JP2692085B2 - Error correction method - Google Patents

Description

Description: TECHNICAL FIELD The present invention relates to an error correction method using a code such as a product code, a concatenated code, or a chain code that is double-encoded with two sequences. SUMMARY OF THE INVENTION The present invention uses a code in which a first code is encoded in a first sequence and a second code is encoded in a second sequence for a predetermined number of symbols. In the error correction method, an error correction method in which an error correction process in a first series and a second error correction process are repeatedly performed,
When the error correction processing in the series of 1 is performed again, the line that could not be error-corrected when the error correction processing in the first series of the previous time was performed is targeted, or the error is corrected in the second error correction processing. By targeting the line including the symbol, unnecessary correction processing time is eliminated and the correction processing time is shortened. [Prior Art] In recent years, a data recording medium having an extremely large recording capacity such as a magneto-optical disk has been developed. In a data recording medium having a very large recording capacity such as a magneto-optical disk, it is important to take measures against burst errors. Therefore, as shown in Japanese Patent Application No. 61-24367, for example, error detection or error correction coding is performed on the data to be recorded on the magneto-optical disk. In a magneto-optical disk, for example, a product code for coding each row and each column is used. That is, the data are arranged in a two-dimensional matrix, the first error correction code C1 is encoded for the data of each row of the two-dimensional array data, and the second error correction code for the data of each column. C2 encoding is performed. In addition to the above, in the magneto-optical disk, the LDC (Long) that performs coding in each data series in the row direction of two-dimensional array data and reads / writes data in the column direction
Distance Code) is also used. Conventionally, decoding of such product codes is performed by decoding C1 sequence and C2 sequence.
This is done by alternately repeating the sequence decoding. That is, first, the data is read in the row direction, and C1 is read for each row.
Series syndrome is required. If an error symbol is generated from this syndrome, error correction is performed for a correctable one. When error correction is impossible, an error flag indicating that an error is included in the symbol of the line is provided for each line or for each symbol of each line. Next, the data is read in the column direction, and the syndrome of the C2 series is obtained for each column. From this syndrome,
If an error symbol has occurred, error correction is performed for correctable errors. Further, erasure correction is performed using the error flag obtained at the time of decoding the C1 sequence. When error correction is impossible, an error flag indicating that an error is included in the symbol of the example is provided for each line or for each symbol of each line. Further, the C1 sequence is decoded again. In this C1 sequence decoding, error correction is performed from the syndrome, and erasure correction is performed using the error flag obtained during C2 sequence decoding. Hereinafter, C2 sequence decoding and C1 sequence decoding are repeatedly performed. [Problems to be Solved by the Invention] As described above, a code encoded by two sequences such as a product code is decoded by alternately repeating C1 sequence decoding and C2 sequence decoding. By the way, in the past, when C1 sequence decoding is performed, C2 sequence decoding is performed, and then C1 sequence decoding is repeated,
The data is read out for all the rows of the series, all the syndromes of each row of the C1 series are obtained, and the decoding is performed for all the rows of the C1 series. Further, even when the decoding of the C2 series is repeated again, the data for all the columns of the C2 series are read, and the syndromes of each column of the C2 series are obtained,
Decoding is performed for all columns of the C2 sequence.
Therefore, when repeating the decoding of the C1 series, it is necessary to read the data of all the rows of the C1 series and obtain the syndrome of all the rows of the C1 series, and when repeating the decoding of the C2 series, Read the data from all columns of the C2 series, and
It takes time to find the syndromes for all columns in the sequence. Therefore, the conventional error correction method has a problem that the processing time is long. Therefore, an object of the present invention is to provide an error correction method which eliminates unnecessary data processing time and shortens the processing time. [Means for Solving the Problems] According to the present invention, the first code is encoded in the first sequence and the second code is encoded in the second sequence for a predetermined number of symbols. An error correction method using a code, which performs error correction processing in the first sequence, and when error correction is impossible, an error flag in the first sequence is applied to an uncorrectable row or each symbol in that row. And the erasure correction processing using the error flag in the first series, and the error flag in the second series cannot be corrected. Steps provided for columns or respective symbols of the columns, and error correction processing in the first series and errors in the second series for the rows in which the error flags in the first series are set And a step of performing erasure correction processing using a flag. The present invention is an error correction method using a code in which a first code is encoded in a first sequence and a second code is encoded in a second sequence for a predetermined number of symbols. And performing error correction processing in the first series and providing an error flag in the first series for the uncorrectable row or each symbol of that row when the error correction is not possible, and for the second series Performing an error correction process and an erasure correction process using the error flag in the first series, and at the same time, providing a collection flag indicating a row direction position including a symbol corrected in the second series; The step of performing again the error correction processing in the first series and the erasure correction processing using the error flag in the second series for the row provided with the flag. It is an error correction method according to claim. [Operation] In the present invention, when the error correction process of the C1 series is repeated, the correction process is performed on the line in which the C1 error flag is set, and when the error correction process of the C2 series is repeated, the C2 error is generated. The correction process is performed on the flagged columns. Further, in the present invention, when the error symbol is corrected in the C2 series, the C1 correction flag is set in the row including the corrected symbol, and when the C1 series error correction process is repeated, the C1 correction flag is set. The correction process is performed for the existing line. When the error symbol is corrected in the C1 sequence, C2 is added to the column containing the corrected symbol.
When the collection flag is set and the C2 series error correction process is repeated, the correction process is performed for the column for which the C2 collection flag is set. Embodiments Embodiments of the present invention will be described in the following order. Recording form of magneto-optical disc b. Decoding of product code c. Description of first embodiment d. Description of second embodiment a. Recording form of magneto-optical disc Extremely recording capacity like magneto-optical disc In a large-sized data recording medium, error correction coding is performed in order to cope with burst errors and the like. FIG. 1 shows an example in which error correction coding is performed using a product code. In the magneto-optical disk, as shown in FIG.
Data is recorded in units of. The error correction coding is performed for each sector S. User data that can be recorded per sector is 512 bytes, for example. As shown in FIG. 1, additional information of, for example, 12 bytes is added to the 512-byte user data as necessary, and further, for example, 4 bytes for 524-byte data of one sector including the additional information. The error detection code EDC of is generated and added. For example, link information to the next sector, data identification information, and the like are inserted in the additional information together with the reserve information. A CRC code is used as the error detection code EDC. In this way, the C1 series error correction coding and the C2 series error correction coding are performed on the 528-byte data to which the additional information and the error detection code EDC are added. That is, FIG. 1 is a matrix block forming a coding unit. The matrix block of 14 rows and 48 columns corresponds to the data size recorded in one sector.
As described above, the 528-byte data to which the additional information and the error detection code EDC are added is arranged in a matrix of 12 rows and 44 columns. Then, the C1 sequence is encoded for each row. The C1 sequence error correction code is, for example, (48,4
4) Reed-Solomon code is used. By encoding the C1 sequence, 4-byte check symbols are generated and added for each row. Furthermore, C2 sequence coding is performed for each column. An example of the C2 sequence error correction code is (14,1
2) Reed-Solomon code is used. By the C2 sequence encoding, 2-byte check symbols are generated and added for each column. In this way, the error-correction-encoded data is the second data.
Recording is performed on the magneto-optical disk as shown in FIGS. In FIG. 2, the magneto-optical disk 1 has, for example, a diameter of
It is 13 cm (5 inches) and has a storage capacity of 300 Mbytes or more on one side. The disk 1 is rotated at a constant angular velocity,
Data is recorded by forming a concentric or spiral track 2 with one track per rotation.
The number of tracks on one side is, for example, 18000 to 20000. As shown in FIG. 3A, each track is divided into n sectors from sector 0 to sector (n-1), for example, 16 or 32 sectors. Each sector has a servo signal area 3 and a data / address signal area 4 as shown in FIG. 3B.
It is divided into and. The servo signal area 3 has a length corresponding to, for example, 2 bytes, and servo signals are recorded in the servo signal area 3 in the form of pits. Data address signal area 4
Has a length of 16 bytes, for example, and data signals and address signals are recorded in the data / address signal area 4. 1
In the data address signal area 4 at the beginning of one sector,
At least the address signal is recorded. b. Decoding of Product Code As described above, the present invention can be used when error correction is performed using a product code that is doubly encoded by two sequences of C1 sequence and C2 sequence. Decoding of such a product code is performed by alternately repeating C1 series decoding and C2 series decoding. That is,
First, the data is read out in the row direction, and the syndrome of the C1 series is obtained for each row. If an error symbol is generated from this syndrome, error correction is performed for a correctable one. In the C1 series, 4-symbol erasure correction is possible and 2-symbol error correction is possible. Therefore, in the first decoding of the C1 sequence, an error within 2 symbols can be corrected. When error correction is impossible, a C1 error flag indicating that an error is included in the symbol of the line is provided for each line or for each symbol of each line. Next, the data is read in the column direction, and the syndrome of the C2 series is obtained for each column. From this syndrome,
If an error symbol has occurred, error correction is performed for correctable errors. In the C2 series, 2-symbol erasure correction is possible and 1-symbol error correction is possible. Therefore, this C2 sequence decoding can be corrected if the error is within 1 symbol. If the error is within 2 symbols, erasure correction can be performed using the C1 error flag obtained when the first C1 sequence is decoded. When error correction is impossible, a C2 error flag indicating that an error is included in the symbol of the column is provided for each row or for each symbol of each row. Further, the C1 sequence is decoded again. In this C1 decoding, if the error is within 2 symbols, it can be corrected, and if the error is within 4 symbols, erasure correction can be performed using the C2 error flag obtained at the time of decoding the C2 sequence. Hereinafter, C2 sequence decoding and C1 sequence decoding are repeatedly performed. c. Description of the first embodiment By the way, when the C1 sequence is decoded, the C2 sequence is decoded, and then the C1 sequence is further decoded, it is not necessary to read the data of all the rows of the C1 sequence. It is not necessary to perform the process again for a row that does not include an error symbol at the time of decoding the first C1 sequence or a row in which all error symbols have been corrected. In the first embodiment of the present invention, when the decoding of the C1 sequence is repeated, the row including the symbol that cannot be corrected when the decoding of the previous C1 sequence is completed is targeted. Further, when the decoding of the C2 sequence is repeated, the column including the symbol that cannot be corrected when the decoding of the previous C2 sequence is completed is targeted. By performing such processing, unnecessary data processing time is eliminated and processing time is shortened. A C1 error flag is set on a line containing a symbol that could not be corrected when the previous decoding of the C1 sequence was completed. Further, a C2 error flag is set in a column including a symbol that could not be corrected when the previous decoding of the C2 sequence was completed. Therefore, in one embodiment of the present invention, when the decoding of the C1 sequence is repeated, the row in which the C1 error flag is set at the previous decoding of the C1 sequence is targeted.
Further, when repeating the combination of the C2 series, the column for which the C2 error flag was set at the time of the previous decoding of the C2 series is targeted. The error correction processing procedure in the embodiment of the present invention will be described with reference to the flowchart shown in FIG. First, the data is read in the row direction, and the decoding process is performed on all the rows in the C1 sequence. This decoding process corrects errors of 2 symbols or less. Then, a C1 error flag is set on a line containing an error of 3 symbols or more (step). The data is read in the column direction, and the decoding process is performed on all columns in the C2 series. This decoding process corrects one symbol error. In addition, erasure correction of 2-symbol error is performed using the C1 error flag. The column that contains more errors is flagged as C2 error (step). It is judged whether all the syndromes of the C1 series and the C2 series have become “0” (step). If all the syndromes are “0”, the decoding process ends. When all the syndromes are not "0", the data of the row in which the C1 error flag is set is read, and the decoding processing is performed on the row in which the C1 error flag is set. This decoding process corrects errors of 2 symbols or less. Then, erasure correction of a 4-symbol error is performed using the C2 error flag. When the correction is made, the C1 error flag of the corrected line is cleared (step). It is judged whether all the syndromes of the C1 series and the C2 series have become “0” (step). If all the syndromes are “0”, the decoding process ends. When all the syndromes are not "0", the data in the column in which the C2 error flag is set is read, and the decoding process is performed on the column in which the C2 error flag is set. This decoding process corrects an error of 1 symbol.
Then, erasure correction of a 2-symbol error is performed using the C1 error flag. When the correction is made, the C2 error flag of the corrected column is cleared (step). Returning to the step, it is judged whether the syndromes are all “0”. If the syndromes are not all "0", the decoding process of the C1 sequence for the row with the C1 error flag and the C2 for the column with the C2 error flag are performed.
The sequence decoding process continues in an alternating manner. d. Description of Second Embodiment In the above-described one embodiment, when the decoding of the C1 sequence is repeated, the row including the symbol that cannot be corrected when the decoding of the previous C1 sequence is completed is targeted. When the decoding of the C2 sequence is repeated, the column including the symbol that could not be corrected when the decoding of the previous C2 sequence was completed is targeted. However, this makes it impossible to deal with an error that was missed during the previous decoding of the C1 series or the decoding of the C2 series. In the second embodiment of the present invention, when the decoding of the C1 sequence is repeated, the row including the symbol corrected in the decoding of the previous C2 sequence is targeted, and when the decoding of the C2 sequence is repeated, The column containing the symbols corrected during the decoding of the previous C1 sequence is targeted. In the second embodiment of the present invention, when the error symbol is corrected during the decoding of the C2 sequence, the C1 correction flag is set in the row containing the corrected error symbol. Also,
When the error symbol is corrected during the decoding of the C1 sequence, the column containing the corrected error symbol is set with the C2 correction flag. When the decoding of the C1 series is repeated, the row for which the C1 collection flag is set at the time of the previous decoding of the C2 series is targeted. When the decoding of the C2 series is repeated, the column for which the C2 collection flag is set at the time of the previous decoding of the C1 series is targeted. For example, as shown in FIG. 5A, three error symbols e ₁ , e ₂ and e ₃ are arranged in a row l 1 and three error symbols are arranged in a row _{l 2} .
It is assumed that e ₄ , e ₅ and e ₆ are lined up. First, the decoding process is performed on all the rows of the C1 sequence. At the time of decoding the first C1 sequence, the error symbols e ₁ , e ₂ , and e ₃ occurring in the row l1 are 3 symbols or more, and the error symbols e ₄ , e ₅ , and e ₆ occurring in the row l2 are 3 symbols. As above, the error symbols e ₁ , e ₂ , e ₃ and the error symbols
e _4, e _5, e ₆ can not be corrected. Next, the decoding process is performed on all the columns of the C2 sequence. When decoding the C2 sequence, the error symbol e ₁ of the column m1 is corrected, the error symbols e ₂ and e _{4 of} the column m2 are corrected using the C1 error flag at the time of C1 decoding, and the error symbols e ₃ and e _{5 of the} column m3 are corrected.
Is corrected using the C1 error flag at the time of C1 decoding, and the error symbol e ₆ of the column m4 is corrected. Then, as shown in FIG. 5B, the C1 correction flag is set in the row l1 including the corrected symbols e ₁ , e ₂ and e ₃ , and the corrected symbols e ₄ and e
_The C1 collection flag is set on line l2 containing ₅ , e ₆ . Then, the decoding of the C1 sequence is repeated again. At this time, the data of the rows l ₁ and l ₂ for which the C1 collection flag is set are read, and the C1 sequence is decoded only for the rows l1 and l2. That is, when C1 sequence decoding is repeated again,
Only the row containing the error-corrected symbol at the time of decoding the C2 sequence has a different result from the process at the time of decoding the sequence.
For a row that does not include an error-corrected symbol at the time of decoding the C2 series, even if the decoding of the C1 series is repeated, the same result as the processing at the time of the previous decoding of the C1 series is obtained. An error correction processing procedure according to the second embodiment of the present invention will be described with reference to the flowchart shown in FIG. First, the data is read in the row direction, and the decoding process is performed on all the rows in the C1 sequence. The C1 series can correct errors of 2 symbols or less. Then, a C1 error flag is set on a line containing an error of 3 symbols or more (step). The data is read in the column direction, and the decoding process is performed on all columns in the C2 series. This decoding process corrects one symbol error. In addition, erasure correction of 2-symbol error is performed using the C1 error flag. A column that contains more errors will be flagged as a C2 error. Along with this, the C1 collection flag is set on the line containing the corrected symbol in the C2 sequence (step). It is judged whether all the syndromes of the C1 series and the C2 series have become “0” (step). If all the syndromes are “0”, the decoding process ends. When all the syndromes are not "0", the data of the line where the C1 collection flag is set is read and
Decoding processing is performed on the rows for which the collection flag is set. This decoding process corrects errors of 2 symbols or less. Then, erasure correction of a 4-symbol error is performed using the C2 error flag. When the correction is made, the C1 error flag of the corrected row is cleared.
Along with this, C2 is added to the column containing the corrected symbols in the C1 sequence.
The collection flag is set (step). It is judged whether all the syndromes of the C1 series and the C2 series have become “0” (step). If all the syndromes are “0”, the decoding process ends. When the syndromes are not all "0", the data in the column for which the C2 collection flag is set is read and C2
The decoding process is performed on the column for which the collection flag is set. This decoding process corrects an error of 1 symbol. Then, erasure correction of a 2-symbol error is performed using the C1 error flag. When the correction is made, the C2 error flag of the corrected column is cleared. Along with this, the C1 collection flag is set on the line containing the corrected symbol in the C2 sequence (step). Returning to the step, it is judged whether the syndrome is all "0". If the syndromes are not all "0", C1 for the line where the C1 collection flag is set
The sequence decoding process and the C2 sequence decoding process for the columns for which the C2 collection flag is set are alternately repeated. The present invention can be applied not only to the product code but also to the case of cross interleaving or the like. (Effect of the invention) According to the present invention, when decoding the C1 sequence, performing the decoding of the C2 sequence, and repeating the decoding of the C1 sequence again, the symbols that could not be error-corrected during the decoding of the previous C1 sequence The row that includes the row or the row that includes the error-corrected symbol when decoding the C2 series is the target of the decoding processing of the C1 series, and the decoding processing is not performed on the other rows. When repeating C2 sequence decoding again, the column containing the symbols that could not be error corrected during the previous C2 sequence decoding, or the column containing the error-corrected symbols during the C1 sequence decoding, was the C2 sequence decoding process. , And the decoding process is not performed for other columns. Therefore, it is possible to shorten the processing time such as the data reading time and the time for obtaining the syndrome when the C1 series decoding is repeated or the C2 series decoding is repeated.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a schematic diagram of an example of a code structure to which the present invention is applicable, FIGS. 2 and 3 are schematic diagrams of an example of a recording format of a magneto-optical disk, and FIG. FIG. 5 is a flow chart used for explaining the first embodiment of the present invention, FIG. 5 is a schematic diagram used for explaining the second embodiment of the present invention, and FIG. 6 is used for explaining the second embodiment of the present invention. It is a flowchart. Description of main symbols in the drawings 1: Magneto-optical disk, 2: Track, 3: Servo signal area, 4:
Data address signal area.

Claims (1)

(57) [Claims] The first code is decoded in the first sequence, the second code is decoded in the second sequence for a predetermined number of symbols, and the decoding process in the first and second sequences is performed. In the error correction method, the error correction process is performed in the first series, and when the error correction is impossible, an error flag in the first series is provided for the uncorrectable row or each symbol of the row. Error correction processing is performed in the second sequence, and when the error correction is impossible, an error flag in the second sequence is provided for the uncorrectable sequence or each symbol of the sequence, and And a step of setting a collection flag in a row including a symbol corrected in the second series, and when performing the error correction processing in the first series again, during the error correction processing in the previous second series. Error correction method comprising the step of correcting error in that row standing collection flag temple target. 2. When performing the error correction processing in the first series again, the erasure correction is performed by using the error flag set at the time of the error correction processing in the previous second series. The error correction method according to the first item of the above. 3. When performing the error correction processing in the first series again, erasure correction is performed using the error flag set at the time of the previous error correction processing in the second series, and the error flag of the corrected row Along with resetting
The error correction method according to claim 2, wherein a collection flag is set for a column including the symbols corrected in the first sequence. 4. Further, when performing error correction processing in the second stream, decoding correction is performed for the column in which the collection flag set in the first stream is set. The error correction method described in the third item of the above.