JPH04278727A - Error correction method and audio decoding and arq transmission methods using said method - Google Patents

Abstract

PURPOSE:To reproduce more powerful data of high reliability by using not only redundancy obtained by error correction code encoding but also redundance of an information source itself to decode the error correction code. CONSTITUTION:An error correction code word consisting of k information bits and a check bit is received. Information bits has 2<k> kinds of bit pattern. A probability qi(i=1,2 to 2<k>) that each information bit pattern is sent is obtained in accordance with an error correction code encoding rule and the reception error correction code word. A probability pi(i=1,2 to 2<k>) that each information bit pattern is generated on the transmission side is obtained in accordance with the property of the information source and reception information bit pattens before the preceding frame. The place (i) which gives a maximum value max(pi.qi) of products of propabilities pi and qi is obtained. The information bit pattern corresponding to the obtained place (i) is determined and stored to decode the error correction code.

Description

[Detailed description of the invention]

0001

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an error correction method used in a transmission device such as a digital car telephone, and a voice decoding method and ARQ transmission method using the error correction method.

0002

2. Description of the Related Art FIG. 4 shows an algorithm for an error correction method on the receiving side of a conventional transmission apparatus of this type. In FIG. 4, first, in step 41, an error correction code word consisting of information bits (the number of bits is k bits) and check bits sent from the transmitting side is received. Check bits are bits added to information bits so that they satisfy certain error correction coding rules, and even if a bit error of a certain number or less occurs in the transmission path, it can be corrected correctly on the receiving side. This is to make it possible to reproduce the correct information bits. Since the information bits are k bits, the bit pattern (combination of bits) of the information bits
There are a total of 2k ways.

Next, each error correction code word (2k
The probability that each information bit pattern would have been transmitted qi (i=
1, 2, 3...2k; i represents the number of the information bit pattern) is calculated in step 42, and then in step 43 the information bit pattern that gives the maximum value is determined,
Further, in step 44, an information bit pattern is determined.

In reality, it is sufficient to find i that gives the maximum value of qi in step 43, so there is no need to find the exact value of qi in step 42, and the received error correction code word and the error with the shortest Hamming distance are All that is needed is to determine which information bit pattern provides the correction code word. Generally, operations in steps 42, 43 and 44 are performed at once using an error correction decoding algorithm depending on the type of error correction code.

[0005] In this way, even in the conventional error correction method described above, bit errors occurring in the transmission path can be correctly corrected to a certain extent by utilizing the redundancy of the error correction code word due to the addition of check bits to the information bits. can do.

[0006]

However, in the above-mentioned conventional error correction method, the probability of occurrence of each information bit pattern of information bits (k bits) depends on the characteristics of the information source and the information bit pattern of the previous frame and the current frame. Because the information bit patterns are not uniform due to their correlation, and the difference in the probability of occurrence of each information bit pattern is not used in error correction, the error correction method uses only redundancy by adding check bits. There is a problem in that when the bit error rate of the transmission path is high, there is a high probability that correct error correction will not be performed.

[0007] The present invention solves such conventional problems by comparing the redundancy of the information source itself, that is, the difference in the probability of occurrence of each information bit pattern, with the redundancy obtained by error correction coding. It is an object of the present invention to provide an error correction method that can perform more powerful error correction when used together, and an audio decoding method and an ARQ transmission method using the same.

[0008]

[Means for Solving the Problems] In order to achieve the above object, the present invention, in error correction decoding on the receiving side, uses the redundancy of the information source itself in addition to the redundancy due to error correction coding. Error correction decoding is performed using the .

[0009]

[Operation] Therefore, according to the present invention, by utilizing the redundancy of the information source itself for error correction decoding in addition to the redundancy due to error correction coding, the error correction is more powerful and reliable. This has the effect that a correction method can be realized.

0010

DESCRIPTION OF THE PREFERRED EMBODIMENTS FIG. 1 shows an algorithm for an error correction method in a first embodiment of the present invention. In FIG. 1, first, in step 11, the information bit (
An error correction codeword consisting of a check bit and a check bit is received. Since there are k bits of information bits, there are a total of 2k bit patterns (combinations of bits) of the information bits.

Next, each error correction code word (2k
The probability that each information bit pattern would have been transmitted qi (i = 1
, 2, 3...2k; i represents the number of the information bit pattern. ) is calculated in step 12.

Next, the characteristics of the information source (each information source, such as an audio signal or a fax signal, has its own characteristics of generating information bits), and the information bit pattern of the previous frame and the information bit pattern of the current frame. Based on the correlation between ) is calculated in step 13.

Next, in step 14, step 12
, 13, the product of pi and qi is pi ・
Find qi (i=1, 2, 3...2k), and find i that gives the maximum value max(pi.qi) of pi.qi.

Once i is determined, in step 15, the information bit pattern corresponding to i is determined and stored. The stored information bit pattern is used as a parameter for determining pi (i=1, 2, 3...2k) in step 13 for the information bit pattern of the next frame.

[0015] As described above, according to the error correction method in the first embodiment, in addition to the redundancy obtained by error correction encoding, error correction can be performed using the redundancy possessed by the information source itself. Yes, more powerful error correction can be performed.

Note that this error correction method is particularly effective when the information source is a voice signal or a fax signal, where the probability of occurrence of information bit patterns is biased and the correlation with the previous frame is relatively high. be.

FIG. 2 shows a second embodiment of the audio decoding method of the present invention using the error correction method of FIG. Steps 21 to 25 are the same as steps 11 to 15 in FIG.
Represents an operation to obtain i that gives ax(pi qi ). Next, in step 26, if the maximum value max(pi ・qi) of pi ・qi is less than the predetermined threshold ε (0<ε≦1), the information bit pattern of the current frame has a bit error. Considering that there is a high probability that there remains, the normal audio decoded output in step 27 is not performed, but the audio decoded output is performed in step 28 by interpolation or the like using the information bit pattern of the previous frame.

As described above, according to the audio decoding method in the second embodiment, the probability that the information bit pattern of a frame in which a bit error remains remains is decoded and output as is, resulting in an abnormal sound being output. can be reduced.

FIG. 3 shows an ARQ (Automatic Repeat Request) transmission method, which is a third embodiment of the present invention, using the error correction method of FIG. Steps 31 to 35 are Step 1 in Figure 1
1 to 15, and by using the same error correction method, p
It represents an operation to find i that gives the maximum value max(pi ・qi ) of i ・qi . Next, in step 36,
The maximum value max(pi ・qi) of that pi ・qi
does not satisfy the predetermined threshold (0<ε≦1), it is assumed that there is a high probability that bit errors remain in the information bit pattern of the current frame, and the information bit pattern of the current frame is changed in step 37. Instead of outputting the pattern as it is, a request for retransmission of the information bit pattern of the current frame is made in step 38.

[0020] In this way, A in the above third embodiment
According to the RQ transmission method, it is possible to reduce the probability that information bits of a frame in which bit errors remain are used as received data signals as they are.

[0021] In FIGS. 1, 2 and 3, p
The maximum value max(pi ・qi) of i ・qi is calculated, but if k becomes a large value, the scale of calculation to obtain all 2k ways of pi and qi becomes very large, so i=1, 2, 3...Instead of finding pi and qi for all of 2k, we find i that gives some of the values of pi in descending order and i that gives some of the values of qi in order of descending order, and find i that both exist. The maximum value of pi ・qi max(pi ・qi )
A quasi-optimal method may be adopted in which i that gives the value i is set as the i to be found.

[0022]

[Effects of the Invention] As is clear from the above embodiments, the present invention provides, in addition to the redundancy obtained by error correction coding,
Since error correction decoding is performed on the receiving side by also utilizing the redundancy of the information source itself, a more powerful and reliable error correction method, an audio decoding method using it, and ARQ are possible.
A transmission method can be realized.

[Brief explanation of the drawing]

FIG. 1 is a flowchart showing an algorithm of an error correction method in a first embodiment of the present invention.

FIG. 2 is a flowchart of a speech decoding method using the error correction method of FIG. 1, showing a second embodiment of the present invention.

FIG. 3 shows a third embodiment of the present invention using the error correction method of FIG.
Flowchart of RQ transmission method

[Figure 4] Flowchart showing the algorithm of a conventional error correction method

[Description of codes] k Number of information bits i Number of information bit patterns qi Probability that each information bit pattern would have been transmitted, determined from the error correction coding rule and the received error correction code word, Pi Characteristics of the information source and Probability that each information bit pattern would have occurred on the transmitting side, determined from the received information bit pattern before the previous frame

Claims (3)

[Claims] Claim 1: When receiving and error-correcting decoding an error correction code word in which check bits are added to k bits of information bits per frame, among the 2k information bit patterns, received information bits from previous frames are The probability of occurrence pi (i=1, 2, 3...2k) of each information bit pattern determined from the pattern and the nature of the transmitted information, and each information bit determined from the received code word bit pattern and line bit error rate. The probability that the pattern was sent qi (i=1
. [Claim 2] When the information to be transmitted is a highly efficient encoded digital audio signal, if the maximum value of pi qi is less than a certain value, the information bits of the frame are not audio decoded, but the information bits of the previous frame are decoded. 2. A speech decoding method using the error correction method according to claim 1, wherein said information bits are also used to perform speech decoding processing. 3. In an ARQ transmission system that can request the transmitting side to retransmit the information bits when a bit error is detected in the information bits obtained by the receiving side,
- An ARQ transmission method using the error correction method according to claim 1, wherein retransmission is requested when the maximum value of qi is less than a certain value.