JP2536861B2 - Multistage decoding method - Google Patents

Description

The present invention relates to a Reed-Solomon code (in order to correct a data error that occurs when data recorded on a recording medium such as a magnetic tape or a magnetic disk is reproduced. The present invention relates to a decoding method when an error correction code such as RS code) is used.

[Conventional technology]

Generally, when recording and reproducing data on a recording medium such as a magnetic tape or a magnetic disk, a data error occurs depending on the state of the recording medium. Data errors include burst errors due to dropouts and random errors due to S / N deterioration. In order to correct these errors, a two-stage coded error correction code is often used. As an example, q
Consider decoding of a two-stage code using an RS code on GF (2 ⁸ ) when = 8. There is a two-stage encoding device shown in FIG. (61) is a C ₂ encoder, (62) is an interleave circuit, and (63) is a C ₁ encoder. First, C ₂ encoding is performed on the original data, then interleaved, and then C ₁ Encode and output. A two-stage decoding device is shown in FIG. In the figure, (71) is C ₁
Decoder, (72) is a deinterleave circuit, (73) is a C ₂ decoder, which performs deinterleave after C ₁ decoding,
Next, C ₂ decoding is performed. Here, as a conventional two-stage encoding method, as shown in FIG. 5, N ₁ × 8 digits in the X direction,
When the data is divided into 8 dimensions in the X direction with respect to the data arranged in two dimensions of N ₂ digits in the Y direction, and the data is treated, after adding the first check code of K ₂ -N ₁ ,
There are two stage encoding method for adding a second check code K ₁ -K _2. Let C ₂ be a (K ₂ , N ₁ ) RS code and C ₁ be a (K ₁ , K ₂ ) RS code. The (α, β) RS code is an RS with information length β and code length α.
Indicates the sign.

Next, specific decoding will be described with reference to FIG. For example, when recording digital voice data on a magnetic tape using a helical scan video tape recorder (hereinafter referred to as VTR), the quantization bit number of voice data is set to 16 bits and the sampling frequency is set to 48 KHz and 2 channels are recorded. In case of NTSC standard TV signal,
Data of 800.8 samples are included in one field period.
If 6 sample data are arranged in the X direction, it will be 134 in the Y direction.
Will be needed. If each sample data is handled in units of 8 bits, 2 × 2 channels will be 4 data, and 4 × 6 = 24 data will be arranged in the X direction. Therefore, N ₁ =
24 As, N ₂ = 134, be performed by the code length 34 in a first direction performs C ₂ encoding code length 30 in a second direction the C ₁ encoding _{thereto, K 2 = 30, K 1} = 34, C ₂ is (30,24) RS code, C ₁ is (3
4,30) RS code.

In the data array, the 1st in the X direction and the mth in the Y direction (l,
Data of m integer, K ₁ ≧ l ≧ 1, N ₂ ≧ m ≧ 1) is represented as a _{l, m} .

First, C ₁ decoding is performed on the data of each row 34 (= K ₁ ) arranged in the first direction. The syndrome is calculated, the position and magnitude of the error are calculated, and the result is output as a flag. This is shifted one by one in the Y direction and repeated in the same manner to perform C ₁ decoding. Then, in the second direction, for example, a _1,1 , a _2,5 ,
_For the 30 (= K ₂ ) data of a _3,9 , ………, a _30,121 ,
Based on performing C ₂ decoding flags C ₁ decoding. The syndrome is calculated from the reproduced data, the position and size of the error are calculated, correction is performed, and a flag is output to the uncorrectable one. This is shifted one by one in the Y direction and repeated in the same manner to perform C ₂ decoding.

For data reproduction, send out 24 (= N ₁ ) data arranged in the first direction while observing the C ₂ decoding flag,
Interpolation is performed on the portion that cannot be error-corrected. With such two-stage decoding, burst error correction
C ₂ code, C _1, C ₂ code is involved for a random error correction, can be efficiently perform error correction.

[Problems to be solved by the invention]

Since the conventional decoding method is configured as described above, since decoding is performed regardless of whether or not encoding is actually performed, erroneous decoding of uncoded data, or If all decoding is performed, there are problems that the processing is complicated and takes time.

The present invention has been made in order to solve the above problems. Since decoding is performed only for the coded one, erroneous decoding can be prevented, and a part of the decoding can be performed. Even if omitted, it is an object of the present invention to provide a multi-stage decoding method capable of simple decoding without having to change other decoding at all.

[Means for solving problems]

The multi-stage decoding method according to the present invention handles N × q digit (N, q: integer) digital information in q digit units, arranges N data in two or more dimensions, and does not overlap in any direction. , From the first direction to the n-th direction (n: integer, n ≧ 2), the i-th encoding (i: integer, 1 ≦ i ≦ n) has a code length Ki (Ki: integer) in the i-th direction. ) In Galois Field G
The code Ci on F (2 ^q ) is configured so that n stages of coding can be performed, and when decoding data to which an identification signal indicating whether each coding has been performed is added, According to the identification signal, it is possible to select whether or not to decode in the corresponding direction.

Another multi-stage decoding method according to the present invention is the multi-stage decoding method according to claim 1, wherein when decoding the data, decoding of some stages out of n stages is omitted. , Is designed to be easily decrypted.

[Action]

According to the present invention, because of the configuration as described above, it is possible to prevent erroneous decoding that decodes unencoded data, and perform decoding for all encoded data. Instead of partially omitting the decoding, it is possible to shorten the decoding time.

Example of Invention

An embodiment of the present invention will be described below with reference to the drawings. n
= 3, the three-stage decoding will be described. Second
The figure is a block diagram showing a three-stage encoder. (21) is
A C ₃ encoder, (22) is an interleave circuit, (23) is a C ₂ encoder, (24) is an interleave circuit, and (25) is a C ₁ encoder. First, C ₃ encoding is performed on the original data,
It is interleaved and C ₂ coded. After further interleaving, C ₁ coding is performed and output. FIG. 3 is a block diagram showing the three-stage decoding apparatus of the present embodiment. (31) is a C ₁ decoder, (32) is a deinterleave circuit, (33) is a C ₂ decoder, and (34) is The deinterleave circuit (35) is a C ₃ decoder. The reproduced data is subjected to C ₁ decoding, deinterleaved, and C ₂ decoded according to the flag information. After further deinterleaving, C ₃ decoding is performed according to the flag information. FIG. 1 is a data array diagram showing the three-stage decoding method of this embodiment. Here as well as the conventional example, NTSC
Helical scan VT that handles standard TV standard signals
Consider the case of recording digital voice data of 16 quantization bits, sampling frequency 48 KHz, and 2 channels in R. Similar to the conventional example, each sample data is handled in 8-bit units, and the data is arranged in two dimensions of 24 (= N ₁ ) in the X direction and 134 (= N ₂ ) in the Y direction, and in the third direction C ₃ with code length 30
If coding is performed, C ₂ coding is performed with a code length of 32 in the _second direction, and C ₁ coding is performed with a code length of 34 in the first direction, K ₃
＝ 30, K ₂ ＝ 32, K ₁ ＝ 34, C ₃ is (30,24) RS code, C ₂ is (32, K
30) RS code and C ₁ is (34,32) RS code. In the data array, the l-th data in the X direction and the m-th data in the Y direction are a _{l, m}
Will be represented.

First, if it is identified that C ₁ encoded in Hetsuda portion relative to the data of the first row 34 arranged in the direction (= K ₁₎ is being performed, it performs a C ₁ decoding. The syndrome is calculated, the position and the size of the error are obtained and correction is performed, and a flag is output for those that cannot be corrected. This is shifted one by one in the Y direction and repeated in the same manner to perform C ₁ decoding.

Next, in the direction of FIG. 2, for example, a _1,1 , a _2,4 , a _3,7 , ...,
a _32,98 of 32 (= K ₂ ) data, in the header part
If it is identified that C ₂ coding is performed perform C ₂ decoding. The syndrome is calculated, the position and size of the error are obtained, correction is performed, and a flag is output for those that cannot be corrected. This is shifted one by one in the Y direction and repeated in the same manner to perform C ₂ decoding. Further, in the direction of FIG. 3, for example, a _1,1 , a _2,5 , a _3,9 , ..., a _30,121 30 (=
If K ₃₎ data to Hetsuda part is identified to C ₃ coding is performed for, performing C ₃ decoding. The syndrome is calculated, the position and size of the error are obtained, correction is performed, and a flag is output for those that cannot be corrected. This is shifted one by one in the Y direction and repeated in the same manner to C ₃
Decrypt. The data is reproduced by sending out 24 (= N ₁ ) data arranged in the first direction while observing the flag output, and interpolating the uncorrectable portion.

Since the multistage decoding method according to the present invention is configured as described above, for example, the number of quantization bits of audio data is
When shifting from 16 bits to 20 bits, the number of data arrays N ₁ lined up in the X direction increases from 24 to 30, but C ₃ coding is discontinued and the check code occupying area by C ₃ coding (the shaded area in Fig. 1). If () is also used as the data area, 30 data can be arranged in the X direction.

Moreover, the entire data array format and capacity are the same, and there is no need to change the decoding method for C ₂ decoding and C ₁ decoding at all. Furthermore, an identification signal indicating that C ₂ coding and C ₁ coding are performed without performing C ₃ coding can prevent erroneous decoding in which decoding is performed in the direction not coded. .

In the encoding device in this case, in FIG. 2, first, the C ₂ encoder (23) performs C ₂ encoding to obtain a (32, 30) RS code, and the interleave circuit (24) performs interleaving, and then C ₁
The encoder (25) performs C ₁ coding to obtain a (34, 32) RS code. The encoder is shown in FIG. 3 as a C ₁ decoder (3
Deinterleave circuit (32) with C ₁ decoding in ₁ )
After deinterleaving in, the C ₂ decoder (33) performs C ₂ decoding. Further, in the above example, the number of quantization bits of voice data is increased, but the number of quantization bits may be left as it is and other data may be inserted.

Further, in the above embodiment, C ₃ encoding and decoding were not performed due to the increase in the amount of information, but even if the encoding is performed, any decoding such as C ₃ is performed depending on the error correction rate and speed of the system. ₃ Without performing the decoding, only the C ₁ and C ₂ decodings may be performed and the simple decoding may be performed.

Although the data array is a secondary array in the above embodiment, it may be a three-dimensional array as shown in FIG. Also,
In a code configured to perform n-stage decoding,
The number of stages without decoding is arbitrarily checked from n-1 stages or more,
Also, it is possible to arbitrarily select which stage.

Further, the data group to be decoded may be either a complete type in which the data within the data array capacity is treated as one unit, or a non-complete type in which the data is sequentially rewritten and encoded.

Also, we considered the case of recording digital audio data in a helical scan type VTR, but the content of the data may be any digital data, the recording medium is tape,
Anything, such as a disk, is fine. Further, the applied device may be other recording device such as a digital audio tape recorder, a floppy disc device, etc. in addition to the VTR. further,
Although the identification signal indicating whether or not the encoding has been performed is added to the header portion, it may be the rear portion of the data, the inside of the data, or a completely different portion.

〔The invention's effect〕

As described above, according to the present invention, the encoding method is configured to enable multi-stage encoding of two or more stages, and the identification signal indicating whether or not each encoding has been performed is added. It is possible to prevent erroneous decryption in the direction that does not perform, and to easily decrypt data by omitting arbitrary decryption, and highly flexible decryption that does not need to be changed for other decryption. There is an effect that the method can be obtained.

[Brief description of drawings]

FIG. 1 is a data array diagram showing a three-stage encoding method according to an embodiment of the present invention, FIG. 2 is a block diagram showing a three-stage encoding device according to an embodiment of the present invention, and FIG. FIG. 4 is a block diagram showing a three-stage decoding apparatus of one embodiment, FIG. 4 is a data array diagram showing a three-stage decoding method of another embodiment of the present invention, and FIG.
FIG. 6 is a data array diagram showing a conventional two-stage decoding method, FIG. 6 is a block diagram showing a conventional two-stage encoding device, and FIG. 7 is a block diagram showing a conventional two-stage decoding device. In the figure, q is the number of digits per data, N is the number of data, K ₁ , K ₂ , ..., Kn are code lengths, and n is the number of coding stages. The same reference numerals in the drawings indicate the same or corresponding parts.

Claims (2)

(57) [Claims] 1. N × q digit (N, q: integer) digital information is treated in q digit units, N data is arranged in two or more dimensions, and the first direction is not overlapped in any direction. To the n-th direction (n: integer, n ≧ 2), the i-th encoding (i: integer, 1 ≦ i ≦ n) is the code length Ki (Ki:
(Integer) is configured so that the code Ci on the Galois field GF (2 ^q ) can be configured to perform n stages of encoding, and an identification signal indicating whether or not each encoding has been performed is added. A multi-stage decoding method characterized in that, when data is decoded, whether or not to decode in a corresponding direction can be selected according to the identification signal. 2. The decoding according to claim 1, wherein when the data is decoded, decoding of some of n stages is omitted and the decoding is simply performed. Multi-stage decoding method.