JPS62159519A - Method and apparatus for error correction - Google Patents

Abstract

PURPOSE:To prevent an error from being overlooked by controlling an area to be corrected not by correction but by replacement depending on the occurrence state of the error. CONSTITUTION:A reproduced digital signal inputted from an input terminal 1 is subjected to 1st stage decoding by a decoder 2 in the row direction at first. Then the 2nd stage of decoding is applied by using decoders 3-5 taking error correction word number in the column direction as 4, 3, 2 based on the result of decoding and the decoding data is stored respectively in memories 6-8. A counter 9 calculates number of columns not corrected by the decoder 4 in the column direction at each, e.g., unit of product sign. Then a selector 10 selects any of the memories 6-8 by using a count C and a decoded data is outputted from an output terminal 11.

Description

[Detailed Description of the Invention] [Field of Application of the Invention] The present invention relates to an error correction method and device thereof, which can increase the correction capability while reducing the probability of error correction, and is suitable for compact discs, digital VTRs, etc. The present invention relates to a method and an apparatus thereof.

[Background of the invention]

In recent years, error correction codes have become widely used in everything from computer storage devices to home audio equipment. FIG. 3 shows the structure of an error correction code called a product code in which error correction parities 31 and 32 are two-dimensionally added to digital data 30 to be recorded or transmitted.

Product codes are explained in detail in, for example, the 73rd page of "Coding Theory" published by Shokodo.

The square in FIG. 3 consists of multiple bits of data, and is hereinafter referred to as a word. The product codes encoded in this way are recorded or transmitted in the order shown by the arrows.

The decoding procedure of the product code will be described below. First, errors are corrected using the error correction parity 31 for each row. If an error that exceeds the correction ability occurs, all words included in that line are made uncertain, and a flag indicating uncertainty is added to each word. Next, errors are corrected for each column using a flag and error correction parity 32 (referred to as erasure correction).

If correctable, the error is corrected and the flag for the word in that column is removed. If the correction capacity is exceeded, a flag is added to all words in that column.

Finally, words without flags are output as they are, and words with flags are output after correcting errors by replacing them with, for example, a highly correlated word from one field before.

Examples of the above error correction procedure described in detail include, for example, JP-A-57-10557, JP-A-57-10558, JP-A-57-10559, and JP-A-57-1.
0560, JP-A-57-10561, etc.

In general, deterioration in image (sound) quality occurs when incorrect correction is made or when a highly correlated signal is substituted without correction. Furthermore, these are in an antinomic relationship, and simply reducing the deterioration of one will simultaneously increase the deterioration of the other, so it is not easy to reduce both of the deteriorations. Therefore, in the above error correction method, both the deterioration due to incorrect correction and the deterioration due to replacement are not reduced to below the allowable level, and it is difficult to reduce the deterioration due to incorrect correction and the deterioration due to replacement to below the allowable level. There still remained problems.

[Purpose of the invention]

An object of the present invention is to provide an error correction method and apparatus that controls the number of errors to be corrected depending on the error occurrence state, and eliminates missed errors and corrections due to unnecessary replacement.

[Summary of the invention]

In the product code shown in Figure 3, missed errors are caused by decoding performed on each row, which causes a slight error of one word or less (hereinafter referred to as error This happens when you miss something (called an oversight).

Such errors can only be overlooked when long, continuous burst errors of several hundred to several thousand bits occur. In addition, if an error is missed by decoding in the row direction, there is a higher probability that an error exceeding the correction ability will occur or an error of multiple words or more will occur in column direction decoding than when no error is missed. . Therefore, by counting the number of columns that cannot be corrected by decoding the product code in the column direction or in which an error of multiple words or more has occurred, it is possible to determine whether errors were missed in the decoding in the row direction. I understand.

Therefore, by changing the number of error words to be corrected in the column direction based on the count value, it is possible to reduce corrections caused by overlooking errors or unnecessary replacement, and to reduce the accompanying deterioration in image (sound) quality. can. Specifically, when the count value is small, the number of error words that can be corrected is increased, and as the count value increases, the number of correctable words is sequentially decreased.

[Embodiments of the invention]

The product code shown in Figure 3 is a Reed-Solomon code on an extended field 0F(2δ) where one word is 8 bits, and a 4-word parity part is added to 32 words of information in the row direction, and a 4-word parity part is added in the column direction. Let us now consider a case in which a 4-word parity section is added to a 28-word information word. In this case, decoding in the row direction can correct errors of 2 words or less.

Further, in column direction decoding, errors of 4 words or less can be corrected by using the row direction decoding results. However, in practice, in order to prevent errors from being overlooked due to incorrect corrections, the number of corrections is kept to one word or less in row direction decoding, and three words or less in column direction decoding, for example.

In the present invention, the number of error words to be corrected by decoding in the row direction is the same as or larger than the conventional method. Decoding in the column direction is performed in two stages. In the first stage, for example, 3
Errors smaller than a word are corrected, and the number of columns that cannot be corrected is counted for each of, for example, 28 columns constituting one product code. In the second stage, the number of error words to be corrected is changed based on this count value and errors are corrected anew. In other words, if the count value is large, there is a high probability that an error will be missed during decoding in the row direction, so by reducing the number of correction words, the error detection ability is increased and the error 1. Reduce misses. If the count value is small, the probability of missing an error in row direction decoding is low;
Increase the error correction ability by increasing the number of correction words,
Reduce uncorrectable errors. Specifically, when the count value is expressed as C.

The number of error correction words is determined as shown in FIG.

FIG. 1 shows an embodiment of the present invention. The reproduced digital signal input from the input terminal 1 is first processed by the first decoder 2 in the row direction.
Decode the steps. Next, based on this decoding result, decoders 3, 4 .
5 performs the second stage decoding, and stores the decoded data in memory 6.
．． 7. Store in 8. The counter 9 counts the number of columns that cannot be corrected by the decoder 4 in the column direction, for example, for each product code unit. Then, as shown in FIG. 5, the selector 10 selects one of the memories 6, 7, and 8 according to the count value C, and
The decoded data is output from the output terminal 11.

FIG. 2 shows another embodiment of the invention. The reproduced digital signal input from the input terminal 13 is first input to the decoder 1 in the row direction.
4 performs the first stage decoding. Next, based on this decoded data, the column-direction decoder 15 performs a second correction with a correction word count of 3.
Decode the steps. The counter 16 counts the number of times that the decoder 15 in the column direction becomes uncorrectable, for example, for each product code unit.

Then, depending on the count value, the decoder 1 in the column direction is
5 sets the number of error words to be corrected, decodes again, and outputs the decoded data from the output terminal 17.

〔Effect of the invention〕

According to the present invention, since the area to be corrected by replacement without correction is controlled according to the error occurrence state, the area to be replaced can be made small and errors can be prevented from being overlooked.

[BRIEF DESCRIPTION OF THE DRAWINGS] Fig. 1 shows an embodiment of the present invention, Fig. 2 shows 1.13
...Input terminal, 11.17...Output terminal, 2゜14
... row direction decoder, 3, 4, 5.15 ... column direction decoder, 9.16 ... counter, 10 ... selector. 30...recording or transmission data, 31.32...parity for error correction.

Claims (1)

[Scope of Claims] 1. An error correction method characterized by counting the number of times errors are detected every fixed period in decoding an error correction code, and controlling the error correction operation based on the counted value. 2. In claim 1, a first stage error correction and detection decoder, a second stage error correction and a plurality of decoders different in the number of error corrections to be detected, and a second stage error correction and detection decoder are provided. It has a circuit for storing error correction data by the decoder in the second stage, and a circuit for counting the number of times a specific decoder in the second stage detects an error every fixed period, and stores it in the storage circuit according to the counted value. An error correction method characterized in that one of a plurality of sets of error correction data is output. 3. In claim 2, the second stage decoder consists of one decoder, the first decoding is performed with a predetermined number of error corrections, and the second decoding is performed according to the count value. An error correction method characterized in that the number of error corrections is changed using the same decoder. 4. It has a row-direction decoder, a column-direction decoder, and a counter, and counts the number of times the row-direction decoder cannot make corrections at each predetermined period, and according to the counted value, the column-direction decoder An error correction device characterized by controlling a correction operation of a decoder. 5. In claim 4, a plurality of column-direction decoders each having a different number of correctable error words and a memory for storing decoding results are provided, and the decoding results of the column-direction decoders are An error correction device characterized in that an error correction device selects and outputs one of the following based on a count value.