JP2957190B2 - Error correction device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Industrial applications]

The present invention provides an error correction method for correcting an information code to minimize the effect of an error when an error generated in data obtained by adding a protection code to the information code is detectable by the protection code and cannot be corrected. It concerns the device.

[Prior art]

Figure 3 is based on the error correction method described in the IEICE 70th Anniversary Commemorative National Congress Lecture Book Volume 6.1323, "Effects of Sound Quality Improvement by Interpolation in PCM Speech" (issued March 15, 1987). FIG. 1 is a configuration diagram showing a conventional error correction device. In the figure, reference numeral 1 denotes a protection capable of correcting N bits, for example, a 1-bit error, and detecting M bits larger than N bits, for example, a 2-bit error in an information code having some correlation between adjacent codes. An input terminal to which data with a sign is input, and an output terminal 2 to which corrected data is output. Reference numeral 3 denotes error detecting means for detecting the occurrence of a one-bit error in the data input to the input terminal 1. Reference numeral 4 denotes a device for correcting the error when the data input to the input terminal 1 has an error. It is a syndrome calculating means for calculating a syndrome. Reference numeral 5 denotes an error correction unit that performs one-bit error correction of data using the syndrome calculated by the syndrome calculation unit 4. 6 calculates the average value of the information codes of the preceding and succeeding data when the two-bit error is detected, if the information codes of the preceding and succeeding data are known. This is an interpolation means for outputting the signal itself to the output terminal 2.
Reference numeral 7 denotes switching means which is switched by the error detecting means 3 and the syndrome 4, and selectively supplies data input to the input terminal 1 to the error correcting means 5, the interpolating means 6, or the output terminal 2. Next, the operation will be described. FIG. 4 is a flowchart showing the procedure of the operation. When data obtained by adding a protection code to an information code arrives at the input terminal 1, the data is input to the error detection means 3 and the syndrome calculation means. When data is input, the error detection means 3 first checks whether a 1-bit error has occurred (step ST1). As a result, when a one-bit error is detected, the switch means 7 is switched to "a", and the syndrome calculated by the syndrome calculation means 4 is sent to the error correction means 5 to perform one-bit error correction (step S1). ST2). The information code corrected by the error correction means 5 is output via the output terminal 2. Conversely, if a one-bit error is not detected, it is checked whether the syndrome calculated by the syndrome calculating means 4 is "0" (step ST3). If the syndrome is "0", an error is detected. As a result, the switch means 7 is switched to "c", and the input information code is output via the output terminal 2 as it is. If the syndrome is not "0", it is determined that there is a 2-bit error, the switch means is switched to "b", and the input information code is sent to the interpolation means 6 to perform an interpolation process (step ST4). The information code obtained by the interpolation processing by the interpolation means 6 is output from the output terminal 2 as a correction code. Therefore, when a two-bit error is detected for a highly continuous information code, an extremely large error can be eliminated by replacing the erroneous information code with the average value of the preceding and succeeding information codes or the preceding information code.

[Problems to be solved by the invention]

Since the conventional error correction device is configured as described above, it is determined that the correct information code is different from the incorrect information code by two bits, but the correct information code is corrected to a different information code. In particular, when the value of the information code does not correspond to the physical value itself, such as when the audio information is compressed using the vector quantization method and the vector code number is used as the information code, Is to correct the decoded data by averaging the proximity data on the physical value,
There is no alternative but to replace the previous information code, and there is a problem that the value is not a large error but is definitely corrected to a value considered to be incorrect. For this reason, there is also a method of obtaining some information codes close to the average value of values before and after, and selecting an information code having the same syndrome as the calculated syndrome and having a two-bit error. According to the method, it is not always possible to reliably obtain an information code having the same syndrome and a 2-bit error, and it has not been a fundamental solution to the problem. SUMMARY OF THE INVENTION The present invention has been made to solve the above-described problems, and an error correction device that does not correct an information code that has detected an M-bit error into an information code that is clearly wrong is provided. The purpose is to gain.

[Means for Solving the Problems]

An error correction device according to the present invention includes: an information code deriving unit that derives a finite number of information codes that are likely to be correct information codes based on an information code and a protection code when an error larger than N bits is detected. And an information code selecting means for selecting an information code having the best correlation with an adjacent information code from the information codes derived by the information code deriving means.

[Operation]

The information code deriving unit in the present invention derives a finite number of information codes that are likely to be correct information codes, selects an information code for correction from among the finite number of information codes, and generates an error code. The information code is not corrected to an obvious information code.

【Example】

An embodiment of the present invention will be described below with reference to the drawings. In FIG. 1, 1 is an input terminal, 2 is an output terminal, 3 is an error detection means, 4 is a syndrome calculation means, 5 is an error correction means, and 7 is a switch means, which are denoted by the same reference numerals in FIG. The detailed description is omitted because it is the same as or equivalent to the conventional one described above. Reference numeral 8 denotes an M-bit error (for example, a 2-bit error) in the data input to the input terminal 1.
In the case where there is, from the syndrome calculated by the syndrome calculating means 4 for N-bit error correction (for example, 1-bit error correction), all the information that is likely to be a correct information code based on the information code and the protection code is obtained. This is an information code deriving unit that derives a code. 9 is the information code deriving means 8
Means for selecting an information code having the best correlation (for example, continuity before and after) with respect to an adjacent information code from the information codes derived by the above. Next, the operation will be described. FIG. 2 is a flowchart showing the procedure of the operation. The data arriving at the input terminal 1 is input to the error detecting means 3 and the syndrome calculating means 4, and the error detecting means 3 detects the presence or absence of a one-bit error (step ST1). When a one-bit error is detected, the error correcting means 5 performs one-bit error correction using the syndrome calculated by the syndrome calculating means 4 (step ST2), and outputs the corrected information code from the output terminal 2. I do. If no one-bit error is detected, it is checked whether the syndrome calculated by the syndrome calculating means 4 is "0" (step ST3). If the syndrome is "0", an error is detected. The information code input as the absence of the information code is output via the output terminal 2 as it is. If the syndrome is not "0", the switch means 7 is switched to "b" assuming that it is a 2-bit error, the input information code is sent to the information code derivation means 8, and the syndrome is calculated by the syndrome calculation means 4. All combinations of two-bit errors that give the syndromes are calculated, and all information codes corresponding to each two-bit error are derived (step ST
Five). The information code calculated by the information code deriving means 8 is input to the information code selecting means 9, and the information code selecting means 9 has the continuity of the preceding information code or the preceding and succeeding information codes among the information codes. A good information code is selected (step ST6). The information code selected in this way is output from the output terminal 2 as a correction code. The information code deriving means 8 can be realized by generating all 2-bit errors and checking whether or not the syndromes match.
In the case of the (7,4) code, there are only four types of two-bit errors that give one syndrome, and by calculating these in advance, the processing of the information code deriving means 8 is simplified. In the above embodiment, the number of error-correctable bits N = 1,
Although the case where the number of error detectable bits M = 2 has been described, the same applies to other cases of N and M. Further, in the above embodiment, one-dimensional audio information only in the time direction has been described. However, for information of two or more dimensions such as image information, the processing in the information code selecting means 9 is performed based on the relationship with the preceding and following information codes. It can be applied by expanding like front, back, left and right.

【The invention's effect】

As described above, according to the present invention, the correction code is selected only from the information codes that are likely to be the correct information codes. Can be reduced, the number of errors can be reduced as compared with the conventional correction method, and by performing selection using the relationship with the adjacent information code,
There is an effect that a large error is removed.

[Brief description of the drawings]

FIG. 1 is a block diagram showing an error correction apparatus according to an embodiment of the present invention, FIG. 2 is a flowchart showing the operation thereof, and FIG.
FIG. 1 is a block diagram showing a conventional error correction device, and FIG. 4 is a flowchart showing its operation. 3 is an error detecting means, 4 is a syndrome calculating means, 5 is an error correcting means, 8 is an information code deriving means, and 9 is an information code selecting means. In the drawings, the same reference numerals indicate the same or corresponding parts.

Claims (1)

(57) [Claims] An information code having a certain correlation between adjacent codes is provided with a protection code capable of correcting an error of up to N bits and detecting an error of up to M bits larger than N bits. Error detecting means for detecting an error;
Information code deriving means for deriving a finite number of information codes that are likely to be correct information codes based on the information code and the protection code when the error detection means detects an error larger than N bits; An error correction apparatus comprising: an information code selection unit that selects an information code having the best correlation with an adjacent information code from information codes derived by an information code derivation unit.