JPH02146820A - Double reed solomon code decoding system - Google Patents

Abstract

PURPOSE:To evade an error at decoding by utilizing the correction capability of a double reed Solomon code to the utmost. CONSTITUTION:A double reed Solomon code subject to the external code processing with a minimum distance of 3 and the internal code processing with a minimum distance of 3 between codes is subject to error correction of 2 or below at internal code decoding, an inner flag is set with error '0' or '1', erasure correction with number the errors of >=2 is applied when the number of errors to which a flag is set is <=2 at the external code decoding to set an external flag, and at the disabled correction, the external flag is set. Moreover, in a data in which both the internal and the external flags are set, the result is deducted by the interpolation processing and a data whose one flag is reset is adopted as it is. Since the strategy of decoding the external code is selected in response to the inner error flag, the code error correction utilizing the correction capability of the double reed Solomon code to the utmost is attained and since the propriety of the final correction is judged from the result of decoding the inner code and the external code, the danger of processing an error of excess of the correction capability as within the correction in the occurrence of the said error is evaded in an excellent way.

Description

DETAILED DESCRIPTION OF THE INVENTION [Industrial Application Field] The present invention is a double Reed-Solomon code decoding system that allows decoding to take full advantage of the correction ability of the double Reed-Solomon code.

[Prior art] A digital VTR that converts video and audio signals into digital signals and records them on magnetic tape via a rotating magnetic head.
Here, we combine the TDM method, which compresses the color signal in the time axis and inserts it into the horizontal retrace period of the luminance signal, and the DPCM method, which converts and encodes the difference signal between the predicted value and the input value and transmits it.
A method of compressing the recording band is often used. In the DPCM method, an image of one field is divided into, for example, 2550 blocks, and each block is further divided into 64 blocks in a matrix.
After partitioning into sample pixels, the video data for each pixel is sampled with a quantization bit number of 2 to 5 depending on the specified mode. The sampled video data and the mode data indicating the quantization mode are both made up of 8 bits as one symbol, and after being arranged on a video memory map according to certain rules, a double Reed-Solomon code ( Double Reed Sol
For example, while reading out in batting order or column order! Recording is performed on a magnetic tape according to the track 1/2 field recording method or the like.

On the other hand, for audio signals, the left and right channel data is, for example, 47.952K, which is 800 times the field frequency.
After sampling at a sampling frequency of Hz and quantizing at 16 bits, one sample for both the left and right channels is divided into the top 8
Divide into bit and lower 8 bits. Next, the audio data symbolized in 8-bit units is arranged on an audio memory map according to certain rules, subjected to double Reed-Solomon encoding processing according to the matrix of the map, and then written to a magnetic tape while being read out in batting order or column order. Record.

By the way, the double Reed-Solomon code processing used in this type of digital VTR consists of C encoding (number of symbols; information length 30, total length 32) with a minimum distance between codes called outer code processing, and inner code processing. C1 encoding with a minimum distance between codes of 5 (number of symbols: information length 66,
total length 70), etc. are being considered as candidates for code error correction processing methods, and the focus is not only on the selection of correction code sequences, but also on how to efficiently decode the double Reed-Solomon code added during recording in the signal reproduction system. There are also decoding methods such as 1,
This was becoming an important consideration that would determine the success or failure of the commercialization of digital VTRs.

[Problems to be Solved by the Invention] In the conventional double Reed-Solomon code decoding system, from the viewpoint of the correction ability of the double Reed-Solomon code, inner code decoding produces 4 errors when the error position is known, but 4 errors when the error position is known. If the error location is known, it is possible to correct up to 2 errors, and in outer code decoding, if the error location is known, it is possible to correct 1 error, and if the error location is known, it is possible to correct up to 2 errors. .

However, conventional decoding methods simply correct 2 or less errors in both decoding processes due to the consideration of using the same calculation algorithm in both the inner code and outer code decoding processes. . For this reason, there is a problem that the correcting ability for burst errors is insufficient, and the high correcting ability of the double Reed-Solomon code prepared as a recording format for digital signals is not fully utilized.

[Means for Solving the Problems] This invention solves the above problems, and is a double Reed-Solomon system that has undergone outer code processing with a minimum inter-code distance of 3 and inner code processing with a minimum inter-code distance of 5. It is a double Reed-Solomon code decoding method that decodes the code in the order of the inner code and the outer code, and in the inner code decoding process, error correction of 2 or less is performed, and the inner error flag is reset when there is no error or when 1 error is corrected. At the same time, an inner error flag is set when 2 errors are corrected or when correction is impossible, and in the outer code decoding process, when the number of errors for which the inner error flag is set is 2 or less, erasure correction of 2 or less is performed, and the error is detected. If there is no error, the outer error flag is reset during error correction or 2 erasure correction, and the outer error flag is set when correction is impossible, and data for which both the inner error flag and outer error flag are set is processed by interpolation processing. This method is characterized in that data for which one of the error flags has been estimated and has been reset is adopted as is.

[Operation] This invention corrects errors of 2 or less in the inner code decoding process for a double Reed-Solomon code that has undergone outer code processing with a minimum intercode distance of 3 and inner code processing with a minimum intercode distance of 5. , the inner error flag is reset when there is no error or 1 error is corrected, and the inner error flag is set when 2 errors are corrected or correction is impossible, and in the outer code decoding process, the number of errors for which the inner error flag is set is When the error is 2 or less, erasure correction of 2 or less is performed, and the outer error flag is reset when there is no error or 1 error correction or 2 erasure correction is made, and the outer error flag is set when correction is not possible. Data for which both error flags are set is estimated by interpolation processing, and data for which either error flag is reset is adopted as is, so that the double Reed-Solomon code originally possesses the same characteristics. Maximize and utilize your correction ability.

[Example] Hereinafter, an example of the present invention will be described with reference to FIG. 1.2. FIG. 1.2 is a flowchart for explaining each embodiment of the inner code decoding method and the outer code decoding method to which the double Reed-Solomon code decoding method of the present invention is applied.

First, video signals and audio signals to be decoded are first encoded (number of symbols; length 30.total length 32), followed by 01 encoding (number of symbols: information length 66, total length 70) with a minimum inter-code distance of 5, called inner code processing, along the row direction. Assume that the inner code is decoded first, and then the outer code is decoded.

In the inner code decoding process, prior to code correction, syndromes 80 to S are calculated from code data A0 to Ao of 70 symbols constituting l block, and these four syndromes 86 to S3 are calculated using primitive polynomials G (x ) = x
”+ x '10x 3+ x''+1 root, α=[00
000010], So= ' Ass+ A@@ten・・+
At+A.

SI-α69A,. +α”Ass+...+α A +
+A tsSt=α13@A,. +αIs@A,,+
11 11+α”A, +A.

S, =a″”A @@+ (X ”’A @@+ ・・
+ ff 3A I + A It is given by a.

Then, steps (101) to (104) shown in FIG.
), if it is found that all four syndromes 80 to S are zero, it is determined not to perform correction, and in step (105), the internal error flag Fl is reset.

On the other hand, S , = S , = 0 big S, ≠ O or S3
If ≠O, it is determined that the error cannot be corrected, and an internal error flag PI is set in step (106). On the other hand, if So≠0, the step (101
), if it is found in steps (107) to (
109), s +”+ s OS t= 0
, S oS! + S ts t= 0, S t
'+ S + S s-〇, in step (110), the error position
After that, one symbol error correction is performed in the following step (111). In addition, in the one-symbol error correction, the error position X and the error pattern Y are 50=YIS+-X+YISt"XI"Yrs,=x,'y.

Because of the relationship, in the end x l = s 1/s.

is obtained, and at this time, the inner error flag Fl is reset.

By the way, steps (107) and (112).

(113) For i, S +”+S as !≠0.
5oS3+ S Is? ≠0.9*”+51S3≠0
When it is determined that the following three coefficients a, b, and c are not zero, step (11
In 4), the coefficients B+ and Bt of the error position equation are calculated. The error position equation is 4 that holds between the error position X,,X, and the error pattern Y,,Y, in two-error correction.
Formula % Formula % <Y, , Y, obtained by eliminating a X 1" 10b X 1+ c = Oa
The original form is a quadratic equation where c = 0. However, each coefficient a, b, c is a = S 1 ” + S 1) S tb = sos
s+s+s*c=s-+S, Ss.

Note that in 2-error correction, the coefficient a of the quadratic term must not be zero, so Bl= (SoSs+5tSt)/(S+”+5oS
t)Bx= (st′+5tss)/(s+”+5o
By introducing the coefficients B, , B, with st), the error positions X+ and Xt can be found as the two roots of the following equation.

W ” + B + W + B v= 0 However, here, instead of solving the error position equation immediately from the coefficient B1.Bt, by applying the following linear transformation, only the constant term is made a function of the syndrome, and then A conversion table indexed by a constant term is used.In other words, the quadratic equation regarding W described above is linearly converted with W=B, Z,
An error position equation is obtained in which only the constant term is a function of the syndrome 80~S, such as Z"+Z+(B*/Bi")=0.

In this final error position equation, the syndrome S0~
Constant B1 calculated from S3. Step Bx (114
), and the solution Z can be read out according to the conversion table indexed by the constant term Bm/B,". Then, in the following step (115), using the solution Z obtained in the previous step, X + = B r Z , X
By setting −B + + X t, the error position Xl,
Find X*.

The error positions X, , X, thus determined are evaluated for validity from multiple angles in the following step (116). That is, the error positions X+ and Xt found in step (115) should originally be included in the 70 symbols, and should not be imaginary solutions, so singular solutions that do not meet these conditions is the step (1
16). That is, if as a result of checking the validity of the obtained error positions X, , X, it is found that they are not valid, the process moves to step (106) and it is determined that they cannot be corrected.

Furthermore, if it is determined that the error positions X, , X, are valid, two-error correction is performed in step (117). In this case, the error pattern Y,, Y, is Y 1 = (S + + X r So ) /
(X t + X + )Y l= S a + Y
*, and the inner error flag F1 is set.

In this way, code data that has undergone a maximum of two error corrections during the inner code decoding process is used for subsequent outer code decoding.

In the outer code decoding process, first, 32 symbols of code data A0 to A3 . Kara syndrome So
, St are calculated, but the two syndromes So and Ss are the primitive polynomial G (x)-x @+ x'+X3+
When the root of X″+1 is α=[00000010], S o = A s + +A s.
+A I+A.

S1=α"A3+++α30 A3゜+ ・ ・ +α
A + + A.

is given by

In outer code decoding, since the above two equations hold, it can be seen that error correction of up to two symbols is possible if the error position is known, and error correction of one symbol is possible otherwise.

That is, as shown in FIG. 2, step (121)
and (+22), syndrome S. , S are both zero, it is determined in step (123) that there is no error, and the outer error flag is reset to F=0.

On the other hand, when S, f: O and S1≠0 and the number of internal error flags PI is 1, steps (124) and (125
), it is determined that there is an l-symbol error, and in step (126), the error position, X, is determined as S, /S. If it is found that the internal error flag PI is set at the determined error position xl, then in step (128) following the judgment step (127), Y,
=S, close! Perform error correction and reset outside error flag F.

On the other hand, if S0≠O,S,=O, the judgment step (1
In step 29), it is checked whether the number of internal error flags F1 is 2, and when it is found that it is, it is determined that there is a 2-symbol error, and in step (130), the internal error flag PI is set. The error position is X, ,X,
shall be. Then, in the following step (131), the outer error flag F is reset and 2 erasure correction is executed. 2 erasure correction, S o, S +, X
Since t, X t are known, the error pattern is Yl-(S++x+so)/(Xt+X+)Y 1
It is determined as = S o + Y t .

Furthermore, negative results determined in steps (127) and (129) are determined to be uncorrectable, and an outside error flag F is set in step (132).

In this way, when the decoding of the outer code following the inner code is completed, data for which both the inner error flag Fl and the outer error flag F are set cannot be corrected completely, so interpolation processing such as intermediate value interpolation is performed. presume. However, data for which either one of the error flags Fl or F has been reset is assumed to have been corrected and is used as is.

In this way, according to the above-mentioned double Reed-Solomon code decoding method, when decoding a double Reed-Solomon code that has undergone outer code processing with a minimum inter-code distance of 3 and inner code processing with a minimum inter-code distance of 5, By selecting the outer code decoding strategy according to the inner error flag F1 revealed in the inner code decoding process, it is possible to perform code error correction that makes full use of the correction ability of the double Reed-Solomon code. Furthermore, since the final decision on whether or not correction is possible is made based on the decoding results of the inner code and outer code, there is a risk that if an error that exceeds the correction ability occurs, it will be mistakenly corrected as an error within the correction range. can be successfully avoided.

In addition, by reading out video data and audio data from the memory map in bat order or column order, which are assigned double Reed-Solomon codes in the row and column directions after being arranged on the memory map, sampling time can be adjusted. In cases where similar data is recorded at a distance from each other, it is possible to provide either the outer code or the inner code with random error correction capability, while the other has burst error correction capability, which allows for comprehensive It is particularly effective for digital recording and reproducing equipment such as digital VTRs that digitally record video and audio signals.

[Effects of the Invention] As explained above, this invention has the advantage that the minimum distance between codes is 3.
A double Reed-Solomon code that has been subjected to outer code processing and inner code processing with a minimum intercode distance of 5 is corrected for errors of 2 or less in the inner code decoding process to ensure that there are no errors or! The inner error flag is reset when an error is corrected, and the inner error flag is set when two errors are corrected or cannot be corrected.In the outer code decoding process, when the number of errors for which the inner error flag is set is less than or equal to 2, the inner error flag is set. performs erasure correction, and resets the outer error flag when there is no error or when one error correction or two erasure corrections are made;
The outer error flag is set when correction is impossible, and data for which both the inner error flag and outer error flag are set is estimated by interpolation processing, and data for which either one of the error flags is reset is left as is. By selecting the outer code decoding strategy according to the inner error flag revealed during the inner code decoding process, code error correction that takes full advantage of the correction ability of the double Reed-Solomon code is possible. Furthermore, since the final decision on whether or not correction is possible is made based on the decoding results of the inner code and outer code, if an error that exceeds the correction capability occurs, it can be incorrectly corrected as an error within the correction range. Therefore, for example, after arranging the data on the memory map, it is possible to avoid the risk of
By reading out video data and audio data with multiple Reed-Solomon codes from the memory map in stroke order or column order, the outer code can be used when data with similar sampling times are recorded at a distance from each other. Alternatively, one of the inner codes can have random error correction ability while the other one has burst error correction ability.This allows extremely high code correction ability to be exhibited overall, and is useful for video signals and audio. This provides excellent effects, such as being particularly effective in digital recording and reproducing equipment such as digital VTRs that digitally record signals.

[Brief explanation of the drawing]

FIG. 1.2 is a flowchart for explaining each embodiment of the inner code decoding method and the outer code decoding method to which the double Reed-Solomon code decoding method of the present invention is applied.

Claims (1)

[Claims] Outer code processing with minimum inter-code distance of 3 and minimum inter-code distance of 5
This is a double Reed-Solomon code decoding method that decodes a double Reed-Solomon code that has undergone inner code processing in the order of the inner code and outer code, and in the inner code decoding process, error correction of 2 or less is performed and there is no error. Or, when one error is corrected, the inner error flag is reset, and when two errors are corrected or when correction is impossible, the inner error flag is set, and in the outer code decoding process, the number of errors for which the inner error flag is set is 2.
Performs erasure correction of 2 or less in the following cases, resets the outer error flag when there is no error or corrects 1 error or 2 erasures, sets the outer error flag when correction is impossible, and then sets the inner error flag and outer error flag. For data with both flags set,
A double Reed-Solomon code decoding method characterized in that data estimated by interpolation processing and for which one of the error flags is reset is adopted as is.