagram showing an example of a conventional error detection correction decoding unit.

In FIG. 4, an explanation will be made with respect to the example of the (n-k-2) correction decoding of the data which was double encoded by the code lengths (n.sub.1, k.sub.1) and (n.sub.2, k.sub.2) in the C.sub.1 (lateral) and C.sub.2 (vertical) directions as shown in FIG. 5. The data block including the error correction code (ECC) shown in FIG. 5 is hereinafter referred to as an ECC block.

In the construction shown in FIG. 5, the data to which errors were mixed by the transmission path is first written into a data memory 301 comprising a RAM on a unit basis of n.sub.1 .times.n.sub.2 blocks. After that, the data is sequentially read out in the C.sub.1 direction by the address operation of the memory 301 and syndromes are calculated by a syndrome calculating unit 302. The error position and size are calculated by an error detection correction processing unit 303 on the basis of the syndromes calculated, thereby correcting the erroneous data in the data memory 301. After such processes were executed with respect to n.sub.1 lines, the similar processes are also executed in the C.sub.2 direction with regard to n.sub.2 lines. After completion of a series of processes, the error corrected data is read out of the data memory 301 on an ECC block unit basis.

In the conventional example, in order to detect and correct the errors in the ECC block shown in FIG. 4, the following number of access times of the data memory 301 is needed. That is, (n.sub.1 .times.n.sub.2) times are necessary to write the input data. (n.sub.1 .times.n.sub.2) times are necessary to read out to calculate the syndromes in the C.sub.1 direction and (n.sub.1 -k.sub.1 -2) times are necessary per maximum lines to write to correct the data. In the C.sub.2 direction as well, (n.sub.1 .times.n.sub.2) times are similarly necessary to calculate the syndromes, (n.sub.2 -k.sub.2 -2) times per maximum lines are necessary to correct the data, and (n.sub.1 .times.n.sub.2) times are necessary to read out the output data. Thus, it is necessary to execute the accessing operations of total 4(n.sub.1 .times.n.sub.2)+.alpha. (.alpha. denotes the number of data to be corrected) times.

As mentioned above, according to the conventional construction, the number of access times of the memory is large. On the other hand, in realization of the high operating speed (cycle time) of the memory, since it is limited, the processing time of the error detection correction unit is long. In the case where the high operating speed is required, it is necessary to construct the circuits in parallel and to suppress the correcting capability or the like. Such a method causes the correcting capability to deteriorate and increases the circuit scale and a large problem occurs. Further, although the above conventional example has been described with respect to the case of processing the data to which the double encoded error correction code was added, in the case of executing the processes corresponding to the double encoding or more, the number of access times of the memory increases in proportion to them. Therefore, it is necessary to further increase the hardware scale of the processing unit.

SUMMARY OF THE INVENTION

In consideration of the above points, it is an object of the invention to provide a method of executing the error correction in a short time.

Another object of the invention is to provide an error correcting method or apparatus which can execute the error correction in a short time and at the high accuracy.

Under such objects, according to a preferred embodiment of the present invention, there is disclosed an error correcting method in a decoding method of digital data having a two-dimensional arrangement including error correction check codes in both of the first and second directions, wherein in at least one of the first and second directions, values in which the result of the arithmetic operation for error correction was added to syndromes are obtained, and the error correction is again executed on the basis of those value.

Still another object of the invention is to provide an error correcting apparatus having a construction suitable to execute the error correction at a high speed.

Still another object of the invention is to provide an error correcting apparatus which can realize the high processing speed of the error correction for the data to which the multiple encoded error correction code was added.

Under the above objects, according to a preferred embodiment of the invention, there is disclosed an error correcting apparatus comprising: a plurality of syndrome calculating units to which input data added with a multiple encoded error correction code is supplied; a syndrome memory to store syndromes calculated by the plurality of syndrome calculating units; a data memory which is connected in parallel with the plurality of syndrome calculating units for the input data; and a correction processing unit to execute the error correction for data in the data memory by using the syndromes stored in the syndrome memory.

The above and other objects and features of the present invention will beccome apparent from the following detailed description and appended claims with reference to the accompanying drawings.

BRIEF DESCRIPTION OF THE DRAWINGS

FIGS. 1(a)-1(c) are diagrams showing a memory construction in an embodiment of the present invention;

FIG. 2 is a block diagram showing a construction of an error correcting apparatus as an embodiment of the invention;

FIG. 3 is a diagram showing a construction of a product code;

FIG. 4 is a diagram showing a data block to which an error correction code was added; and

FIG. 5 is a block diagram showing an example of a construction of a conventional error correcting apparatus.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

In the first embodiment of the present invention which will be described hereinbelow, in the decoding of digital data having a two-dimensional arrangement having a digital data train of n.sub.1 words including error correction detection check codes of k.sub.1 words in the first direction and a digital data train of n.sub.2 words including error correction detection check codes of k.sub.2 words in the second direction, there is disclosed a data decoding method whereby in the case of repetitively executing the error correction in the first direction and the error correction in the second direction, there is provided memory means for storing the values in which the arithmetic operation corresponding to the error correcting operation which was executed in the first direction was applied to syndromes as a result of the error correction in the first direction and the values in which the arithmetic operation corresponding to the error correcting operation which was executed in the second direction was applied to the syndromes as a result of the error correction in the second direction. However, the invention is not obviously limited to such a construction.

An embodiment of the invention will be described hereinbelow.

It is assumed that all of the following arithmetic operations in the embodiment relate to the arithmetic operations on the Galois field. In the embodiment, the code words of c.sub.1 or c.sub.2 are based on the Reed Solomon code and a check matrix H is as follows. ##EQU1## where d is a hamming distance.

Assuming that the code words C are as follows

C=(c.sub.n-1, c.sub.n-2, . . . c.sub.1, c.sub.0) it is the reception code words such that c=(c.sub.n-1 e.sub.n-1, . . . , c.sub.1 +e.sub.1, c.sub.0 +e.sub.0) to which error patterns E E=(e.sub.n-1, e.sub.n-2, e.sub.1, e.sub.0) were added on the transmission path are decoded. When all of the error patterns are 0, C=C and all of the syndromes are set to 0. Such syndromes S.sub.j (j=0 to d-1) are expressed by ##EQU2##

Since there are n.sub.2 code words in the A direction, assuming that a hamming distance is set to d.sub.1, the number of syndromes corresponding to those n.sub.2 code words is

n.sub.2 .times.(d.sub.1 -1) On the other hand, assuming that a hamming distance is set to d.sub.2, k.sub.1 .times.(d.sub.2 -1) syndromes exist in the B direction.

All of the syndromes are stored into memories A and B shown by S(c.sub.1) and S(c.sub.2) in FIG. 1. In the embodiment, in each error correcting process, after the error correction in step 4) was executed,

e.sub.x .multidot..alpha..sup.xj in consideration of the syndrome number j is added to the error pattern e.sub.x and error position .alpha..sup.x which were estimated for the relevant syndrome.

That is, assuming that the new syndrome is set to S.sub.j ',

S.sub.j '=S.sub.j +e.sub.x .alpha..sup.xj

The new syndrome S.sub.j ' is written into the memory A or B and is used as a basic syndrome in the next, that is, the second and subsequent repetitive error correcting operations.

Due to this, in the second (t=2) and subsequent repetitive error correcting operations, there is no need to execute the arithmetic operation of each syndrome. The content of the memory S(c.sub.1) or S(c.sub.2) coincides with the value of the syndrome at each time point.

Therefore, the time which is necessary for the processes of repetitive t times is calculated in a manner similar to the above and becomes as follows.

{(n.sub.1 +n.sub.1)n.sub.2 +(n.sub.2 +n.sub.2)k.sub.1 }+{n.sub.1 n.sub.2 +n.sub.2 k.sub.1 }(t-1)=(t+1) (n.sub.1 n.sub.2 +n.sub.2 k.sub.1)(2) Thus, it will be understood that the processing time is reduced by only the time of (t-1)(n.sub.1 n.sub.2 +n.sub.2 k.sub.1) as compared with the processing time shown in the equation (1) in the conventional example.

Although the above embodiment has been described with respect to the calculation of the processing time, for instance, in the case of executing the processes in steps 1 and steps 2 to 4 by the pipeline processes or the like, the processing time is merely slightly reduced; however, the circuits for such pipeline processes can be remarkably decreased. On the other hand, in the case of executing only steps 2 to 4 by the pipeline processes, assuming that the circuit scale is the same, the processing time can be obviously reduced into about half the time.

As described above, according to the method of the embodiment, a data decoding method which can extremely reduce the processing time can be provided.

A construction of an error correcting apparatus in another embodiment of the invention will now be described.

FIG. 2 is a diagram showing a construction of an error correcting apparatus for executing one correction decoding of the double encoded Reed Solomon code as shown in FIG. 4. Data to which the double encoded Reed Solomon code was added is supplied to an input terminal 114 through a transmission path. The input data is written into a data memory 112 on a data block unit basis of eight bits. The data memory 112 has a capacity which can store the data in which the error correction codes were eliminated from data to be arithmetically operated in a syndrome calculating unit, which will be explained hereinlater. Simultaneously with the writing of the input data into the data memory 112, syndromes of C.sub.1 and C.sub.2 are simultaneously calculated in syndrome calculating units 101 and 105. The syndrome calculating unit 101 on the C.sub.1 side comprises: adders 103, 103', and 103"; one-data delays 102, 102', and 102"; and multipliers 104 and 104'. Assuming that reception data is set to W.sub.i and the root of the generating polynomial is set to .alpha..sup.m, syndromes S.sub.0, S.sub.1, and S.sub.2 can be calculated by the following equations (1) to (3). ##EQU3##

On the other hand, in the syndrome calculating unit 105 on the C.sub.2 side, one-line delays 108, 108', and 108" to execute the arithmetic operations in the vertical direction are added to one-data delays 106, 106', and 106". The calculating unit 105 calculates syndromes by the arithmetic operations of the above equations (1) to (3) in a manner similar to the case of C.sub.1. Reference numerals 107, 107', and 107" denote adders and 109 and 109' indicate multipliers.

The calculated syndromes are sequentially written into a syndrome memory 110. The syndromes are rewritten by the syndrome memory 110 as will be explained hereinlater, so that the parallel processes can be accomplished. The error correcting operation will now be described hereinbelow.

At a time point when the error correcting operation is executed, the writing of the input data into the data memory 112 and the writing of the syndromes of C.sub.1 and C.sub.2 into the syndrome memory 110 have already been completed. The correcting operation is executed by a series of processes such that an error detection correction processing unit 111 reads out the syndromes from the syndrome memory 110 and calculates the error position and size and rewrites the erroneous data in the input data in the data memory 112 on the basis of the result of the error position and size calculated. Such processes are sequentially executed with respect to C.sub.1 and C.sub.2. However, since the data in the data memory 112 is rewritten by the correcting operation on the C.sub.1 side which is first executed, the values of the syndromes on the C.sub.2 side of the rewritten data differ. Therefore, in a manner similar to FIG. 1 described, upon completion of the correcting operation on the C side, the product of the size of error of the data corrected by C.sub.1 and the error position on the C.sub.2 side of the data is added to the original syndromes on the C.sub.2 side of the data, thereby rewriting the syndromes, thus corrected, in the syndrome memory 110. Due to this, even if the calculations of the syndromes of C.sub.1 and C.sub.2 are simultaneously executed, the accurate error correcting operation can be executed. At the time point when the correcting processes of C.sub.1 and C.sub.2 was finished, data is output from the data memory 112 through an output terminal 115. All of timing signals which are used in the apparatus in the embodiment are generated by a timing controller 113.

According to the above embodiment, since the hardware is constructed so as to calculate the (n-k-2) syndromes simultaneously with the writing of the input data, the similar processes can be executed for the time which is about 1/(n-k-2) shorter than that in the conventional apparatus.

According to the construction of the above embodiment, syndromes of a plurality of codes can be simultaneously calculated by a plurality of syndrome calculating units and the data accessing operations to the data memory can also be simultaneously executed. This is because the syndromes are rewritten syndrome memory. That is, even if the calculations of the syndromes of the multiple codes are simultaneously executed,the error correcting processes can be correctly executed in the syndrome memory on the basis of the rewritten syndromes. Thus, the number of access times to the data memory remarkably decreases and the high error correction processing speed can be realized.

The above embodiment has been described with respect to the example of the double encoding. However, even in the case of handling the data which was subjected to the triple or more encoding, by similarly arranging a plurality of syndrome calculating units in parallel, the error correcting processes can be executed without increasing the number of access times of the memory.

As described above, according to the error correcting apparatus of the embodiment, it is possible to realize the high error correction processing speed for the data to which the multiple encoded error correction code was added without increasing the hardware scale.

Claims

1. An error correcting apparatus comprising:

syndrome calculating means for calculating first and second syndromes from input data including information codes and error correction check codes;

a data memory which is connected in parallel with said syndrome calculating means for storing the information codes;

a syndrome memory to store syndromes calculated by said syndrome calculating means; and

error correction means for correcting errors of the information codes stored in said data memory by using the first and second syndromes stored in said syndrome memory and for correcting the second syndrome stored in said syndrome memory by using a result of an error correction executed on information codes stored in said data memory by using the first syndrome.

2. An apparatus according to claim 1, wherein said syndrome calculating means comprises:

a delay to delay the input data; and

a multiplier to multiply the data delayed by said delay and a predetermined coefficient.

3. An apparatus according to claim 1, wherein said data memory has a capacity large enough to store the data in which error correction check codes are eliminated from the input data.

4. An apparatus according to claim 1, wherein said syndrome memory is constructed as a memory for exclusive use.

5. An apparatus according to claim 1, wherein said error correction means calculates error position and size by using the syndromes.

6. An apparatus according to claim 1, wherein said syndrome calculating means calculates the first syndrome from an error correction check code (C.sub.1) provided in a first direction and the second syndrome from an error correction check code (C.sub.2) provided in a second direction when the information codes are two-dimensionally arranged.

7. An apparatus according to claim 1, wherein said data memory stores the information codes concurrently with the calculating by said syndrome calculating means.

8. An apparatus according to claim 1, wherein said syndrome calculating means calculates the first and second syndromes concurrently.

9. An error correcting method comprising:

a calculating step for calculating first and second syndromes from input data including information codes and error correction check codes;

a first storing step for storing the information codes into a data memory;

a second storing step for storing the first and second syndromes calculated in said calculating step into a syndrome memory;

a first correcting step for correcting errors of the information codes stored in the data memory by using the first syndrome stored in the syndrome memory;

a second correcting step for correcting the second syndrome stored in the syndrome memory by using a result of an error correction performed in said first correcting step using the first syndrome; and

a third correcting step for correcting errors of the information codes stored in the data memory by using the second syndrome corrected in said second correcting step.

10. A method according to claim 9, wherein the first syndrome is calculated from an error correction check code (C.sub.1) provided in a first direction and the second syndrome is calculated from an error correction check code (C.sub.2) provided in a second direction when the information codes are two-dimensionally arranged.

11. A method according to claim 9, wherein storage of the information codes is performed concurrently with performance of said calculating step.

12. A method according to claim 9, wherein the first and second syndromes are calculated concurrently.