CA2041214A1 - Multi-level error correction system - Google Patents
Multi-level error correction systemInfo
- Publication number
- CA2041214A1 CA2041214A1 CA002041214A CA2041214A CA2041214A1 CA 2041214 A1 CA2041214 A1 CA 2041214A1 CA 002041214 A CA002041214 A CA 002041214A CA 2041214 A CA2041214 A CA 2041214A CA 2041214 A1 CA2041214 A1 CA 2041214A1
- Authority
- CA
- Canada
- Prior art keywords
- symbols
- redundancy
- section
- level
- code
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Abandoned
Links
- 238000012937 correction Methods 0.000 title claims description 63
- 238000000034 method Methods 0.000 claims description 26
- 230000000875 corresponding effect Effects 0.000 claims description 18
- 230000004224 protection Effects 0.000 claims description 11
- 230000000717 retained effect Effects 0.000 claims description 4
- 238000004193 electrokinetic chromatography Methods 0.000 description 58
- 238000010586 diagram Methods 0.000 description 6
- 230000003247 decreasing effect Effects 0.000 description 2
- 238000001514 detection method Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 241001362574 Decodes Species 0.000 description 1
- 241000760358 Enodes Species 0.000 description 1
- 238000007796 conventional method Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
Classifications
-
- H—ELECTRICITY
- H03—ELECTRONIC CIRCUITRY
- H03M—CODING; DECODING; CODE CONVERSION IN GENERAL
- H03M13/00—Coding, decoding or code conversion, for error detection or error correction; Coding theory basic assumptions; Coding bounds; Error probability evaluation methods; Channel models; Simulation or testing of codes
- H03M13/29—Coding, decoding or code conversion, for error detection or error correction; Coding theory basic assumptions; Coding bounds; Error probability evaluation methods; Channel models; Simulation or testing of codes combining two or more codes or code structures, e.g. product codes, generalised product codes, concatenated codes, inner and outer codes
Abstract
ABSTRACT
A system separates data into "i" interleaved sections of "k" symbols each and encodes each section using a relatively weak (n,k) ECC. It generates for each k symbols an n symbol code word containing "n-k" or "r" redundancy symbols, or a total of "ri" redundancy symbols. The system segments the ri redundancy symbols into "m" multi-symbol sections, namely, R(1), R(2), ... , R(m), which contain corresponding symbols from each of the i code words. The system then encodes each of these sections using various ECC's and generates Q(2), ..., Q(m) redundancy symbols, respectively. The system then records just the data, the section R(1) redundancy symbols and the Q(2), ...,Q(m) redundancy symbols. When the system later retrieves the data, it decodes it code word-by-code word using the level-1 ECC and fills in the R(2), ..., R(m) redundancy symbols in each code word. It also corrects, if possible, any errors in the data. If the level-1 decoding does not correct all the errors, the system uses the Q(2) redundancy symbols and the level-2 ECC to fill in missing section R(2) coda word symbols and tries to correct the corresponding code words using the level-1 ECC. If necessary, the system fills in redundancy symbols in the next section, i.e., section R(3), using the Q(3) symbols and the level-3 ECC. It then uses the level-1 ECC to correct more errors in the data, and so on.
A system separates data into "i" interleaved sections of "k" symbols each and encodes each section using a relatively weak (n,k) ECC. It generates for each k symbols an n symbol code word containing "n-k" or "r" redundancy symbols, or a total of "ri" redundancy symbols. The system segments the ri redundancy symbols into "m" multi-symbol sections, namely, R(1), R(2), ... , R(m), which contain corresponding symbols from each of the i code words. The system then encodes each of these sections using various ECC's and generates Q(2), ..., Q(m) redundancy symbols, respectively. The system then records just the data, the section R(1) redundancy symbols and the Q(2), ...,Q(m) redundancy symbols. When the system later retrieves the data, it decodes it code word-by-code word using the level-1 ECC and fills in the R(2), ..., R(m) redundancy symbols in each code word. It also corrects, if possible, any errors in the data. If the level-1 decoding does not correct all the errors, the system uses the Q(2) redundancy symbols and the level-2 ECC to fill in missing section R(2) coda word symbols and tries to correct the corresponding code words using the level-1 ECC. If necessary, the system fills in redundancy symbols in the next section, i.e., section R(3), using the Q(3) symbols and the level-3 ECC. It then uses the level-1 ECC to correct more errors in the data, and so on.
Description
2 ~
A M~L~ I,EVEL ~RROR CORRECTION ~;Y~TEN
FIELD OF THE~INVENTION
The invention relates generally to error correction of data and more particularly to improved techniques for multi-level encoding and decoding.
BACKGROUND
The importance of error correction coding of data in digital computer systems has increased greatly as the density of the data recorded on mass storage medial more particularly magnetic disks, has increased. With higher recording densities, a tiny imperfection in the recording surface of a disk can corrupt a large amount of data. In order to avoid losing that data, error correction codes ("ECC's") are employed to, as the name implies, correct the erroneous data.
Before a strin~ of data s~mbols is recorded on a disk, it is mathematically encoded to form ECC, or redundancy, sym-bols. The redundancy symbols are then appended to the data string to form code words - data symbols plus redundancy sym-bols - and the code words are then stored on the disk. When the stored data is to be accessed from the disk, the code words containing the data symbols are retrieved from the disk and mathematically decoded. During decoding any errors in the data are detected and, if possible, corrected through manipulation of the redundancy symbols [For a detailed description of decod ing see Peterson and Weldon, Error Correctinq Codes, 2d Edi-tion, MIT Press, 1972].
Stored digital code words can contain m~ltiple errors.
One of the most effective types of ECC used for the correction of multiple errors is a Reed-Solomon code [For a detailed des-cription of Reed-Solomon codes, see Peterson and Weldon, Error 2~:L2 ~ ~
Correctinq Codes]. Error detection and correction techniques for Reed-Solomon ECC's are well known. Id. One such technique begins with again encoding the code word data to generate a new set of redundancy symbols and then comparing this set of redundancy symbols with the redundancy symbo].s in the retrieved code word, i.e. the set of redundancy symbols generated by the pre-storage encoding of the data, to detect any errors in the retrieved code word. ~For a detailed discussion of this error detection technique, see United States Patent 4,413,339 issued to Riggle and Weng~
Errors in stored data often occur in pattsrns such as multi-symbol bursts. Such errors may be caused by, for exam-ple, an imperfection in the recording medium. Various enc:oding and decoding schemes have been developed to correct common er-ror patterns. One scheme developed for bursts is interleaving.
Interleaving involves separating the data into a number of sec tions and, usiny an ECC, separately encoding each section to form a code word. Each code word thus contains a section of the data and a related set of redundancy symbols. The inter-leaved data and the redundancy symbols from all the sections are then recorded. Typically, the interleaved data symbols are recorded contiguously, followsd by the redundancy s~nbols.
Burst errors, which affect a number of contiguous tinterleaved) data symbols, cause a small number of errors in each of several code words.
When the data and redundancy symbols are later retrieved, the various sets o~ redundancy symbols are used to correct errors in the data sections associated with them. Thus a given set of redundancy symbols protects a portion of the data, and presumably, corrects only a portion of any error burst. Accordingly, an ECC which is designed to correct "x"
erroneous symbols can correct bursts which are longer than x symbols by correcting them section-by-section. The ECC
selected for such a scheme can be less powerful than one which must correct burst errors without partioning them. The ad-2~12~
vantages to using a less powerful ECC are simpler encod~ing/decoding hardware and (typically) faster correction. [For a detailed description of interleaving encocling, see Peterson and Weldon, Brror Correctinq Codes~.
Another scheme used to correct common error patterns is multi-leval encoding. Multi-level encoding involves encoding data once using an ECC and/or a technique which is designed to correct the most common error patterns and then encoding the data a next time using another ECC and/or technique which is designed to correct the next most common error patterns, and so on. Multi-level decoding involves correcting the data using the first level ECC or technique and, if errors then remain, correcting them using the second level ECC or technique, and so on.
For axample, a first level of encoding may consist of encoding the data with a relatively weak ECC. A second level of encoding may consist of again encoding the data with a more powerful ECC. Such a two-level scheme is disclosed in United States Patents 4,706,250 and 4,525,838 to Patel. The Patel schems disclosed in the two patents first encodes the data using an ECC which corrects ea errors. Patel separates a block of data into "i" sections, or sub-blocks, of "k" symbols each and separately encodes each k-symbol section using a first ECC.
This level of encoding generates "p" redundancy symbols ~or each data section, or a total of "ip" redundancy symbols.
Patel next adds together (exclusive OR's) the cor--responding data symbols in each sub-block and treats the resulting k symbols as an additional "data" section. Patel then encodes these symbols using a more powerful ECC which can correct up to eb errors (eb 2 ea). This second encoding gener-ates "s" additional redundancy symbols. Patel stores the data and the ip and s redundancy symbols, that is, the redundancy symbols generated during the two levels of encoding. He does not store the k additional "data" symbols.
The Patel scheme can correct up to ea errors in each 2 ~
suh-block using the first level, or level-l, ECC and up to eb errors in any one sub-block using the level-2 ECC. Accordingly Patel attempts first to correct any errors in the data using the level-l ECC. If all the errors are corrected using this ECC, he stops the error correction decoding operation and does not use the level~2 ECC. If all but one sub-bloc~ is cor-rected, the level-2 code is used in an attempt to correct it.
Patel thus exclusive OR's the corresponding data symbols in each section which include the symbols corrected using the levPl-l ECC and the erroneous symbols which the level-l code did not correct, to form k "datal' symbols which correspond to the k-symbol additional data section generated during encoding.
Patel then attempts to correct the errors in these k symbols using the level-2 code.
If there are eb or fewer errors, the level-2 code determines the locations of the erroneous symbols within the k-symbol "data" section and generates the associated error values, that is, the symbols which must be exclusive ORId with the erroneous symbols to correct them. Patel translates the error locations to the sub-block which contains the errors and corrects them using the generated error value symbols.
If the data contains more than eb errors per sub-block, or more than one sub-block with more than ea errors, the Patel scheme can not correct the arrors. Thus, as more errors occur because of increased recording densities, it is desirable to employ an error correcting scheme which is capable of correct-ing a greater number of errorsO
One solution is to use a more powerful ECC at each level. This creates three problems. First, the more powerful ECCs generate additional redundancy symbols. $hus more storage space must be allocated to the data, and less data can be re-corded in a given storage space. Second, the more powerful ECCs require more complex, and thus more expensive, encoding and decoding hardware and/or software. Third, the ECCs require more time to correct the errors than the ECCs which use fewer 2~21~
redundancy symbols.
As data transfer technologies improve, and the speed with which data may be retrisved from a disk increases, a slower, more complex ECC acts as a limit on the speed with which data may be transferred. Accordingly, wh~t is needed is a more powerful encoding scheme which can quickly correct er-rors, is easily implemented and does not require excess amounts of storage space.
SUMMARY
The invention is a multi-level encoding scheme which uses in higher levels of encoding the r~dundancy symbols gener-ated during a first level of encoding such that all the levels together form a powerful and quick error correction scheme.
The multi-level encoding scheme need not retain all o~ the redundancy symbols because these symbols can be re~generated from a selected number of retained symbols, as explained below.
A system using the inventive multi-level encoding scheme first separates the dat~ intv "i" sections of "k" symbols each, or i sub-blocks, and encodes each sub-block using an (n,k) ECC. The data symbols may or may not be interleaved~ This first level of encoding generates for each X data symbols an "n" symbol code word containing l'n-k", or "r" redundancy sym-bols, or a total of "ri" redundancy symbols.
The system next segments the ri redundancy symbols into "m" multi-symbol sections, namely, R(1), R(2), ... , R(m), which contain corresponding symbols from each of the i code words. The system then encodes each of these sections using either the same ECC, or, preferably, ECC's of decreasing strength. Thus it encodes, for example, a first redundancy symbol, wj+l, from each of the i level-l code words and genera-tes q2 redundancy symbols, the wi+l redundancy symbols and q2 generated redundancy forming a level-2 code word. Next it en-codes a second redundancy symbol, wj+2, from each of the i code 2 ~
words and generates another q2 redundancy symbols, forming a second level-2 code word, and so forth. The level-2 redundancy svmbols corresponding to the various sections R(2)...R(m), are in sactions Q(2), ..., Q(m), respectively. The system then xecords the data, the section R(1) redundancy symbols and the section Q(2), ...,Q(m) redundancy symbols. It does not record the redundancy symbols in sections R(2), ... , R(m), however, because these symbols can be re-generated usîng the recorded R(l) and Q(2), ..., Q(m) redundancy symbols, as set forth be-low.
When data is retrieved from storage, it is decoded code word-by code word using the level-l ECC. The system usas the retrieved section R(1) symbols as redundancy symbols and treats the "missing" section R~2)...R~m) symbols as erasure errors.
The level-l ECC is thus used to fill in the redundancy symbols in sections R(2), ..., R(m) for each level-1 code word and also to correct, if possible, any errors in the retrieved data and section R(l) redundancy symbols. If the level-l decoding does not correct all the errors and erasures using only the retrieved section R(1) symbols as redundancy symbols, the sys-tem uses the section Q(2) redundancy symbols and the level-2 ECC to fill in missing section R(2) redundancy symbols. Next, the system tries to correct the corresponding level-l code words using the level-1 ECC and the retrieved and filled-in redundancy symbols in sections R(1) and R(2). If necessary, the system fills in any uncorrected redundancy symbols in the next section, i.e., section R(3), using the section Q(3) sym- -bols and the level-3 ECC. It then uses the lPvel-1 ECC and the redundancy symbols in sections R(1), R(2) and R(3) to correct more errors in the data, and so on.
BRIEF DESCRIPTION OF THE DRAWINGS
The above and further advantages of the invention may be better understood by referring to the following description in 2 ~ ~
conjunction with the accompanying drawings, in which:
Figure 1 is a block diagram of interleaved code words generated in accordance with a first le~el of encoding;
Figure 2 is a block diagram of the code words of Figure 1 further encoded using a second and third level of encoding;
Figure 3 is a block diagram of portions of the code words of Figure 2 which are retained in memory;
Figure 4 is a block diagram of retrieved code words in-cluding re-generated redundancy symbols;
Figure 5 is a block diagram of an encoding system which encodes data to form code words in accordance with a secolld em-bodiment of the current invention; and Figure 6 is a block diagram of the code words generated by the encoder shown in Fiyure 5.
DETAILED DESCRIPTION
With reference to Eigure 1, a system (not shown) sepa-rates a string of "ki" data symbols into "i" sections, or sub-blocks, of "k'l symbols each. The system may first interleave the ki data symbols. These i sections are arrayed as columns 10 in the drawing. The system performs a first level encoding using an (n,k) ECC and conventional techniques, and encodes each section of data to form a level-1 code word consisting of k data symbols and "n-k", or "r", redundancy symbols. The code words correspond to coiumns c1, c2~ ~ ci of the array in the drawing. The system thus generates "ri" redundancy symbols in the first level of encoding.
The system next segments the ri redundancy symbols into m sections 12-16, R(1), R(2), ..., R(m), of arbitrary size, wh~re m=3 in the exemplary embodiment. Each section, preferab-ly, contains at least two redundancy symbols, which is the num-ber of symbols required to correct one erroneous data symbol.
Thus each section 12-16 contains at least two rows in the array depicted in the drawing.
2~
Referring now to Figure 2, the system encodes the sym-bols in section Rt2) using a second level of encoding. It thus encodes as "informationl' symbols the corresponding redundancy symbols from each of the i code words, that is, each row, IIW
of section R(2), using a second ECC. This second level of en-coding generates for each row wj+1, ..., wp of section R(2), "q2" redundancy s~mbols~ Each row consisting of i information symbols and q2 redundancy symbols forms a level-2 code word.
The system thus generates in this level of encoding a total o~
(p-j~l)q2 redundancy symbols. These symbols are depicted as section Q(2) in the drawing. Generally, number of symbols in section Q(2~ c number o~ symbols in section R(2)~
The system next performs a third level of encoding by similarly encoding the symbols in section R(3). It thus en-codes the corresponding redundancy symbols from each of the i code words using an ECC which, preferably, is less powerful than the level-~ code. The third level of encodinq generates for each row wp+1, wp+2...wr of section R(3), "q3" redundancy symbols. Each row consisting of i "information" symbols and q3 redundancy symbols forms a level-3 code word. The system thus generates in this level of encodin~ a total of (r-p~l)q3 redundancy symbols. These symbols are depicted as seckion Q(3) in the drawing. Generally, the number of symbols in section Q(3) ~ the number of symbols in section R(3).
Referring now to Figure 3, the system stores in memory (not shown) the ki data symbols and the section R~1~, Q(2) and Q(3) redundancy symbols. It does not store the section R(2) and R(3) redundancy symbols, however, because these symbols can be re-generated using the stored section R(1), Q(2) and Q(3) redundancy symbols, as explained below. The section R(2) and R(3) symbols are thus intentionally "erased." Each of the erased symbols is treated during decoding as an erasure, that is, an error with a known location.
Optionally, the system may also encode the stored sec-tions R(l), Q(2)...Q(m) redundancy symbols using one or more 2~:L21~
ECCs, which generates redundancy symbols ECK~l), ECK(2) and ECK(3). Thess symbols are then stored along with the data and the section R(1), Q(2) and Q(3~ symbols. During decoding, these (optional) redundancy symbols may be used to correct any errors in the stored redundancy symbols before the redundancy symbols are used to re-generate the erased symbols. The op-tional ~ymbols ara thus an added layer of protection for both the stored symbols and the erased symbols.
The data symbols are protected by m levels of encoding.
The storage space required for the m levels of redundancy sym-bols is less than the space which would be reguired aIl the ri level-l redundancy symbols. Accordingly, the scheme offers more protection than prior multi-level encoding schemes and uses less storage space.
As is understood by those skilled in the art, the data and redundancy symbols may be stored, or recorded, in any format. The block format shown in the figures is for explana-tory purposes only.
When the stored information is later retrieved, the sys-tem uses the retrieved section R(1) redundancy s~mbols to re-generate for each code word the 'lerased" section R~2) and R(3) level-l redundancy symbols using conventional error and erasure correcting techniques. It also uses the section R(l) redundancy symbols to correct err~rs in the retrieved data and redundancy symbols. Figure 4 illustrates several re-generated columns, e.g., c1, c2, etc.
If the system does not generate a valid code word for a particular section of data, i.e., a particular column, it treats the associated section R(2) and R(3) symbols as errors with known locations, e.g., columns ce1, Ce2...Cef~ It then decodes each row of section R(2) usiny the appropriate section Q(2) redundancy symbols and the level-2 ECC and fills in the erased section R~2) symbols in the columns ce1, ce2 ... cew, as appropriate. The system next tries to correct any remaining errors in the retrieved data and section R(1) symbols of these 2 ~
lu--columns using the level-l ECC and the appropriate section R(1) and R(2) redundancy symbols.
If the system can not correct all the errors and re-generate all the erased redundancy symbols using the level-l ECC and the sections R(1) and R(2) redundancy symbols, it uses the section Q(3) redundancy symbols and the level-3 ECC to fill in any missing section R(3) symbols. It then uses these filled in symbols and the section R(1) and R(2) redundancy symbols with the level-1 ECC to correct more errors. If necessary, and if there are more levels, the system repeats the correction procedure for each such level.
Before the system re-generates the erased section R(2) and R(3) symbols, it uses the optional ECC symbols, if any, to correct any errors in the retrieved section R(1), Q(2) and Q(3~
redundancy symbols. Next the system performs the level-1 error and erasure correction to fill in the erased redundancy symbols and correct errors the retrieved data in section R(l) redundancy symbols. It them follows the procedures set forth above to correct the errors.
If the system performs its higher levels of decoding, it is encountering badly corrupted data. The system may thus sig-nal to a system operator that the data is badly corrupted, and the system operator may then examine the recording medium and salvage the stored data before it becomes irretrievably lost.
Prior art systems which use only two levels of decoding are no~
so warned of corrupted data, and thus, they may eventually lose data to corrupted media.
Figures 5 and 6 illustrate a second embodiment of the current invention which can be used in conjunction with an ex-isting one-level encoding system. The one-level system encodes the ki data symbols after (optionally) interleaving them. It encodes them in k-symbol sections in encoder 18 to form code words in accordance with an ECC which has a consecutive root generator polynomial, gl(x):
(X+~Xa) * (X+~Xa+l) * (X+~a+2) * (X+lXa+3) * * (X+C~a+9) 2~ 12~ ~
where * represents Galois Fiel.d multiplication and ~ is a primitive polynomial of the Galois Field GF(2t) used to enode the data.
The code words consist of k data symbols and s redundancy symbols and are arrayed as columns 10 in Figure 6, with the si redundancy symbols for all the code words depicted as section S(l) in the drawings. The code words are next ap-plied to a second encoder 20 for a second level of encoding.
This level of ~ncoding is associated to with an ECC which has as a generator polynomial, g2(x), an expanded version of g1(x):
(x a! ) (x~lx ) * . . ~, * (~C~aa) * . . . * (x~ya+EI) * (x+a~a+~;+l) * * (X ~- a~8 I d I--gl(X) This seaond ECC is more powerful, that is, it corrects more er-rors in a given amount of data than the ECC which is associated with gl(x).
The data symbols have already been encoded in accordance with g1(x) by the one-level encoder 18 to produce the section S(l) redundancy symbols~ Accordingly, the code word data sym-bols and S(l) redundancy symbols are treated as information symbols and the second encoder 20 actually encodes the symbols in accordance with the factors of g2(x) which are not part of g1(x). Thus encoder 20 encodes the s~mbols in accordance with a polynomial g*2(x) which is equal to g2~x)/gl(x). This second encoding generates, for each code word, redundancy symbols depicted as section S(2) in Figure 6. The section S(2) redundancy symbols alone do not provide adequate error pro-tection to the data because the polynomial, g2*(x~, used to generate them may not have as many consecutive roots in GF(2t) as gl(x). However, the section S(2) symbols together with the corresponding section S(1) symbols protect the data symbols with the relatively powerful ECC having as a generator polyno-mial g2(x).
Third, fourth, ..., m levels of encoding may be per-formed in the same manner. A third level of encoding, for ex-2~12~
ample, uses a generator polynomial g3*(x) = ~3(X)/g2tX)~
where g3(x) is an expanded version of g2~x). Thus encoders for these levels, for example, encoder 22 for m-level encoding, en-code the data symbols and the corresponding section S(l)...S(m-1) redundancy symbols in accordance with a polynomial g*m(x) which is equal to the guotient of gmtx), which is the expanded version of gm_1(x), divided by the generator polynomial associ-ated with the previous level, ~m-l (X) . The level m genera-tor polynomial thus includes consecutive roots, also, and pro-tects the data with an even more powerful ECCo The redundancy symbols generated for each level of en-coding are depicted in sections S(l)...S(m) in Eigure 6. These redundancy symbol sections correspond to the section R(1), R(2)...R(m) of Figure 2. Referring again to Figure 6, the redundancy ~ymbols in æections S(2)...S(m), where m=3 in this embodiment, are then encoded, as lnformation symbols, to gener-ate the redundancy symbol depicted in sections Q(2) and Q(3), as discussed above with reference Figure 3.
After the data and the levels-2 and 3 redundancy symbols have been encoded to generate the symbols depicted in sections Q(2) and Q(3), the system further encodes the redundancy sym-bols in sections Q(2) and Q(3), using one or moxe ECCs, and generates redundancy symbols ECK(2) and ECK(3). The system then stores these redundancy symbols along with the data sym-bols, the section S(1) redundancy symbols and the section Q(2) and Q(3) redundancy symbols. It does not store the section S(2) and S(3) redundancy s~mhols, however, because these sym-bols can be regenerated using the stored redundancy symbols.
This second embodiment is different than the first, in that the system is required to encode the section Q(2)...Q(m) redundancy symbols. The system may (optionally) encode the section S(l) symbols to form section ECK(1) symbols, as in the first embodi-ment.
The data symbols lO are protected by all m levels of en-coding. The mth level protects them using gm(x) which is much 2 ~ ~
more powerful ECC than the one used in level-1, without requir~
ing the storage of the larga number of redundancy symbols.
Prior systems requira for such protection the storage o~ all the redundancy symbols from all levels of encoding.
Deccding is performed by the system in levels, as dis-cussed above with respect to the first embodiment. The system starts with the level-l ECC and uses higher levels, as neces-sary, to regensrate the erased symbols and correct additional errors. Each level provides a much more powerful error correc-tion mechanism, due to the use of a generator polynomial which has more terms than the one used in the previous level. Also, each level uses redundancy symbols filled by the codes a~socia-ted with appropriate section Q(2)...Q(m) and thus uses error free redundancy symbols. Accordingly correction can be per-formed relatively quickly.
The inventive multi level encoding scheme protects data symbols with powerful ECCs without, as expected with prior art systems, increasing the complexity of encoders/decoders, re-quiring additional storage space, or drastically increasing the time required for error correction. The scheme encodes redundancy symbols from the various lower encoding levels and treats them as erasures. The encoding scheme can thus give the data symbols powerful error protection without the storage of the number of redundancy symbols required with prior systems.
The invantive system uses the various levels of redundancy symbols during the multi-level decoding to correct both errors and erasures. Each decoding level corrects only a portion of the erasures and/or errors in the retrieved symbols.
Accordingly, error correction can be performed relatively guickly using, in higher levels of decoding, either ECCs of decreasing complexity, as in the first embodiment, or powerful ECC's which are required to correct relatively few errors, as in the second embodiment.
The foregoing description has been limited to two specific embodiments of this invention. It will be apparent, 2~:12 ~ ~
however, that variations and modifications may be made to the invention, with the attainment of some or all of its ad-vantages. Therefore, it is the object of the appended claims to cover all such variations and modifications as come within the true spirit and scope of the invention.
A M~L~ I,EVEL ~RROR CORRECTION ~;Y~TEN
FIELD OF THE~INVENTION
The invention relates generally to error correction of data and more particularly to improved techniques for multi-level encoding and decoding.
BACKGROUND
The importance of error correction coding of data in digital computer systems has increased greatly as the density of the data recorded on mass storage medial more particularly magnetic disks, has increased. With higher recording densities, a tiny imperfection in the recording surface of a disk can corrupt a large amount of data. In order to avoid losing that data, error correction codes ("ECC's") are employed to, as the name implies, correct the erroneous data.
Before a strin~ of data s~mbols is recorded on a disk, it is mathematically encoded to form ECC, or redundancy, sym-bols. The redundancy symbols are then appended to the data string to form code words - data symbols plus redundancy sym-bols - and the code words are then stored on the disk. When the stored data is to be accessed from the disk, the code words containing the data symbols are retrieved from the disk and mathematically decoded. During decoding any errors in the data are detected and, if possible, corrected through manipulation of the redundancy symbols [For a detailed description of decod ing see Peterson and Weldon, Error Correctinq Codes, 2d Edi-tion, MIT Press, 1972].
Stored digital code words can contain m~ltiple errors.
One of the most effective types of ECC used for the correction of multiple errors is a Reed-Solomon code [For a detailed des-cription of Reed-Solomon codes, see Peterson and Weldon, Error 2~:L2 ~ ~
Correctinq Codes]. Error detection and correction techniques for Reed-Solomon ECC's are well known. Id. One such technique begins with again encoding the code word data to generate a new set of redundancy symbols and then comparing this set of redundancy symbols with the redundancy symbo].s in the retrieved code word, i.e. the set of redundancy symbols generated by the pre-storage encoding of the data, to detect any errors in the retrieved code word. ~For a detailed discussion of this error detection technique, see United States Patent 4,413,339 issued to Riggle and Weng~
Errors in stored data often occur in pattsrns such as multi-symbol bursts. Such errors may be caused by, for exam-ple, an imperfection in the recording medium. Various enc:oding and decoding schemes have been developed to correct common er-ror patterns. One scheme developed for bursts is interleaving.
Interleaving involves separating the data into a number of sec tions and, usiny an ECC, separately encoding each section to form a code word. Each code word thus contains a section of the data and a related set of redundancy symbols. The inter-leaved data and the redundancy symbols from all the sections are then recorded. Typically, the interleaved data symbols are recorded contiguously, followsd by the redundancy s~nbols.
Burst errors, which affect a number of contiguous tinterleaved) data symbols, cause a small number of errors in each of several code words.
When the data and redundancy symbols are later retrieved, the various sets o~ redundancy symbols are used to correct errors in the data sections associated with them. Thus a given set of redundancy symbols protects a portion of the data, and presumably, corrects only a portion of any error burst. Accordingly, an ECC which is designed to correct "x"
erroneous symbols can correct bursts which are longer than x symbols by correcting them section-by-section. The ECC
selected for such a scheme can be less powerful than one which must correct burst errors without partioning them. The ad-2~12~
vantages to using a less powerful ECC are simpler encod~ing/decoding hardware and (typically) faster correction. [For a detailed description of interleaving encocling, see Peterson and Weldon, Brror Correctinq Codes~.
Another scheme used to correct common error patterns is multi-leval encoding. Multi-level encoding involves encoding data once using an ECC and/or a technique which is designed to correct the most common error patterns and then encoding the data a next time using another ECC and/or technique which is designed to correct the next most common error patterns, and so on. Multi-level decoding involves correcting the data using the first level ECC or technique and, if errors then remain, correcting them using the second level ECC or technique, and so on.
For axample, a first level of encoding may consist of encoding the data with a relatively weak ECC. A second level of encoding may consist of again encoding the data with a more powerful ECC. Such a two-level scheme is disclosed in United States Patents 4,706,250 and 4,525,838 to Patel. The Patel schems disclosed in the two patents first encodes the data using an ECC which corrects ea errors. Patel separates a block of data into "i" sections, or sub-blocks, of "k" symbols each and separately encodes each k-symbol section using a first ECC.
This level of encoding generates "p" redundancy symbols ~or each data section, or a total of "ip" redundancy symbols.
Patel next adds together (exclusive OR's) the cor--responding data symbols in each sub-block and treats the resulting k symbols as an additional "data" section. Patel then encodes these symbols using a more powerful ECC which can correct up to eb errors (eb 2 ea). This second encoding gener-ates "s" additional redundancy symbols. Patel stores the data and the ip and s redundancy symbols, that is, the redundancy symbols generated during the two levels of encoding. He does not store the k additional "data" symbols.
The Patel scheme can correct up to ea errors in each 2 ~
suh-block using the first level, or level-l, ECC and up to eb errors in any one sub-block using the level-2 ECC. Accordingly Patel attempts first to correct any errors in the data using the level-l ECC. If all the errors are corrected using this ECC, he stops the error correction decoding operation and does not use the level~2 ECC. If all but one sub-bloc~ is cor-rected, the level-2 code is used in an attempt to correct it.
Patel thus exclusive OR's the corresponding data symbols in each section which include the symbols corrected using the levPl-l ECC and the erroneous symbols which the level-l code did not correct, to form k "datal' symbols which correspond to the k-symbol additional data section generated during encoding.
Patel then attempts to correct the errors in these k symbols using the level-2 code.
If there are eb or fewer errors, the level-2 code determines the locations of the erroneous symbols within the k-symbol "data" section and generates the associated error values, that is, the symbols which must be exclusive ORId with the erroneous symbols to correct them. Patel translates the error locations to the sub-block which contains the errors and corrects them using the generated error value symbols.
If the data contains more than eb errors per sub-block, or more than one sub-block with more than ea errors, the Patel scheme can not correct the arrors. Thus, as more errors occur because of increased recording densities, it is desirable to employ an error correcting scheme which is capable of correct-ing a greater number of errorsO
One solution is to use a more powerful ECC at each level. This creates three problems. First, the more powerful ECCs generate additional redundancy symbols. $hus more storage space must be allocated to the data, and less data can be re-corded in a given storage space. Second, the more powerful ECCs require more complex, and thus more expensive, encoding and decoding hardware and/or software. Third, the ECCs require more time to correct the errors than the ECCs which use fewer 2~21~
redundancy symbols.
As data transfer technologies improve, and the speed with which data may be retrisved from a disk increases, a slower, more complex ECC acts as a limit on the speed with which data may be transferred. Accordingly, wh~t is needed is a more powerful encoding scheme which can quickly correct er-rors, is easily implemented and does not require excess amounts of storage space.
SUMMARY
The invention is a multi-level encoding scheme which uses in higher levels of encoding the r~dundancy symbols gener-ated during a first level of encoding such that all the levels together form a powerful and quick error correction scheme.
The multi-level encoding scheme need not retain all o~ the redundancy symbols because these symbols can be re~generated from a selected number of retained symbols, as explained below.
A system using the inventive multi-level encoding scheme first separates the dat~ intv "i" sections of "k" symbols each, or i sub-blocks, and encodes each sub-block using an (n,k) ECC. The data symbols may or may not be interleaved~ This first level of encoding generates for each X data symbols an "n" symbol code word containing l'n-k", or "r" redundancy sym-bols, or a total of "ri" redundancy symbols.
The system next segments the ri redundancy symbols into "m" multi-symbol sections, namely, R(1), R(2), ... , R(m), which contain corresponding symbols from each of the i code words. The system then encodes each of these sections using either the same ECC, or, preferably, ECC's of decreasing strength. Thus it encodes, for example, a first redundancy symbol, wj+l, from each of the i level-l code words and genera-tes q2 redundancy symbols, the wi+l redundancy symbols and q2 generated redundancy forming a level-2 code word. Next it en-codes a second redundancy symbol, wj+2, from each of the i code 2 ~
words and generates another q2 redundancy symbols, forming a second level-2 code word, and so forth. The level-2 redundancy svmbols corresponding to the various sections R(2)...R(m), are in sactions Q(2), ..., Q(m), respectively. The system then xecords the data, the section R(1) redundancy symbols and the section Q(2), ...,Q(m) redundancy symbols. It does not record the redundancy symbols in sections R(2), ... , R(m), however, because these symbols can be re-generated usîng the recorded R(l) and Q(2), ..., Q(m) redundancy symbols, as set forth be-low.
When data is retrieved from storage, it is decoded code word-by code word using the level-l ECC. The system usas the retrieved section R(1) symbols as redundancy symbols and treats the "missing" section R~2)...R~m) symbols as erasure errors.
The level-l ECC is thus used to fill in the redundancy symbols in sections R(2), ..., R(m) for each level-1 code word and also to correct, if possible, any errors in the retrieved data and section R(l) redundancy symbols. If the level-l decoding does not correct all the errors and erasures using only the retrieved section R(1) symbols as redundancy symbols, the sys-tem uses the section Q(2) redundancy symbols and the level-2 ECC to fill in missing section R(2) redundancy symbols. Next, the system tries to correct the corresponding level-l code words using the level-1 ECC and the retrieved and filled-in redundancy symbols in sections R(1) and R(2). If necessary, the system fills in any uncorrected redundancy symbols in the next section, i.e., section R(3), using the section Q(3) sym- -bols and the level-3 ECC. It then uses the lPvel-1 ECC and the redundancy symbols in sections R(1), R(2) and R(3) to correct more errors in the data, and so on.
BRIEF DESCRIPTION OF THE DRAWINGS
The above and further advantages of the invention may be better understood by referring to the following description in 2 ~ ~
conjunction with the accompanying drawings, in which:
Figure 1 is a block diagram of interleaved code words generated in accordance with a first le~el of encoding;
Figure 2 is a block diagram of the code words of Figure 1 further encoded using a second and third level of encoding;
Figure 3 is a block diagram of portions of the code words of Figure 2 which are retained in memory;
Figure 4 is a block diagram of retrieved code words in-cluding re-generated redundancy symbols;
Figure 5 is a block diagram of an encoding system which encodes data to form code words in accordance with a secolld em-bodiment of the current invention; and Figure 6 is a block diagram of the code words generated by the encoder shown in Fiyure 5.
DETAILED DESCRIPTION
With reference to Eigure 1, a system (not shown) sepa-rates a string of "ki" data symbols into "i" sections, or sub-blocks, of "k'l symbols each. The system may first interleave the ki data symbols. These i sections are arrayed as columns 10 in the drawing. The system performs a first level encoding using an (n,k) ECC and conventional techniques, and encodes each section of data to form a level-1 code word consisting of k data symbols and "n-k", or "r", redundancy symbols. The code words correspond to coiumns c1, c2~ ~ ci of the array in the drawing. The system thus generates "ri" redundancy symbols in the first level of encoding.
The system next segments the ri redundancy symbols into m sections 12-16, R(1), R(2), ..., R(m), of arbitrary size, wh~re m=3 in the exemplary embodiment. Each section, preferab-ly, contains at least two redundancy symbols, which is the num-ber of symbols required to correct one erroneous data symbol.
Thus each section 12-16 contains at least two rows in the array depicted in the drawing.
2~
Referring now to Figure 2, the system encodes the sym-bols in section Rt2) using a second level of encoding. It thus encodes as "informationl' symbols the corresponding redundancy symbols from each of the i code words, that is, each row, IIW
of section R(2), using a second ECC. This second level of en-coding generates for each row wj+1, ..., wp of section R(2), "q2" redundancy s~mbols~ Each row consisting of i information symbols and q2 redundancy symbols forms a level-2 code word.
The system thus generates in this level of encoding a total o~
(p-j~l)q2 redundancy symbols. These symbols are depicted as section Q(2) in the drawing. Generally, number of symbols in section Q(2~ c number o~ symbols in section R(2)~
The system next performs a third level of encoding by similarly encoding the symbols in section R(3). It thus en-codes the corresponding redundancy symbols from each of the i code words using an ECC which, preferably, is less powerful than the level-~ code. The third level of encodinq generates for each row wp+1, wp+2...wr of section R(3), "q3" redundancy symbols. Each row consisting of i "information" symbols and q3 redundancy symbols forms a level-3 code word. The system thus generates in this level of encodin~ a total of (r-p~l)q3 redundancy symbols. These symbols are depicted as seckion Q(3) in the drawing. Generally, the number of symbols in section Q(3) ~ the number of symbols in section R(3).
Referring now to Figure 3, the system stores in memory (not shown) the ki data symbols and the section R~1~, Q(2) and Q(3) redundancy symbols. It does not store the section R(2) and R(3) redundancy symbols, however, because these symbols can be re-generated using the stored section R(1), Q(2) and Q(3) redundancy symbols, as explained below. The section R(2) and R(3) symbols are thus intentionally "erased." Each of the erased symbols is treated during decoding as an erasure, that is, an error with a known location.
Optionally, the system may also encode the stored sec-tions R(l), Q(2)...Q(m) redundancy symbols using one or more 2~:L21~
ECCs, which generates redundancy symbols ECK~l), ECK(2) and ECK(3). Thess symbols are then stored along with the data and the section R(1), Q(2) and Q(3~ symbols. During decoding, these (optional) redundancy symbols may be used to correct any errors in the stored redundancy symbols before the redundancy symbols are used to re-generate the erased symbols. The op-tional ~ymbols ara thus an added layer of protection for both the stored symbols and the erased symbols.
The data symbols are protected by m levels of encoding.
The storage space required for the m levels of redundancy sym-bols is less than the space which would be reguired aIl the ri level-l redundancy symbols. Accordingly, the scheme offers more protection than prior multi-level encoding schemes and uses less storage space.
As is understood by those skilled in the art, the data and redundancy symbols may be stored, or recorded, in any format. The block format shown in the figures is for explana-tory purposes only.
When the stored information is later retrieved, the sys-tem uses the retrieved section R(1) redundancy s~mbols to re-generate for each code word the 'lerased" section R~2) and R(3) level-l redundancy symbols using conventional error and erasure correcting techniques. It also uses the section R(l) redundancy symbols to correct err~rs in the retrieved data and redundancy symbols. Figure 4 illustrates several re-generated columns, e.g., c1, c2, etc.
If the system does not generate a valid code word for a particular section of data, i.e., a particular column, it treats the associated section R(2) and R(3) symbols as errors with known locations, e.g., columns ce1, Ce2...Cef~ It then decodes each row of section R(2) usiny the appropriate section Q(2) redundancy symbols and the level-2 ECC and fills in the erased section R~2) symbols in the columns ce1, ce2 ... cew, as appropriate. The system next tries to correct any remaining errors in the retrieved data and section R(1) symbols of these 2 ~
lu--columns using the level-l ECC and the appropriate section R(1) and R(2) redundancy symbols.
If the system can not correct all the errors and re-generate all the erased redundancy symbols using the level-l ECC and the sections R(1) and R(2) redundancy symbols, it uses the section Q(3) redundancy symbols and the level-3 ECC to fill in any missing section R(3) symbols. It then uses these filled in symbols and the section R(1) and R(2) redundancy symbols with the level-1 ECC to correct more errors. If necessary, and if there are more levels, the system repeats the correction procedure for each such level.
Before the system re-generates the erased section R(2) and R(3) symbols, it uses the optional ECC symbols, if any, to correct any errors in the retrieved section R(1), Q(2) and Q(3~
redundancy symbols. Next the system performs the level-1 error and erasure correction to fill in the erased redundancy symbols and correct errors the retrieved data in section R(l) redundancy symbols. It them follows the procedures set forth above to correct the errors.
If the system performs its higher levels of decoding, it is encountering badly corrupted data. The system may thus sig-nal to a system operator that the data is badly corrupted, and the system operator may then examine the recording medium and salvage the stored data before it becomes irretrievably lost.
Prior art systems which use only two levels of decoding are no~
so warned of corrupted data, and thus, they may eventually lose data to corrupted media.
Figures 5 and 6 illustrate a second embodiment of the current invention which can be used in conjunction with an ex-isting one-level encoding system. The one-level system encodes the ki data symbols after (optionally) interleaving them. It encodes them in k-symbol sections in encoder 18 to form code words in accordance with an ECC which has a consecutive root generator polynomial, gl(x):
(X+~Xa) * (X+~Xa+l) * (X+~a+2) * (X+lXa+3) * * (X+C~a+9) 2~ 12~ ~
where * represents Galois Fiel.d multiplication and ~ is a primitive polynomial of the Galois Field GF(2t) used to enode the data.
The code words consist of k data symbols and s redundancy symbols and are arrayed as columns 10 in Figure 6, with the si redundancy symbols for all the code words depicted as section S(l) in the drawings. The code words are next ap-plied to a second encoder 20 for a second level of encoding.
This level of ~ncoding is associated to with an ECC which has as a generator polynomial, g2(x), an expanded version of g1(x):
(x a! ) (x~lx ) * . . ~, * (~C~aa) * . . . * (x~ya+EI) * (x+a~a+~;+l) * * (X ~- a~8 I d I--gl(X) This seaond ECC is more powerful, that is, it corrects more er-rors in a given amount of data than the ECC which is associated with gl(x).
The data symbols have already been encoded in accordance with g1(x) by the one-level encoder 18 to produce the section S(l) redundancy symbols~ Accordingly, the code word data sym-bols and S(l) redundancy symbols are treated as information symbols and the second encoder 20 actually encodes the symbols in accordance with the factors of g2(x) which are not part of g1(x). Thus encoder 20 encodes the s~mbols in accordance with a polynomial g*2(x) which is equal to g2~x)/gl(x). This second encoding generates, for each code word, redundancy symbols depicted as section S(2) in Figure 6. The section S(2) redundancy symbols alone do not provide adequate error pro-tection to the data because the polynomial, g2*(x~, used to generate them may not have as many consecutive roots in GF(2t) as gl(x). However, the section S(2) symbols together with the corresponding section S(1) symbols protect the data symbols with the relatively powerful ECC having as a generator polyno-mial g2(x).
Third, fourth, ..., m levels of encoding may be per-formed in the same manner. A third level of encoding, for ex-2~12~
ample, uses a generator polynomial g3*(x) = ~3(X)/g2tX)~
where g3(x) is an expanded version of g2~x). Thus encoders for these levels, for example, encoder 22 for m-level encoding, en-code the data symbols and the corresponding section S(l)...S(m-1) redundancy symbols in accordance with a polynomial g*m(x) which is equal to the guotient of gmtx), which is the expanded version of gm_1(x), divided by the generator polynomial associ-ated with the previous level, ~m-l (X) . The level m genera-tor polynomial thus includes consecutive roots, also, and pro-tects the data with an even more powerful ECCo The redundancy symbols generated for each level of en-coding are depicted in sections S(l)...S(m) in Eigure 6. These redundancy symbol sections correspond to the section R(1), R(2)...R(m) of Figure 2. Referring again to Figure 6, the redundancy ~ymbols in æections S(2)...S(m), where m=3 in this embodiment, are then encoded, as lnformation symbols, to gener-ate the redundancy symbol depicted in sections Q(2) and Q(3), as discussed above with reference Figure 3.
After the data and the levels-2 and 3 redundancy symbols have been encoded to generate the symbols depicted in sections Q(2) and Q(3), the system further encodes the redundancy sym-bols in sections Q(2) and Q(3), using one or moxe ECCs, and generates redundancy symbols ECK(2) and ECK(3). The system then stores these redundancy symbols along with the data sym-bols, the section S(1) redundancy symbols and the section Q(2) and Q(3) redundancy symbols. It does not store the section S(2) and S(3) redundancy s~mhols, however, because these sym-bols can be regenerated using the stored redundancy symbols.
This second embodiment is different than the first, in that the system is required to encode the section Q(2)...Q(m) redundancy symbols. The system may (optionally) encode the section S(l) symbols to form section ECK(1) symbols, as in the first embodi-ment.
The data symbols lO are protected by all m levels of en-coding. The mth level protects them using gm(x) which is much 2 ~ ~
more powerful ECC than the one used in level-1, without requir~
ing the storage of the larga number of redundancy symbols.
Prior systems requira for such protection the storage o~ all the redundancy symbols from all levels of encoding.
Deccding is performed by the system in levels, as dis-cussed above with respect to the first embodiment. The system starts with the level-l ECC and uses higher levels, as neces-sary, to regensrate the erased symbols and correct additional errors. Each level provides a much more powerful error correc-tion mechanism, due to the use of a generator polynomial which has more terms than the one used in the previous level. Also, each level uses redundancy symbols filled by the codes a~socia-ted with appropriate section Q(2)...Q(m) and thus uses error free redundancy symbols. Accordingly correction can be per-formed relatively quickly.
The inventive multi level encoding scheme protects data symbols with powerful ECCs without, as expected with prior art systems, increasing the complexity of encoders/decoders, re-quiring additional storage space, or drastically increasing the time required for error correction. The scheme encodes redundancy symbols from the various lower encoding levels and treats them as erasures. The encoding scheme can thus give the data symbols powerful error protection without the storage of the number of redundancy symbols required with prior systems.
The invantive system uses the various levels of redundancy symbols during the multi-level decoding to correct both errors and erasures. Each decoding level corrects only a portion of the erasures and/or errors in the retrieved symbols.
Accordingly, error correction can be performed relatively guickly using, in higher levels of decoding, either ECCs of decreasing complexity, as in the first embodiment, or powerful ECC's which are required to correct relatively few errors, as in the second embodiment.
The foregoing description has been limited to two specific embodiments of this invention. It will be apparent, 2~:12 ~ ~
however, that variations and modifications may be made to the invention, with the attainment of some or all of its ad-vantages. Therefore, it is the object of the appended claims to cover all such variations and modifications as come within the true spirit and scope of the invention.
Claims (19)
1. A method of protecting "ki" data symbols against er-rors, said method including the steps of:
A. segmenting the ki data symbols into i sections of k symbols each and encoding each of the data sections in accordance with a first error correction code to generate "i" level-1 code words, each of said level-1 code words containing "k" data symbols and "r"
redundancy symbols;
B. sectioning said r redundancy symbols in each level-1 code word into "m" multi-symbol sections R(1), R(2), .... , R(m);
C. encoding the corresponding redundancy symbols in section R(2) from each of said level-1 code words in accordance with an error correction code to generate level-2 code words, each of said level-2 coda words containing as information symbols "i" redundancy symbols generated using said first error correction code and "q2" redundancy symbols, the symbols so generated being associated with a section Q(2) of redundancy symbols;
D. repeating step C for each of the remaining R(3)...R(m) redundancy symbol sections, said encod-ing generating sections Q(3), ... , Q(m) redundancy symbols; and E. retaining said data symbols, said section R(1) redundancy symbols and said sections Q(2), Q(3), ...
, Q(m) redundancy symbols.
A. segmenting the ki data symbols into i sections of k symbols each and encoding each of the data sections in accordance with a first error correction code to generate "i" level-1 code words, each of said level-1 code words containing "k" data symbols and "r"
redundancy symbols;
B. sectioning said r redundancy symbols in each level-1 code word into "m" multi-symbol sections R(1), R(2), .... , R(m);
C. encoding the corresponding redundancy symbols in section R(2) from each of said level-1 code words in accordance with an error correction code to generate level-2 code words, each of said level-2 coda words containing as information symbols "i" redundancy symbols generated using said first error correction code and "q2" redundancy symbols, the symbols so generated being associated with a section Q(2) of redundancy symbols;
D. repeating step C for each of the remaining R(3)...R(m) redundancy symbol sections, said encod-ing generating sections Q(3), ... , Q(m) redundancy symbols; and E. retaining said data symbols, said section R(1) redundancy symbols and said sections Q(2), Q(3), ...
, Q(m) redundancy symbols.
2. The error protection method of claim 1, wherein each of said R(2), ...R(m) sections of said r redundancy symbols is encoded using the same error correction code.
3. The error protection method of claim 1, wherein each of said redundancy symbol sections is encoded using a different error correction code, and said error correction codes used to encode redundancy sections R(3), ..., R(m), respectively, are designed to correct fewer errors than the codes used to encode the previous section R(2) ... R(m-1).
4. The error protection method of claim 1, wherein said method further includes the step of interleaving the ki data symbols.
5. The error protection method of claim 1, wherein said method further includes the steps of:
F. treating said section R(2), R(3), ..., R(m) symbols which are not retained in step E as erasurss;
G. filling in, in each level-1 code word, as many of said erasures and correcting as many errors in said code words as possible using said first error cor-rection code;
H. treating said data symbols as error-free, if all of said redundancy symbols in said level-1 code word are filled in;
I. filling in, in each level-1 code word in which said section R(2) symbols not filled in, said erasures in section R(2) using said section Q(2) redundancy sym-bols and said error correction code used to encode said section R(2) symbols;
J. correcting, using said first error correction code and said section R(1) and R(2) redundancy symbols, any errors in said level-1 code words in which said section R(2) symbols were filled in by step I;
K. if any redundancy symbols in said level-1 code words are not filled in:
i. filling in said section R(3) erasures using said section Q(3) redundancy symbols and said error correction code used to encode said sec-tion R(3) symbols, and ii. correcting, using said first error correction code and said section R(1), R(2) and R(3) sym-bols, any errors in said level-1 code words in which said section R(3) symbols were filled in by this step J; and L. repeating step J for each remaining section of said r redundancy symbols.
F. treating said section R(2), R(3), ..., R(m) symbols which are not retained in step E as erasurss;
G. filling in, in each level-1 code word, as many of said erasures and correcting as many errors in said code words as possible using said first error cor-rection code;
H. treating said data symbols as error-free, if all of said redundancy symbols in said level-1 code word are filled in;
I. filling in, in each level-1 code word in which said section R(2) symbols not filled in, said erasures in section R(2) using said section Q(2) redundancy sym-bols and said error correction code used to encode said section R(2) symbols;
J. correcting, using said first error correction code and said section R(1) and R(2) redundancy symbols, any errors in said level-1 code words in which said section R(2) symbols were filled in by step I;
K. if any redundancy symbols in said level-1 code words are not filled in:
i. filling in said section R(3) erasures using said section Q(3) redundancy symbols and said error correction code used to encode said sec-tion R(3) symbols, and ii. correcting, using said first error correction code and said section R(1), R(2) and R(3) sym-bols, any errors in said level-1 code words in which said section R(3) symbols were filled in by this step J; and L. repeating step J for each remaining section of said r redundancy symbols.
6. The error protection method of claim 5, wherein each of said R(2), ..., R(m) sections of said r redundancy symbols is encoded using the same error correction code.
7. The error protection method of claim 5, wherein each of said redundancy symbol sections is encoded using a different error correction code, and said error correction codes used to encode redundancy sections R(3), ..., R(m) are designed to cor-rect fewer errors than the codes used to encode the data, and sections R(2), ... R(m-1), respectively.
8. An error correction system for encoding "ki" data symbols to protect said data symbols against errors, said sys-tem including:
A. a first encoder means for separating the data into "i" sections of k symbols each and encoding the sec-tions in accordance with a first error correction code to generate i level-1 code words, each of said level-1 code words containing "k" data symbols and "r" redundancy symbols;
B. means for sectioning said r redundancy symbols in each level-1 code word into "m" multi-symbol sec-tions R(1), R(2), ...., R(m);
C. a second encoder means for encoding the correspond-ing redundancy symbols in section R(2) from each of said level-1 code words in accordance with an error correction code to generate level-2 code words, each of said level-2 code words containing as information symbols "i" redundancy symbols generated using said first error correction code and "q2" redundancy sym-bols, the symbols so generated being associated with a section Q(2) of redundancy symbols, said second encoder means encoding each of the remaining redundancy symbol sections to generate sections Q(3), ... , Q(m) of redundancy symbols, respective-ly; and D. means for retaining said data symbols, said section R(1) redundancy symbols and said sections Q(2), Q(3), ... , Q(m) redundancy symbols.
A. a first encoder means for separating the data into "i" sections of k symbols each and encoding the sec-tions in accordance with a first error correction code to generate i level-1 code words, each of said level-1 code words containing "k" data symbols and "r" redundancy symbols;
B. means for sectioning said r redundancy symbols in each level-1 code word into "m" multi-symbol sec-tions R(1), R(2), ...., R(m);
C. a second encoder means for encoding the correspond-ing redundancy symbols in section R(2) from each of said level-1 code words in accordance with an error correction code to generate level-2 code words, each of said level-2 code words containing as information symbols "i" redundancy symbols generated using said first error correction code and "q2" redundancy sym-bols, the symbols so generated being associated with a section Q(2) of redundancy symbols, said second encoder means encoding each of the remaining redundancy symbol sections to generate sections Q(3), ... , Q(m) of redundancy symbols, respective-ly; and D. means for retaining said data symbols, said section R(1) redundancy symbols and said sections Q(2), Q(3), ... , Q(m) redundancy symbols.
9. The error correction system of claim 8, wherein said second encoder means encodes each of said R(2), ..., R(m) sec-tions of said r redundancy symbols using the same error correc-tion code.
10. The error correction system of claim 8, wherein said second encoder means encodes each of said R(2), ..., R(m) sec-tions of said r redundancy symbols using for each section an error correction code which is designed to correct fewer errors than the code used to encode the symbols of the previous sec-tion.
11. The error correction system of claim 8, wherein said first encoder means interleaves the ki data symbols before separating them into said i sections.
12. The error correction system of claim 8, wherein said system further includes:
E. a first correcting means for:
i. correcting said section R(2), ..., R(m) redundancy symbols as erasures and correcting any erroneous retained data and section R(1) symbols using said level-1 error correction code, and ii. treating said data symbols as error-free, if all of said redundancy symbols in said level-1 code word are filled in; and F. a second correcting means for:
i. filling in, in turn, said section R(2), ...., R(m) symbols not filled in by said first cor-recting means using said section Q(2), ....
Q(m) redundancy symbols and said error correc-tion code used to encode said section R(2), ..., R(m) symbols, respectively, and ii. applying to said first correcting means for further error correction, after each of said sections R(2)...R(m) is filled in, said level-1 code words.
E. a first correcting means for:
i. correcting said section R(2), ..., R(m) redundancy symbols as erasures and correcting any erroneous retained data and section R(1) symbols using said level-1 error correction code, and ii. treating said data symbols as error-free, if all of said redundancy symbols in said level-1 code word are filled in; and F. a second correcting means for:
i. filling in, in turn, said section R(2), ...., R(m) symbols not filled in by said first cor-recting means using said section Q(2), ....
Q(m) redundancy symbols and said error correc-tion code used to encode said section R(2), ..., R(m) symbols, respectively, and ii. applying to said first correcting means for further error correction, after each of said sections R(2)...R(m) is filled in, said level-1 code words.
13. A method of generating error correction redundancy symbols corresponding to "ki" data symbols, said method includ-ing the steps of:
A. sectioning the ki data symbols into i sections and encoding the sections over a Galois Field GF(2t) using a first error correction code to generate "i"
level-1 code words, each of said level-1 code words containing "k" data symbols and "s" redundancy sym-bols, said first error correction code having a gen-erator polynomial, g1(x), with consecutive roots in GF(2t):
g1(x)=(x+.alpha.a)*(x+.alpha.a+l)*(X+.alpha.a+2)*(x+.alpha.a+3)*. .*(x+.alpha.a+s) where * represents Galois Field multiplication and .alpha. is a primitive polynomial of GF(2t), said redundancy symbols being associated with a section S(1) of redundancy symbols;
B. encoding, to generate a second set of redundancy symbols as-sociated with a section S(2), said data symbols in each of said i level-1 code words and the corresponding symbols in section S(1) using a second error correction code having a generator polynomial, g*2(x), which is equal to the quotient of a genera-tor polynomial, g2(x) g (x) = (x+.alpha.a-d)*(x+.alpha.a-d+1)*...*(x-.alpha.a)*...*(x+.alpha.a+s)*
(x+.alpha.a+s+1) *... * (x+.alpha.as+d) divided by said generator polynomial, g1(x), said polynomial g2(x) being an expansion of g1(x);
C. repeating Step B a predetermined number of times and gener-ating sections S(3)...S(m) of redundancy symbols;
D. further encoding said section S(2)...S(m) redundancy sym-bols to generate associated sections of redundancy symbols Q(2)...Q(m);
E. encoding each of said sections Q(2)...Q(m) of redundancy symbols to generate sections ECK(2)...ECK(m) redundancy sym-bols; and F. retaining said data symbols, said section S(1) redundancy symbols, said section Q(2)...Q(m) redundancy symbols and said ECK(2)...ECK(m) redundancy symbols.
A. sectioning the ki data symbols into i sections and encoding the sections over a Galois Field GF(2t) using a first error correction code to generate "i"
level-1 code words, each of said level-1 code words containing "k" data symbols and "s" redundancy sym-bols, said first error correction code having a gen-erator polynomial, g1(x), with consecutive roots in GF(2t):
g1(x)=(x+.alpha.a)*(x+.alpha.a+l)*(X+.alpha.a+2)*(x+.alpha.a+3)*. .*(x+.alpha.a+s) where * represents Galois Field multiplication and .alpha. is a primitive polynomial of GF(2t), said redundancy symbols being associated with a section S(1) of redundancy symbols;
B. encoding, to generate a second set of redundancy symbols as-sociated with a section S(2), said data symbols in each of said i level-1 code words and the corresponding symbols in section S(1) using a second error correction code having a generator polynomial, g*2(x), which is equal to the quotient of a genera-tor polynomial, g2(x) g (x) = (x+.alpha.a-d)*(x+.alpha.a-d+1)*...*(x-.alpha.a)*...*(x+.alpha.a+s)*
(x+.alpha.a+s+1) *... * (x+.alpha.as+d) divided by said generator polynomial, g1(x), said polynomial g2(x) being an expansion of g1(x);
C. repeating Step B a predetermined number of times and gener-ating sections S(3)...S(m) of redundancy symbols;
D. further encoding said section S(2)...S(m) redundancy sym-bols to generate associated sections of redundancy symbols Q(2)...Q(m);
E. encoding each of said sections Q(2)...Q(m) of redundancy symbols to generate sections ECK(2)...ECK(m) redundancy sym-bols; and F. retaining said data symbols, said section S(1) redundancy symbols, said section Q(2)...Q(m) redundancy symbols and said ECK(2)...ECK(m) redundancy symbols.
14. The method of claim 13, wherein said method further includes interleaving the ki data symbols.
15. The method of claim 13, wherein said method further includes the steps of:
F. correcting said data symbols using said section S(1) redundancy symbols and said first error cor-rection code;
G. if errors in said data symbols remain filling in, each code word, said section S(2) redundancy sym-bols using said section Q(2) redundancy symbols and said error correction code used to generate said section Q(2) redundancy symbols;
H. correcting said erroneous data symbols using said filled in section S(2) redundancy symbols and said second error correction code having said generator polynomial g*2(x);
I. if errors in said data symbols remain, filling in, in each code word, the next said section of redundancy symbols, S(3)...S(m) using said cor-responding section Q(3)...Q(m) redundancy symbols and error correction code associated with said level; and J. correcting any errors using an error correction code associated with that level of encoding.
F. correcting said data symbols using said section S(1) redundancy symbols and said first error cor-rection code;
G. if errors in said data symbols remain filling in, each code word, said section S(2) redundancy sym-bols using said section Q(2) redundancy symbols and said error correction code used to generate said section Q(2) redundancy symbols;
H. correcting said erroneous data symbols using said filled in section S(2) redundancy symbols and said second error correction code having said generator polynomial g*2(x);
I. if errors in said data symbols remain, filling in, in each code word, the next said section of redundancy symbols, S(3)...S(m) using said cor-responding section Q(3)...Q(m) redundancy symbols and error correction code associated with said level; and J. correcting any errors using an error correction code associated with that level of encoding.
16. A system for generating error correction redundancy symbols corresponding to "ki" data symbols, said system in-cluding:
A. a first encoding means for separating said ki data symbols into i sections and encoding the i sections over a Galois Field GF(2t) using a first error cor-rection code to generate "i" level-1 code words, each of said level-1 code words containing "k" data symbols and "s" redundancy symbols, said first er-ror correction coda having a generator polynomial, g1(x), with consecutive roots in GF(2t):
g1(x) = (x+.alpha.a) * (x+.alpha.a+l) * (x+.alpha.a+2) * (x+.alpha.a+3) *...* (x+.alpha.a+s) where * represents Galois Field multiplication and .alpha.
is a primitive polynomial of GF(2t), said redundancy symbols being associated with a section S(1) of redundancy symbols;
B. encoding means for encoding, to generate a section of redundancy symbols S(2), said data symbols and said section S(1) redundancy symbols in each of said i level-1 code words using a second error cor-rection code having a generator polynomial, g*2(x), which is equal to the quotient of g2(x) which is an expanded version of g1(x):
g2(x) = (x+.alpha.a-d)*(x+.alpha.a-d+1)*...*(x+.alpha.a)*...*(x+.alpha.a+s)*
(x+a+s+1)* ... *(x+.alpha.a+s+d) divided by g1(x);
C. encoding means for encoding said section S(2) redundancy symbols to generate a section Q(2) of redundancy symbols;
D. encoding means for encoding said section Q(2) redundancy symbols to generate a section ECK(2) redundancy symbols; and E. means for retaining said data symbols, said section S(1) redundancy symbols, said saction Q(2) and sec-tion ECK(2) redundancy symbols.
A. a first encoding means for separating said ki data symbols into i sections and encoding the i sections over a Galois Field GF(2t) using a first error cor-rection code to generate "i" level-1 code words, each of said level-1 code words containing "k" data symbols and "s" redundancy symbols, said first er-ror correction coda having a generator polynomial, g1(x), with consecutive roots in GF(2t):
g1(x) = (x+.alpha.a) * (x+.alpha.a+l) * (x+.alpha.a+2) * (x+.alpha.a+3) *...* (x+.alpha.a+s) where * represents Galois Field multiplication and .alpha.
is a primitive polynomial of GF(2t), said redundancy symbols being associated with a section S(1) of redundancy symbols;
B. encoding means for encoding, to generate a section of redundancy symbols S(2), said data symbols and said section S(1) redundancy symbols in each of said i level-1 code words using a second error cor-rection code having a generator polynomial, g*2(x), which is equal to the quotient of g2(x) which is an expanded version of g1(x):
g2(x) = (x+.alpha.a-d)*(x+.alpha.a-d+1)*...*(x+.alpha.a)*...*(x+.alpha.a+s)*
(x+a+s+1)* ... *(x+.alpha.a+s+d) divided by g1(x);
C. encoding means for encoding said section S(2) redundancy symbols to generate a section Q(2) of redundancy symbols;
D. encoding means for encoding said section Q(2) redundancy symbols to generate a section ECK(2) redundancy symbols; and E. means for retaining said data symbols, said section S(1) redundancy symbols, said saction Q(2) and sec-tion ECK(2) redundancy symbols.
17. The system of claim 16, wherein said system further in-cludes:
F. encoding means for encoding, in levels, said data symbols and said sections of redundancy symbols S(i) generated in previous levels of encoding to generate a next section S(i+1) of redundancy sym-bols using an error corre.cting code having a gener-ator polynomial, g*i+1(x), which is equal to the quotient generator polynomial, gi+1(x), which is an expansion of the generator polynomial gi(x) associ-ated with the preceding level, divided by said gen-erator polynomial gi(x);
G. encsding means for encoding each of said sections S(i) of redundancy symbols to generate another sec-tion of redundancy symbols Q(i);
H. encoding means for encoding each of said Q(i) sec-tions of redundancy symbols to generate ECK(i) sec-tions of redundancy symbols; and I. means for retaining said data symbols, said section S(1) redundancy symbols, said section Q(2), ....
Q(i) of redundancy symbols, and said sections ECK(2), ...., ECK(i) sets of redundancy symbols.
F. encoding means for encoding, in levels, said data symbols and said sections of redundancy symbols S(i) generated in previous levels of encoding to generate a next section S(i+1) of redundancy sym-bols using an error corre.cting code having a gener-ator polynomial, g*i+1(x), which is equal to the quotient generator polynomial, gi+1(x), which is an expansion of the generator polynomial gi(x) associ-ated with the preceding level, divided by said gen-erator polynomial gi(x);
G. encsding means for encoding each of said sections S(i) of redundancy symbols to generate another sec-tion of redundancy symbols Q(i);
H. encoding means for encoding each of said Q(i) sec-tions of redundancy symbols to generate ECK(i) sec-tions of redundancy symbols; and I. means for retaining said data symbols, said section S(1) redundancy symbols, said section Q(2), ....
Q(i) of redundancy symbols, and said sections ECK(2), ...., ECK(i) sets of redundancy symbols.
18. The system of claim 16, wherein said system further includes:
F. means for correcting said data symbols using said section S(1) redundancy symbols and said first er-ror correction code;
G. if errors in said data symbols remain, means for filling in, in each code word said section S(2) redundancy symbols using said section Q(2) redundancy symbols and said error correction code used to generate said section Q(2) redundancy sym-bols;
H. correcting said erroneous data symbols using said filled in section S(2) redundancy symbols and said second error correcting code having said generator polynomial g*2(x), which is equal to the quotient of generator polynomial, g2(x), which is an expan-sion of said level-1 generator polynomial, divided by said level-1 generator polynomial.
F. means for correcting said data symbols using said section S(1) redundancy symbols and said first er-ror correction code;
G. if errors in said data symbols remain, means for filling in, in each code word said section S(2) redundancy symbols using said section Q(2) redundancy symbols and said error correction code used to generate said section Q(2) redundancy sym-bols;
H. correcting said erroneous data symbols using said filled in section S(2) redundancy symbols and said second error correcting code having said generator polynomial g*2(x), which is equal to the quotient of generator polynomial, g2(x), which is an expan-sion of said level-1 generator polynomial, divided by said level-1 generator polynomial.
19. The system of claim 17, wherein said system further includes:
J. means for correcting said data symbols using said section S(1) redundancy symbols and said first er-ror correction code;
K. if errors in said data symbols remain, means for filling in, in each code word said section S(2) redundancy symbols using said section Q(2) redundancy symbols and said error correction code used to generate said Q(2) redundancy symbols;
L. correcting said erroneous data symbols using said filled in section S(2) redundancy symbols and said second error correction code having said generator polynomial g*2(x), which is equal to the quotient of generator polynomial g2(x), which is an expan-sion of said level-1 generator polynomial, divided by said level-1 generator polynomial;
M. if errors in said data symbols remain, means for filling in, in each code word said section of redundancy symbols S(i) generated by said next en-coding step using said corresponding section Q(i) redundancy symbols and said corresponding error correction code;
N. correcting said erroneous data symbols using said filled-in section S(i) redundancy symbols and said corresponding error correction code having as a generator polynomial g*i(x) which is equal to the quotient of said generator polynomial gi(x), which is in expansion of the previous level generator polynomial, divided by said previous level genera-tor polynomial.
J. means for correcting said data symbols using said section S(1) redundancy symbols and said first er-ror correction code;
K. if errors in said data symbols remain, means for filling in, in each code word said section S(2) redundancy symbols using said section Q(2) redundancy symbols and said error correction code used to generate said Q(2) redundancy symbols;
L. correcting said erroneous data symbols using said filled in section S(2) redundancy symbols and said second error correction code having said generator polynomial g*2(x), which is equal to the quotient of generator polynomial g2(x), which is an expan-sion of said level-1 generator polynomial, divided by said level-1 generator polynomial;
M. if errors in said data symbols remain, means for filling in, in each code word said section of redundancy symbols S(i) generated by said next en-coding step using said corresponding section Q(i) redundancy symbols and said corresponding error correction code;
N. correcting said erroneous data symbols using said filled-in section S(i) redundancy symbols and said corresponding error correction code having as a generator polynomial g*i(x) which is equal to the quotient of said generator polynomial gi(x), which is in expansion of the previous level generator polynomial, divided by said previous level genera-tor polynomial.
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| US07/521,224 | 1990-05-09 | ||
| US07/521,224 US5224106A (en) | 1990-05-09 | 1990-05-09 | Multi-level error correction system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CA2041214A1 true CA2041214A1 (en) | 1991-11-10 |
Family
ID=24075899
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CA002041214A Abandoned CA2041214A1 (en) | 1990-05-09 | 1991-04-25 | Multi-level error correction system |
Country Status (5)
| Country | Link |
|---|---|
| US (1) | US5224106A (en) |
| EP (1) | EP0458468B1 (en) |
| JP (1) | JPH04230124A (en) |
| CA (1) | CA2041214A1 (en) |
| DE (1) | DE69121307T2 (en) |
Families Citing this family (30)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP0481128B1 (en) * | 1990-10-16 | 1998-01-14 | Koninklijke Philips Electronics N.V. | Data processor system based on an (N, k) symbol code having symbol error correctibility and plural error mendability |
| US5734826A (en) * | 1991-03-29 | 1998-03-31 | International Business Machines Corporation | Variable cyclic redundancy coding method and apparatus for use in a multistage network |
| US5379305A (en) * | 1992-07-20 | 1995-01-03 | Digital Equipment Corporation | Error correction system with selectable error correction capabilities |
| US5502732A (en) * | 1993-09-20 | 1996-03-26 | International Business Machines Corporation | Method for testing ECC logic |
| JPH08509351A (en) * | 1994-02-16 | 1996-10-01 | フィリップス エレクトロニクス ネムローゼ フェンノートシャップ | Method and apparatus for error-correctable data transmission based on semi-cyclic code |
| AU2905695A (en) * | 1994-06-21 | 1996-01-15 | Michael J. Seo | Method and system for encoding and decoding signals using a fast algebraic error correcting code |
| JP3328093B2 (en) * | 1994-07-12 | 2002-09-24 | 三菱電機株式会社 | Error correction device |
| JPH08293802A (en) * | 1995-04-13 | 1996-11-05 | Internatl Business Mach Corp <Ibm> | Interleaving type error correction method |
| US5898710A (en) * | 1995-06-06 | 1999-04-27 | Globespan Technologies, Inc. | Implied interleaving, a family of systematic interleavers and deinterleavers |
| US5983383A (en) * | 1997-01-17 | 1999-11-09 | Qualcom Incorporated | Method and apparatus for transmitting and receiving concatenated code data |
| GB2321374A (en) * | 1997-01-21 | 1998-07-22 | Ico Services Ltd | Spread spectrum satellite communication |
| US6514140B1 (en) * | 1999-06-17 | 2003-02-04 | Cias, Inc. | System for machine reading and processing information from gaming chips |
| MXPA03009531A (en) * | 2001-04-19 | 2004-12-06 | Graham Packaging Co | Multi-functional base for a plastic wide-mouth, blow-molded container. |
| CN1611027A (en) * | 2001-12-28 | 2005-04-27 | 皇家飞利浦电子股份有限公司 | Unequal error protection using forward error correction based on reed-Solomon codes |
| US7274642B2 (en) * | 2002-02-20 | 2007-09-25 | Sony Corporation | Data recording medium, data recording method and apparatus, data reproducing method and apparatus, data transmitting method, and data receiving method |
| US7546510B2 (en) * | 2003-12-30 | 2009-06-09 | Sandisk Il Ltd. | Compact high-speed single-bit error-correction circuit |
| US20050289433A1 (en) * | 2004-06-25 | 2005-12-29 | Itschak Weissman | Discrete universal denoising with error correction coding |
| TWI415416B (en) * | 2005-09-19 | 2013-11-11 | St Ericsson Sa | Apparatus and method for error correction in mobile wireless applications incorporating erasure table data |
| US8219886B1 (en) | 2006-01-20 | 2012-07-10 | Marvell International Ltd. | High density multi-level memory |
| US7721178B2 (en) * | 2006-06-01 | 2010-05-18 | International Business Machines Corporation | Systems, methods, and computer program products for providing a two-bit symbol bus error correcting code |
| EP1887446A1 (en) * | 2006-08-02 | 2008-02-13 | Siemens Aktiengesellschaft | Method for serial asynchronous transfer of data in an arrangement for the surveillance, control and regulation of a technical installation of a building automation system |
| KR101254185B1 (en) * | 2006-08-03 | 2013-04-18 | 삼성전자주식회사 | Apparatus for recording data in information recording medium in which extra ECC is applied or reproducing data from the medium, and the method therefor |
| US8627167B1 (en) * | 2007-01-08 | 2014-01-07 | Marvell International Ltd. | Methods and apparatus for providing multi-layered coding for memory devices |
| US8145975B2 (en) * | 2008-02-28 | 2012-03-27 | Ip Video Communications Corporation | Universal packet loss recovery system for delivery of real-time streaming multimedia content over packet-switched networks |
| US8402351B1 (en) * | 2009-02-09 | 2013-03-19 | Marvell International Ltd. | Error correction coding with nuisance data |
| FR2961613B1 (en) * | 2010-06-18 | 2012-07-27 | Commissariat Energie Atomique | MEMORY PROTECTION METHOD CONFIGURABLE AGAINST PERMANENT AND TRANSIENT ERRORS AND APPARENT DEVICE |
| US8832524B2 (en) * | 2011-09-22 | 2014-09-09 | Violin Memory, Inc. | System and method for correcting errors in data using a compound code |
| WO2015099568A1 (en) * | 2013-12-27 | 2015-07-02 | Евгений Иванович ПРЯХИН | Encoding and decoding digital information in the form of ultra-compressed nano-barcode |
| US9553611B2 (en) * | 2014-11-27 | 2017-01-24 | Apple Inc. | Error correction coding with high-degree overlap among component codes |
| US11764813B1 (en) * | 2022-06-07 | 2023-09-19 | Western Digital Technologies, Inc. | Extendable parity code matrix construction and utilization in a data storage device |
Family Cites Families (23)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US3629824A (en) * | 1970-02-12 | 1971-12-21 | Ibm | Apparatus for multiple-error correcting codes |
| GB2095440B (en) * | 1981-03-23 | 1985-10-09 | Sony Corp | Digital television signal processing |
| US4413339A (en) * | 1981-06-24 | 1983-11-01 | Digital Equipment Corporation | Multiple error detecting and correcting system employing Reed-Solomon codes |
| BE890280A (en) * | 1981-09-09 | 1982-03-09 | Belge Lampes Mat Electr Mble | ENCODING AND DECODING DEVICE BASED ON A CONVOLUTION CODE |
| GB2132393B (en) * | 1982-12-17 | 1986-05-14 | Sony Corp | Methods and apparatus for correcting errors in binary data |
| NL8300249A (en) * | 1983-01-25 | 1984-08-16 | Philips Nv | ERROR CORRECTION METHOD FOR TRANSMITTING DATA ORDERED DATES AND DEVICES FOR CARRYING OUT THE METHOD |
| US4525838A (en) * | 1983-02-28 | 1985-06-25 | International Business Machines Corporation | Multibyte error correcting system involving a two-level code structure |
| JP2533076B2 (en) * | 1983-04-30 | 1996-09-11 | ソニー株式会社 | Encoding method for error correction |
| JPS6069917A (en) * | 1983-09-26 | 1985-04-20 | Pioneer Electronic Corp | Data transmission system |
| JPH0812612B2 (en) * | 1983-10-31 | 1996-02-07 | 株式会社日立製作所 | Error correction method and apparatus |
| US4769818A (en) * | 1984-05-30 | 1988-09-06 | Canadian Patents And Development Limited-Societe Canadienne Des Brevets Et D'exploitation Limitee | Method and apparatus for coding digital data to permit correction of one or two incorrect data packets (bytes) |
| JPS61154227A (en) * | 1984-12-26 | 1986-07-12 | Mitsubishi Electric Corp | Two-stage coding method |
| JPH06101207B2 (en) * | 1985-03-13 | 1994-12-12 | ソニー株式会社 | Error detection and correction method |
| US4706250A (en) * | 1985-09-27 | 1987-11-10 | International Business Machines Corporation | Method and apparatus for correcting multibyte errors having improved two-level code structure |
| CA1264091A (en) * | 1986-01-10 | 1989-12-27 | Yoichiro Sako | Generator for error correcting code and decoder for the code |
| US4763330A (en) * | 1986-05-06 | 1988-08-09 | Mita Industrial Co., Ltd. | Syndrome calculating apparatus |
| DE3750456T2 (en) * | 1986-09-22 | 1995-04-20 | Nec Corp | Error correcting, bit serial decoder. |
| US4782490A (en) * | 1987-03-16 | 1988-11-01 | Cythera Corporation | Method and a system for multiple error detection and correction |
| US4856003A (en) * | 1987-05-07 | 1989-08-08 | Digital Equipment Corporation | Error correction code encoder |
| US4833679A (en) * | 1987-08-31 | 1989-05-23 | International Business Machines Corporation | Method and apparatus with improved error correction and error information availability |
| US4949200A (en) * | 1988-03-31 | 1990-08-14 | Digital Equipment Corporation | Method and apparatus for encoding and mapping magnetic disk sector addresses |
| US4989211A (en) * | 1988-05-12 | 1991-01-29 | Digital Equipment Corporation | Sector mis-synchronization detection method |
| DE3838234A1 (en) * | 1988-11-11 | 1990-05-17 | Broadcast Television Syst | METHOD AND CIRCUIT FOR DETECTING AND CORRECTING ERRORS IN DATA WORDS |
-
1990
- 1990-05-09 US US07/521,224 patent/US5224106A/en not_active Expired - Lifetime
-
1991
- 1991-04-25 CA CA002041214A patent/CA2041214A1/en not_active Abandoned
- 1991-04-26 DE DE69121307T patent/DE69121307T2/en not_active Expired - Fee Related
- 1991-04-26 EP EP91303775A patent/EP0458468B1/en not_active Expired - Lifetime
- 1991-05-08 JP JP3102380A patent/JPH04230124A/en active Pending
Also Published As
| Publication number | Publication date |
|---|---|
| US5224106A (en) | 1993-06-29 |
| DE69121307D1 (en) | 1996-09-19 |
| EP0458468B1 (en) | 1996-08-14 |
| DE69121307T2 (en) | 1997-01-23 |
| EP0458468A1 (en) | 1991-11-27 |
| JPH04230124A (en) | 1992-08-19 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CA2041214A1 (en) | Multi-level error correction system | |
| US5465260A (en) | Dual purpose cyclic redundancy check | |
| KR100913965B1 (en) | Coding and decoding of partially a priori known information | |
| US6615387B1 (en) | Method and apparatus for error detection | |
| US7188295B2 (en) | Method and apparatus for embedding an additional layer of error correction into an error correcting code | |
| KR930008683B1 (en) | Reed solomon error correction code encoder | |
| US5712863A (en) | Randomizing encoder for digital data storage | |
| US6504493B1 (en) | Method and apparatus for encoding/decoding data | |
| EP0998791A1 (en) | Concurrent row/column syndrome generator for a product code | |
| WO1993018589A1 (en) | Data recovery after error correction failure | |
| US5948117A (en) | Modified Reed-Solomon error correction system using (W+i+1)-bit representations of symbols of GF(2w+i) | |
| EP1421699A2 (en) | Enhanced coding for informed decoders | |
| US5889792A (en) | Method and apparatus for generating syndromes associated with a block of data that employs re-encoding the block of data | |
| WO1999036913A1 (en) | Methods and apparatus for data storage on magnetic media including error correction codes | |
| US5822336A (en) | Modified reed solomon code selection and encoding system | |
| US7174497B2 (en) | Method of storing or decoding a stream of bits | |
| US6898754B2 (en) | Error detection for data storage and transmission | |
| US20050229085A1 (en) | Method and apparatus for embedding an additional layer of error correction into an error correcting code | |
| CA2246423C (en) | Encoder for digital data storage | |
| US5889794A (en) | Two-level error correction encoder | |
| JP4135497B2 (en) | Data reproduction method and data reproduction apparatus | |
| CN1599981A (en) | Simple decoding method and apparatus | |
| JPH08509351A (en) | Method and apparatus for error-correctable data transmission based on semi-cyclic code | |
| KR20040014723A (en) | Encoding apparatus and method, and decoding apparatus and method for correcting error | |
| CA2466569A1 (en) | Encoder for digital data storage |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| EEER | Examination request | ||
| FZDE | Discontinued |