JPH05500581A - Error correction method and device - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

[Detailed description of the invention]

Error correction method and device background 1. Field of invention The present invention relates to a method and apparatus for error correction of stored or transmitted data, and in detail, the entire information block or multiple information blocks recorded on the storage medium. The present invention relates to a method and apparatus for correcting errors caused by loss of data. 2. Prior art and other considerations For decades, digital information has been stored on magnetic media, including magnetic tape and magnetic disks. It has been recorded. Unfortunately, data storage in or reading data from a medium Noise generated in the output causes errors. To alleviate this problem, seeds Various encoding techniques specially encode transmitted or stored data to provide error correction capabilities. It is developed for encoding. Generally, digital information is stored in the form of physical data blocks, with each A block consists of multiple bits. Prior art error encoding techniques allow information or Within each block set of tsage bits, check bits are used to form a codeword. Granted. The check bits for the codeword are a predetermined code, such as a Reed-Solomon code. is derived by an encoder that operates on a set of information bits using a code. encoder conveys the desired characteristics to the codeword, so that the continuation of the block (by reading it, the codeword The codeword is decoded in such a way that errors are identifiable and correctable. decoder An example is given in Chr. istoper P, established in U.S. Patent No. 4.845.713 to Zook. referenced here. By acting on blocks, you can create two or more marks. The codes are combined to produce a stronger code and an error correction tree is created for the block. A second level of wages has been established. Such a combination of codes is called "two-dimensional code" ” or “product code” in the literature. Errors driven by product codes An example of a correction system is "Multibus Error Correction Process and Apparatus for Product Codes" U.S. Patent No. 4.845.714 to Christopher P. Zook entitled Established in No. Current error correction mechanisms generally correct errors that occur within blocks of recorded information. Enough to correct. Additionally, not all blocks may be readable. . Even worse, some blocks may not be readable. Ella - Error correction at the block level where correction bits are embedded in blocks is It is completely useless when one or more whole blocks are lost. One example of a block loss problem occurs in the environment of spiral scan recording. Spiral In a scanning arrangement, a moving magnetic tape is partially wrapped around a rotating drum. as a result, the head positioned on the drum will move around the drum as it rotates. adjacent to. The write head on the drum is relative to the direction of tape movement Physically records data on tape in a series of discrete strips oriented at an angle do. Data is typically formatted into multiple physical blocks per strip be converted into An example of a spiral scan recording system is disclosed in U.S. Pat. In National Patent No. 4.835.638 and U.S. Patent No. 4.845.577 provided. In the spiral scanning technique, there are two movements with respect to the tape. The first operation is the strip diagonal steps on the tape to write and/or read blocks on the tape. This is the movement of the head according to the trip. The second movement is reel to reel tape This is a linear motion of the tape on the guide. Therefore, in the helical scanning technique, there are two types of errors that occur as a result of the two movements. There is. The first type of error destroys multiple blocks in the strip. The second type of error refers to blocks at the same location in consecutive strips (i.e. i.e. blocks aligned in the direction of tape movement). Current error correction mark The technology does not handle the destruction of multiple blocks. Therefore, in order to correct errors related to recording or reading information on a storage medium, It is an object of the present invention to provide a method and apparatus for this purpose. An advantage of the invention is that all data stored on or read from a storage medium A method and apparatus are provided that facilitate correction of errors associated with loss of blocks. Yet another advantage of the invention is that during storage to or reading from a storage medium Method and apparatus for facilitating reconstruction of entire strips or rows of lost data blocks Preparation for installation. summary A method of encoding and decoding information data to be transmitted to a storage medium, comprising: Group G of physical blocks written in the body is a user data information block and a complementary Contains both an auxiliary and error correction block. Each block included in group G (B,... e) belongs to row 1 and column 0. The information data is reserved for the user data information block. Formatted to selected bit positions. Auxiliary error - correction block The value of the bit position that is generated by performing an exclusive OR operation on the values in corresponding bit positions in the Ru. By strategically selecting user data blocks included in the sag group Thus, the method facilitates recovery of entire blocks and even rows or columns of blocks. In a preferred mode of the invention, the columns of each group G are each The information data block is written by the system as a spiral 1~rip on the magnetic tape. block in the strip preceding the strip containing the auxiliary error correction block. recorded. Each strip comprises k number of blocks. Each group G is n A number of strips are provided. Further, according to a preferred embodiment of the invention, m number of error correction blocks El are blocks It is generated to be included in the BI+c group. In this respect, i is The range is 0 to m-1, m=qxk, and q is an integer constant. Each auxiliary gill - The value of each bit position included in the corrected block El is is derived from the sum of bits with corresponding bit positions in . Each auxiliary gill The error correction block has a bit that indicates that the block is an auxiliary error correction block. remembers the cut. In the strip following the strip of the auxiliary correction block The recorded block is marked with a bit indicating that the block is not an auxiliary correction block. I remember that. Using conventional error correction techniques, error correction bits are stored in user data blocks. and inserted into the auxiliary error correction block at other bit positions. Brief description of the drawing The foregoing and other objects, features and advantages of the invention will be apparent from the preferred embodiments illustrated in the accompanying drawings. It is clear from the following detailed explanation by Mr. In this case, the reference characters refer to various drawings. refer to the same part throughout. The drawings, which are not necessarily to scale, illustrate the principles of the invention. Emphasis is placed on that. FIG. 1 is a schematic representation of the head arrangement on a drum of a spiral scanning system according to an embodiment of the invention; It is a diagram. FIG. 2 shows a helical spiral of the embodiment of FIG. 1 for recording strips on magnetic tape. 1 is a schematic diagram of a scanning system. FIG. 3 shows the AUXECC generator included in the helical scanning system of the embodiment of FIG. FIG. FIG. 4 is a diagram illustrating an arrangement for facilitating head placement on a drum according to the embodiment of FIG. 1 is a schematic diagram of a orbital scanning system. FIG. 5 shows the images recorded or read out by the helical scanning system of the embodiment of FIG. 1 is a schematic diagram depicting the format of a magnetic tape; FIG. 6 shows a magnetic tape recorded by the helical scanning system of the embodiment of FIG. 1 is a schematic illustration of a plurality of strips, and further includes a helical scanning system of the embodiment of FIG. The moving path of the head provided on the drum is shown. 7A-70 are images recorded by the helical scanning system of the embodiment of FIG. A schematic diagram depicting the format of track 1 and track 2 of a magnetic tape being read. be. FIG. 8 shows the images recorded or read out by the helical scanning system of the embodiment of FIG. Describes the format of the physical data blocks contained on a strip of magnetic tape. This is a schematic diagram. FIG. 9 shows the images recorded or read out by the helical scanning system of the embodiment of FIG. The physical block header of a physical data block contained on a strip of magnetic tape. FIG. FIG. 10 shows the recording on magnetic tape by the helical scanning system of the embodiment of FIG. 1 is a schematic diagram depicting a group of strips including an AUXECC strip; FIG. 11 shows the recording on magnetic tape by the helical scanning system of the embodiment of FIG. 2 is a schematic diagram depicting a group of strips including an AUXECC strip Ru. Figure 12 depicts an alternative method of generating the AUXECC block.

【drawing A brief explanation of

FIG. 1 shows a spiral scanning system for recording digital information on magnetic tape 32. The drum and transport portion of the system 30 are shown. The spiral scanning system 30 is a tape transport machine. It includes a structure 34 and a rotating drum 36. Tape transport mechanism 34 includes a capstan 42 and two tape guides 44 . In the conventional manner, capstan 42 is directed in the direction of tape movement as indicated by arrow 46. In order to move the tape 32 in this direction, it is rotated by a capstan drive motor (not shown). In the illustrated embodiment, capstan 42 rotates to transport tape 32 at a speed on the order of one-half inch per second. Drum and Head Structure The drum 36 rotates about a drum axis 50. The drum is in a drum mode (not shown). It is rotated by a motor at a speed of about 1800 rpm. A tachometer, not shown, detects the rotational speed of the drum motor shaft and thus the drum and generates the DRUM 5YNC signal. As shown in FIG. 2, the drum shaft 50 and thus the drum 36 are oriented at an angle with respect to the edge of the tape 32 and the direction of travel. de The ram 36 has an upper drum surface 52 and a lower drum surface 54, both of which are flat. Dora The drum axis 50 is orthogonal to both the drum top surface 52 and the drum bottom surface 54. The drum 36 also has a peripheral surface 56 surrounding the circumference of the drum 36. Mounted on the peripheral surface 56 of the drum 36 are two sets of heads, specifically a first set of heads including write heads W1 and W2 and a second set of heads including read heads R1 and R2. Additionally, the peripheral surface 56 of the drum 36 includes a Bohead S is installed. Heads W1, W2, R1 and R2 are mounted to produce a helical strip in magnetic tape 32 in the manner shown in FIG. Heads W1 and W2 essentially simultaneously write first and second tracks of data, tracks T1 and T2, respectively, on tape 32. Heads R1 and R2 rotate 180 degrees after tracks T1 and T2 are written, respectively. It is positioned to read out tracks T1 and T2. In this regard, FIG. 6 shows a write head moving over tracks T1 and T2 to depict the head placement with respect to the tracks. Although read heads W1, W2 and read heads R1, R2 are shown, read heads R1, R2 and write heads W1, W2 are tracked simultaneously in the manner depicted in FIG. It is understood that this is not on the internet. Similarly, FIG. for tracks T1 and T2 and for read heads R1, R2 or write It simply shows the position of the servo head S, regardless of the working heads R1 and R2. In the above respect, the four geometrical factors regarding the heads are: the angular separation of the heads about the drum circumferential surface 56, the axial position of the heads with respect to each other, the head width, and the azimuth orientation of the heads due to the heads W1 and W2. Simultaneous writing and corresponding This strategy allows subsequent reading by heads R1 and R2, respectively. child These geometric factors are further explained in US patent application Ser. FIG. 6 shows a plurality of helical strips recorded on the magnetic tape 32 by the helical scanning system 30, the ultimate movement paths of the heads W1, W2, R1 and R2, and the spiral strips recorded on the magnetic tape 32 by the helical scanning system 30. The servo head S on the lip is shown. The direction of head movement is depicted by arrow 78 in FIG. Thus, a dual channel helical scanning system 20 is provided, with the first channel being The second channel includes heads W1 and R1, and the second channel includes heads W1 and R2. The use of two channels effectively doubles the data transfer rate since twice the number of tracks are written to the tape each revolution of drum 36. For inspection purposes, heads R1 and R2 read back two simultaneously written tracks approximately 180 degrees after the tracks have been written. writing and reading The functions are performed exclusively, thereby eliminating crosstalk problems. System Structure FIG. 4 shows a full helical scanning system 30 including read heads R1, R2, write heads W1, W2, and servo head S. The helical scanning system 30 Additionally, there is a control microprocessor 100 which primarily communicates with the mailbox 102, a 5C8I interface 104, a data buffer manager 106, an encoder/formater 108, a decoder/default formatter 110, and an AUXE CC program generator. 111. The data buffer manager 106 stores data including one megabyte of DRAM. It has a data buffer, is 9-bit wide, dual-ported, and is organized as circular memory. Data Noku Sofa-Manager 106 and 5C8I Interface 104, encoder/former 108, decoder/deformer 110, and Data transfer between the AUXECC processor 111 and the AUXECC processor 111 is performed asynchronously or synchronously. The logical user data block is a data block for tape recording. The file is formatted into a physical program by running the file 1061. The AtJXECC block generator 111 generates an auxiliary error correction program to read the error correction strips from the tape 32. The structure and operation of the AUX ECC block generator are described here. Encoder/formater 108 receives data processors from data node manager 106 and AUXECC block generator 111. The encoder/former 108 adds error correction code (ECC) information and performs synchronization. Many functions including inserting a marker, inserting a search field, and inserting a servo field. It performs various functions, and performs inter-node ordering of nodes. The encoder/former 1081 sends the data processor and the attached information to the RLL modulator 112 to convert each 8-bit word into a 10-bit word. Executes the execution length code write of the data stream. The 10-bit word is sent to bit serializer 1141. Bi Sotong The realizer 114 connects a write drive circuit 116 (for write head W1) and a write drive circuit (for write head W2) through a FIFO register 118. It is connected to the dynamic circuit 120. The function of FIFO register 118 is to This will be discussed later in connection with a description of the write operation of stem 30. Read heads R1 and R2 are coupled to preamplifiers 130 and 132, respectively, to amplify the read signal. Preamplifiers 130 and 132 are coupled to signal conditioning circuits 134 and 136, respectively. Signal conditioning circuits 134 and 136 include circuits for amplitude sensing, equalization, and data timing and detection. The signal conditioning circuit 134 is coupled to a FIFO register 138. The register is connected to the serial to parallel controller at 140. The signal conditioning circuit 136 is directly coupled to the serial to serial controller 140. A serial-to-parallel converter 140 is coupled to an RLL demodulator 142 and a no<9-no detection circuit 143. The RLL read modulator essentially performs the inverse operation of the corresponding RLL write modulator 112. A pattern detector 143 analyzes the incoming data stream to recognize synchronization fields. monitor the stream. When the pattern detector 143 recognizes a predetermined number of synchronization fields at predetermined intervals, the pattern detector 143 generates a BKRDY signal. Furthermore, the pattern detector 143 serially transmits a synchronization signal necessary for the operation of the controller 140. 140 is supplied to the parallel controller. RLL demodulator 142 is coupled to decoder/deformer 110. A decoder/deformator 110 assembles data blocks and performs error correction. Microprocessor 100 communicates with servo microprocessor 150 and motion control system 152 through mailbox 102. The motion control system 152 includes a dedicated master for communicating with a drum servo 156, a capstan servo 158, a reel control circuit 160, and a mechanical controller 162. Including microprocessor. Additionally, upon receipt of the B KRDY signal generated by pattern detector 143, motion control system 152 includes circuitry for generating a HEAD 5YNC signal, referred to herein as HEAD 5Y NC generator 1.64. HEAD 5YNC generator 164 is described in U.S. patent application Ser. HEAD 5YNC signal is read The signal is high when heads R1 and R2 are on the group strip of tracks 1 and 2. The HEAD 5YNC signal is low when write heads W1 and W2 are over the track 1 and track 2 strips. The motion control system 152 also includes a drum tachometer, a caps tank tachometer, The sensor interface circuit communicates with sensor interface circuits for various elements not shown, including a reel tachometer, a reel tachometer, an end of tape (EOT) detector, and a start of tape (BOT) detector. As shown in FIG. 4, the tape transport system 34 is coupled to a drum servo 156, a capstan servo 158, a reel control 160, and a mechanical controller 162. Servo head S couples its output signal to preamplifier 172 . The output of the servo preamplifier 172 filters the amplified signal and outputs the servo preamplifier 172. is applied to a filter and detection circuit 174 that detects the signal. Filter and detection circuit 174 is a servo control included in motion control system 152. It is connected to racking circuit 175. Details of the servo tracking circuit 175 are shown in U.S. patent application Ser. Tape Format FIG. 5 depicts the format of magnetic tape 32 for helical scanning system 30. The tape 32 has a translucent leader material attached to the magnetic medium (=i) and is positioned at a point i. It has a physical beginning of tape (PBOT) 300. (The text indicated by arrow 78 (in the tape transport direction) downstream of the PBOT 300. There is a heavy spiral strip. Of course, the spiral strips contain information that is written by write heads W1 and W2 and read by read heads R1 and R2. including information. The format of the helical strip is discussed subsequently in connection with FIGS. 6 and 7. At the end of tape 32, there is a physical end of tape (PEOT) 302. Because the helical scanning system 30 is a dual azimuth system, it cannot be placed in writing mode. , the odd numbered spiral strips are written in the first azimuth direction by the write head W1. The even numbered helical strips are written at the second azimuth angle A1, and the even numbered spiral strips is written at angle A2. In read and read back modes, odd numbers The helical strip of No. is written at a first azimuth angle A1 by the writing head W1 and the even numbered spiral strips are read out according to the second azimuth angle A2. is read out. As identified above, the first azimuth angle A1 is +20 degrees. 2nd Azima The angle A2 is -10 degrees. As used hereinafter, the write head W1 The spiral strip written or read by read head R1 is referred to as "Track 1". Similarly, writing is performed by writing head W2. Once loaded or read out by the read head R2, the rotation strip is 2). For purposes of this invention, both track 1 and track 2 are physical data blocks. The data block 317 includes a plurality of data blocks 317 known as blocks. In the illustrated embodiment Both track 1 and track 2 contain eight data blocks 317. Other aspects of the track 1 and track 2 formats are understood with reference to U.S. patent application Ser. FIG. 8 shows the format of data blocks 317 provided on tracks 1 and 2. Showing the mat. The format of data block 317 is the same regardless of whether data block 317 is written to track 1 or track 2. As shown in FIG. 8, each data block 317 (known as an "rphysical data block") includes a physical block header 350 (14 bytes), a data area 352 (1024 bytes), and an error correction code (rEccJ) area 354. (400 bytes), and a cyclic redundancy check (rcRcJ') area 356 (2 bytes). FIG. 9 shows the format of the physical block header 350 of each physical data block 317. Draw the mat. FIG. 9 shows the hexadecimal system for physical block header 350. Header 350 indicates the format of the information contained in block 317, e.g. used to determine the format of the file and the manner in which the information is stored. As mentioned above, header 350 is 14 bytes. Lower 4 bits of byte O of header 350 The block includes a value (identifier BLOCK TYPE) indicating the type of block 317. The following hexadecimal values relate to different block formats. That is, rO HJ represents user data. rAHJ represents a file mark. "C H" represents a logical start of tape (LBOT) write with retry. rD HJ represents LBOT write without retry. rEHJ expresses the gap. And rFHJ represents the end of data. As mentioned above, the BLOCK TYPE value of roHJ means that the physical block 317 contains user data. Such block 317 is a user data database. Also referred to as a lock. The user data block is a physical block containing up to 1024 bytes of user data in the data area 352. Of interest to the present invention is the fact that physical block header 350 includes an AUXECC field and an IN 1 ECC field. In this regard, the AUXECC field The code (byte 0, bit 7) indicates that the information contained in the data area of 8 consecutive blocks includes auxiliary FCC data for the preceding 128 data blocks. The INIECC field (byte O, bit 6) indicates that eight consecutive blocks The first 8 blocks of a set of 128 blocks of 8 auxiliary FCC data written to the Indicates that it is Tsuku. The WRTRTY field (byte 0, bit 5) indicates whether this block has been previously written. Structure: AUXECC Circuit AUXECC circuit 111 includes AUXECC RAM 702 used to store data for generating AUXECC 7° trips on magnetic tape 32. FIG. 10 shows a segment of tape 32 and the recorded 16 spiral strips. 704, particularly the group G of strips 704o-705, fi. Regarding the configuration of strip group G, strips 7040-70414 are The management data block (described above) is a strip containing user information data and error correction information. It is a pool. Strip 704+a is an AUXECC strip produced in the manner described below, providing an additional layer or dimension of error correction capability. child In this respect, the AUXECC strip 704+s is configured such that conventional error correction circuitry provided in the decoder/default formatter 110 decodes one or more blocks. It is used for error correction purposes when this is not possible. FIG. 11 shows a plurality of strip groups G1, ie, group G. and group G n +1. last strip strips 704+s of each group G are connected to the AUXECC strips 704+s. strip, and for illustrative purposes is shown somewhat darker than the other strips in the group. Thus, multiple consecutive groups G are set up on tape 32. You can see that you are kicked. Returning to the structure of AUXECC circuit 111 described in FIG. 3, AUXECC RAM 702 is sized to accommodate eight blocks. In addition to the AUXECCRAM 70 2, the AUXECC circuit 111 also includes an address generator 710 and a controller. 712, an address multiplexer (MUX) 71.4, an XOR circuit represented by an XOR gate 716, and a data transmission MUX 718. The stream of 8-bit data is routed to the line designated DATA-IN in Figure 3. and enters the AUXECC circuit 111. In the write mode, the data on line DATA-IN is The data is received from the buffer manager 106 (see FIG. 4). In read mode, the data on line DATA-IN is a stream of data received from RLL demodulator 1 42. The 8-bit data on line DATA-IN is coupled to input terminal 0 of MUX 718 and XOR circuit 716. Although not explicitly shown as such, XOR gate 716 represents eight XOR gates, one gate for each bit of information maintained on line DATA-IN. Similarly, MUX718 has 8 multiplexes. Represents a bank of lexers, one for each bit of information on line DATA-IN. It is maintained at The output terminal of the XOR circuit 716 is coupled to the data input bin of the AUXECCRAM 702 (8-'). The eight output bins of AUXECCRAM 702 are coupled to input terminals l of MU X718. Furthermore, the eight output pins of AUXECCRAM 702 are coupled to the input terminals of eight respective AND gates represented by AND gates 726, whose output terminals are connected to the input terminals of eight respective AND gates represented by AND gates 726. are connected to each XOR gate. Address generator 710 generates addresses for application to AUXECCR AM 702 via address MUX 714 . Address generator 710 includes byte counter 730 . It includes an OR gate 731, a block counter 732, a strip counter 734, and an adder 736. AUXECC circuit 111 is notified via line BYTE CLOCK when a new byte is applied on line DATA-IN. In this regard, circuitry is provided to pulse the line BYTE CLOCK for each transmission of a bit carried on the lines DATA, -IN. The BYTE CLOCK signal is the count enable bit of the byte counter 730. applied to the The data output bin of byte counter 730 is tied to terminal 0 of address MUX 714. As can be seen, the address MUX 714 actually represent multiple multiplexers corresponding to the number of bits in the address. The carry output of byte counter 730 is coupled to the count enable bin of block counter 732 through OR gate 731. block counter -732 couples the three data output pins to the first port of adder 736. The carry output bin of block counter 732 is coupled to the enable count bin of strip counter 734. Strip counter 734 couples the three lowest data output bins to the second port of adder 736 and logic controller 712. The three lowest output bins of the adder are connected by lines 437 to the three highest input bins at terminal 0 of address MUX 714. FIG. 3 also shows an intra-block error correction circuit 740 and a decoder/default formatter. 110, but is shown in FIG. 3 for convenience. error correction episode FCC 740 is coupled to apply a signal to the select bin of MUX 714 and logic controller 712 . FCC circuit 740 is also coupled to apply a BLOCK COUNT INCREMENT pulse on line 741 to OR gate 731 whenever a block of data cannot be decoded. FCC circuit 40 is also coupled to MUX 714 for applying the address of the error block to terminal 1 of MUX 714. Additionally, FCC circuit 740 connects line 743 to and is connected to receive the BAD BLOCK ADDRESS value. The select pin of data transmit MUX 718 is coupled to the output of logic controller 712 by line 742. Logic controller 712 connects one of terminals 0 or 1 to MUX 718 to receive data. Logic controller 712 also has an output terminal coupled to a second terminal of AND gate circuit 726 by line 724. Operation: AUXECC Encoding As mentioned above, in spiral scan technology there are two types of errors resulting from two types of operation. The first type of error destroys multiple blocks in the strip. Ru. The second type of error refers to blocks at the same location in consecutive strips (all In other words, the blocks (blocks aligned in the tape movement Hoko) are destroyed. As discussed above with reference to FIGS. 10 and 11, to overcome the problem of lost blocks and strips and rows of blocks, the AUXECC block generator of the present invention The generator 111 sends an AUXECC block to all group G strips. A strip 704+s of blocks is generated. AUXECC strip 704+s is physically the last strip in the group. AUXECC block E included in AUXECC strip 704+s. is obtained by performing an exclusive or ("XOR") operation with other blocks in subgroup i of group G. Each AUXECC plot El is Members of loop i and other members of subgroup 1 are specifically chosen in the manner described below to allow recovery of all strips and rows. Describes performing exclusive or (XOR) operations with other blocks of a subgroup and each bit of a block in a subgroup is This means that it is XORed with the corresponding bit in the lock. FIG. 10 shows each strip 704 divided into eight blocks. Each block is generally B11. It is expressed as closing. Each block belongs to a row rrJ (running parallel to the direction of tape movement indicated by arrow 46) and a column or strip rcJ. For example, block B2. shown in FIG. . is the block in row 2, strip O, with the general block name B2. . is actually shown as block C2,0 with subscripts r, c. A subgroup formulation and arrangement for the illustrated embodiment of the helical scanning system 30 is shown in FIG. Instead of a general block name, each block is labeled with the name it belongs to. Alphabetic letters A, P, C, D, U to indicate subgroups. Indicated as having one of F, G or H (letters P and U are used to avoid confusion with general identifiers). For subgroup alphabets, blocks belong to group G. A subscript is added to indicate the row and strip number in which it is physically located. With the organization of FIG. 10, subgroup i=A has blocks A0. . , A7,1, Ao, 2, A6,3, A4,4, A316, A2.6, A1,7, Ao, 8, A711, A$+ 10SAL II, A4. +2, A3. +3, A2.1 4 and A1. Consists of +5. As another example, subgroup i=U is U4. O, U3. 1, and u, , +5. Thus, the subgroup formulation and arrangement for the helical scanning system 30 of the illustrated embodiment is similar to the block skin 4. Obtained by shift selection. Shift selection is specifically implemented by an atnosier generator 710 included in the AUXECC circuit 111, as described further below. Shift selection of blocks contained within allows recovery of entire rows and columns. A general explanation of subgroup formulation relationships is given. Each group G has rnJ number of strips. Suppose that it consists of groups or columns and rkJ number of rows. In addition, the AUXECC block of rmJ number Suppose that a block E is generated for a group G. In this case m=qxk (ie, multiply by qi::k), where q is an integer. In the user data block Each bit b1 in the error block E1 corresponding to the user data is logically XORed with the corresponding bit of the user data block included in block group i. Due to the above formulation relationship, each bit position included in each block E1 Place b! The value of is a bit with b bit positions in all blocks B, , c. is derived from the logical XOR of the Here [Equation 1] i= (r+k (c mod q)+INT (c/q) m. m In this case INT(c/Q) is the largest integer not exceeding c/q. Thus, given Solving Equation 1 for i for block BI+e, the result for that block is generate a group classification l. General Equation 1 uses k (number of rows per strip) = 8, n (number of strips per group G) = 16, and m (number of AUXECC blocks) - 8 to calculate the AUXECC circuit of the illustrated embodiment. This is realized by 111. Address generator 710 specifically implements the block knot mechanism shown in Equation 1 and FIG. During a recording operation, data to be encoded and recorded on tape is provided from data buffer manager 106 to AUXECC circuit 111 on line DATA-IN. be provided. As each byte is applied on line DATA-IN, the BYTE CLOCK signal is pulsed so that byte counter 730 counts the number of bytes. Byte counter 730 counts up to the maximum number of bytes in the block (1048). The count of counter 730 is applied to terminal O of address MUX 714 as the 10 least significant bits of the address. When byte counter 730 exceeds its maximum counting capacity, as occurs when all bytes of a block are counted, the carry signal of byte counter 730 is The number is used to increment block counter 732. block counter The count of adder 732 is applied to adder 736. When block counter 732 reaches a maximum count of 8 (the number of blocks in the strip), a block counter carry signal is applied to strip counter 734. Using the carry signal from the block counter, strip counter 734 counts the number of strips of data passing through AUXECC circuit 111 for group G. The contents of strip counter 734 are applied to adder 736 and logic controller 712. Adder 736 adds block counter 732 and strip counter 734 to block counter 732 and strip counter 734. add the two numbers represented by the values in the module, and 8) functions to execute the operation. A modulo 8 operation is performed by taking the three least significant bits of the sum. The three least significant bits of the sum obtained by adder 736 are applied to terminal O of address multiplexer 71,4 as the most significant bit of the address. The first 15 steps of incoming data for each group G (on line DATA-IN) In response to a trip, logic controller 712 uses the data transmit MUX 71.8. The user data is passed to an encoder/formater 108 where, among other things, block header information is added. For these first 15 strips, each incoming bar The bit also has corresponding other blocks for other blocks in the same subgroup i. (ie, bytes with the same byte position). Control logic 712, which has access to the contents of strip counter 734, causes MUX 718 to encode/former the user data for the first 15 strips of group G by applying a signal representing MUX 718 terminal 0 on line 742. Have Matsuta 108 send it. As described above, each byte of user data is XORed with the corresponding other byte. The XOR operation is accomplished using XOR circuit 716. The bits of the incoming byte of user data are summed with the corresponding bits in the same subgroup i. and the total is stored in AUXECCRAM 702. The total has the contents of the appropriate byte sum in AUXECCRAM 702 as human power, and as second human power. This is accomplished using an XOR circuit 716 with an incoming data byte. "1" again is the one-bit sum of digital bits that have one of the “0” values. It will be understood by those skilled in the art that the logical XOR operation performed on the Address generator 710 determines with which byte in AUXECCRAM 702 the incoming data byte is summed to implement the subblock selection arrangement of the invention. Address generator 710 performs a modulation in adder 736 with respect to counts in block counter 732 and strip counter 734. This is done by performing a Juro 8 operation. Address generator 710 essentially adds the incoming data to the AUXECC strip where it is summed according to Equation 1. Select the byte address in The operations of address generator 710 result in the subgroup classification shown in FIG. After 15 strips of incoming data have been received, logic controller 712 A signal is sent on line 742 to instruct the data transmitter MUX 718 to send the contents of AUXECCRAM 720 to the encoder/formater 1.08. child At this point, the contents of AUXECCRAM 702 are AUXECC strips 704+s. After the contents of AUXECCRAM 702 are sent, the encoder/file Formatter 108 receives the used data that has been similarly processed by encoder/formater 118. A header 350 and error correction area 354 (see FIG. 8) are provided for each AUXECC block contained in AUXECC strip 704+s as was done for each of the user data blocks. To prepare the header 317 of the AUXECC block, the logic controller sends a signal to the record header generator in encoder/formatter 118 that causes encoder/formatter 118 to set the AUXE CC bit to header 350. enable. As shown in FIG. 9, the AUXECC hit occurs at byte O, p(.7) of the AUXECC record header 350. The error correction information produced by formatter 108 is due to the particular intra-block error correction polynomial that the helical scanning system operates on. Although an intra-lock FCC technique has been described, it should be understood that the choice of intra-block FCC technique does not affect the intra-block error correction of the present invention. Whenever the encoding of a new group G of strips begins, the first strip of the group For strips (i.e., strip “0”), logic controller 712 sends that strip to the block ladder generator in encoder/formater 108. block is on the first strip following the AUXECC strip. know At this point, logical controller 712 blocks the INIECC bit. A signal is generated on line 744 that is used by the header generator to set lock header 350. INIECC bit (byte 01 bit Point 6) is shown in FIG. Additionally, the signal on line 744 from logic controller 712 is connected to group G. is used by AND gate 726 to reset the contents of AUXECCRAM 702 for the first strip in the memory. When the signal on line 744 turns off AND gate 726 for the first strip, the data on line DATA-IN goes directly to AUXECCRA, M2O3. I don't go to meet you. It is clear that for systems that have the ability to rewrite bad blocks to tape, the AUXECC block generation element described above must be suppressed when the bad blocks to be rewritten are sent on the DATA-IN line. be. The above discussion of write mode indicates that the address MUX 714 It is assumed that the address generated by the controller 710 is sent to the AUXECCRAM 702. Otherwise, only in the read mode described below and when the intrablock FCC circuit 740 is unable to decode the block. Operation: Decryption During read or decryption mode, incoming data is ultimately from a readout circuit other than jar 106. All incoming data during read mode is first applied to an intrablock ECC circuit 740, which decodes each block using, for example, a conventional error correction code or product code. Incoming data is then applied on line DATA-IN for application to AUXECC circuit 111. When incoming data enters AUXECC circuit 111 on line DATA-IN, each For group i, bit b8 is summed in AUXECCRAM 702. It will be done. The summation and addressing operations are essentially the same as those described for write operations, except for the fact that corresponding bits of bytes in all 16 strips of group G, including AUXECC strip 704+5, are summed. Assuming no read errors occur, at the end of the 16th strip of the group, i.e. the read of AUXECC strip 70'Ls of the group. After reading, the contents of AUXECCRAM702 are all zeros, and there is no error. reflect the new readout. When the intra-block FCC circuit 740 is unable to decode an error block, the The address of the good block is determined by ECC circuit 740. This point smells Then, the address of the bad block is determined by performing a modulo 8 operation by the ECC circuit 740, as is done by the address generator 710. Locks are enabled in conjunction with the inventive sub-block arrangement. This defective block address is applied to terminal 1 of address MUX 714. At the same time, ECC circuit 740 pulses line 741 and increments block counter 732 to compensate for the omission of bad blocks from DATA-IN in the address calculation. After attempting to decode all groups G, ECC circuit 740 sends an ERRORDETECT signal to address MUX 714. Further, the ECC circuit 740 sends the defective block address prepared as described above to the address MUX 714. The ERRORDETECT signal applied to the selected bin of the ATOINOS MUX 714 preempts the normal operation of the AUXECC circuit 111. Related to this preemption, The address of the good block is printed in AUXECCRAM702 via MUX714. added. Using the bad block address, the MUX selects the block in AUXECCRAM 702 that corresponds to subgroup i to which the bad block belongs. Ru. At this point, the contents of the selected block in AUXECCRAM 702, excluding the bad block, are the sum of the remaining blocks in subgroup 1 to which the bad block belongs. The sum is the bad blocks and the selected block address. This means that the contents of the AUXECCRAM 702 for the response are a corrected defective block. The preceding discussion shows that exactly one block correction in group G. However, it should be understood that the AUXECC circuit 111 of the present invention can be used to reconfigure multiple blocks. In this regard, the intra-block FCC times For any group G that the path 740 cannot decode, After attempting to decode each block, the steps described above are corrected by the FCC circuit 740. This is done for each block found to be invalid. That is, the ECC circuit provides a bad block address to address MUX 714 for each bad block while operation of the input data stream is temporarily preempted. Thus, a plurality of full blocks are reproduced according to the invention. For this purpose, strategic classification of blocks into relevant subgroups and The XOR of the corresponding bits allows reconstruction of all strips of m adjacent blocks in a recording block or row. Although one particular embodiment of the subgroup classification of the present invention has been described, other applications for the classification are possible in other embodiments. For example, FIG. In this case, m=18, n=20, and n=6. In FIG. 13, AU XECC blocks are provided in strips 17, 18 and 19. Although the invention has been shown and described in detail with reference to preferred embodiments, the invention has not been construed in form or detail. It will be understood by those skilled in the art that various modifications may be made without departing from the spirit and scope of the invention. FIG. 9 Copy and translation of written amendment) Submission (Article 184-8 of the Patent Law) dated May 6, 1992

Claims (17)

[Claims] 1. In a method for encoding information data transmitted to a storage medium, the block Br , c, and format the information data into at least some of the groups of Br,c belongs to row r and column c, r is in the range from 0 to k-1, and c is not 0. n-1, and each block Br,c contains bit position bx, where x must be in the range from integer Y to integer Z, and must be included in the group of blocks Br, c. m number of error correction blocks Ei are generated, where i is in the range of 0 to m-1. m = qxk, q is an integer constant, and the error correction block Ei each also includes a bit position bx, where x ranges from an integer Y to an integer Z. In this case, the value of each pit position bx included in each block Ei is is derived from the sum of bits with bit position bx in block Br,c, and this case i = (r + k (c mod q) + INT (c/q) mod m where INT ( c/q) is the largest integer not exceeding (c/q). 2. In a method for encoding information data transmitted to a storage medium, the block Br , c, and format the information data into at least some of the groups of Br,c belongs to row r and column c, r is in the range from 0 to k-1, and c is not 0. n-1, and each block Br,c contains bit position bx, where x is in the range from integer Y to integer Z, and the blocks Br, c are subgrouped. and error correction for m numbers included in the group of blocks Br,c. Blocks Ei are generated, each block Ei is included in one of the subgroups, and each block Ei is included in one of the subgroups. A block group includes one of the error correction blocks Ei, where i is from 0 to m- 1, m = qxk, q is an integer constant, and the error correction block Each of the blocks Ei also includes a bit position bx, where x is in the range of integers Y to Z. In this case, the value of each bit position bx included in each block Ei is bit position bx for all blocks Br, c included in the block group. and performing an exclusive OR operation on the values of the bits. 3. The m number of error correction blocks Ei are blocks Br recorded on the storage medium. , c, which is the last block in the group. Law. 4. The storage medium is a magnetic tape, and the blocks Br, c are recorded on the storage medium. further comprising: recording the same row in the spiral strip; Recording all blocks Br,c with m/q numbers, followed by and recording the error correction block Ei in a lip. The method described in. 5. An indication that the block is an error correction block is given to the error correction block Ei. 5. The method according to claim 4, further comprising providing at. 6. Error correction bit included in the additional bit position of the block Br, c To obtain bit position bx (where x is in the range of integer Y to integer Z) Claims 1 or 2 further comprising the use of a code word to affect Method described. 7. The subgroup is such that blocks in the same column are not in the same subgroup. organized, where the subgroup is a sequence of m adjacent blocks in the same row. 3. The method of claim 2, wherein the instances are organized in such a way that they are not in uniform subgroups. 8. In an apparatus for encoding information data to be transmitted to a storage medium, a block Br , c. each block Br,c belongs to row r and column c, where r ranges from 0 to k-1. , c ranges from 0 to n-1, and each block Br,c has bit position b x, where x ranges from an integer Y to an integer Z; Generate m number of error correction blocks Ei included in the group of blocks Br and c Here, i is in the range from 0 to m-1, and m = qxk. , q is an integer constant, and each error correction block Ei also has a bit position bx, where x ranges from an integer Y to an integer Z, in which case each block The value of each bit position bx included in Ei is a bit value in all blocks Br, c. is derived from the sum of bits with bit position bx, in this case i = (r + k (c mod q) + INT (c/q) mod m where INT ( c/q) is the largest integer not exceeding (c/q). 9. In an apparatus for encoding information data to be transmitted to a storage medium, a block Br , c. each block Br,c belongs to row r and column c, where r ranges from 0 to k-1. , c ranges from 0 to n-1, and each block Br,c has bit position b x, where x ranges from an integer Y to an integer Z; means for organizing the blocks Br,c into subgroups; means for generating m number of error correction blocks Ei included in the group of Yes, each block Eif is included in one of the subgroups, and each subgroup has the error - one of the correction blocks Ei, where i ranges from 0 to m-1 and m = qxk, qf is an integer constant, and each error correction block Ei is also bit position bx, where x ranges from an integer Y to an integer Z, in which case each The value of each bit position bx included in block Ei is included in the same subgroup. exclusive to the value of the bit with bit position bx for all blocks Br,c and means derived by performing an OR operation. 10. further comprising means for recording said block Br,c on said storage medium; In this case, the m number of error correction blocks Ei are the blocks B recorded on the storage medium. The last block in the group of r, c according to claim 8 or 9. Device. 11. The error correction pitch contained in the additional bit position of the block Br, c bit position bx (where x is an integer between Y and Z). Claim 8 further comprising means for using the code word to affect or the device according to 9. 12. The subgroup has the same sequence of m adjacent blocks in the same row. in the same row, in which case the subgroup is Claims organized such that sequences of adjacent blocks are not in the same subgroup 9. The device according to 9. 13. In a device that encodes information data sent to magnetic tape, blocks are for formatting the information data into at least some of the groups of Br,c; each block Br,c belongs to row r and column c, where r is 0 to k-1. range, c is a range from 0 to n-1, and each block Br,c is a bit position. means including a location bx, where x is in the range of an integer Y to an integer Z; Generate m number of error correction blocks Ei included in the group of blocks Br and c Here, i is in the range from 0 to m-1, and m = qxk. , q is an integer constant, and each error correction block Ei also has a bit position bx, where x ranges from an integer Y to an integer Z, in which case each block The value of each bit position bx included in Ei is a bit value in all blocks Br, c. is derived from the sum of bits with bit position bx, in this case i = (r + k (c mod q) + INT (c/q) mod m where INT ( c/q) is the largest integer not exceeding (c/q), and k on the magnetic tape means for recording all blocks Br,c in a number of spiral strips. Yes, all blocks Br,c in the same column are identical to each other on the magnetic tape. The error correction block Ei is recorded in n number of strips. means for recording as a strip of the last m/q number of. 14. In a method of decoding information data obtained from a storage medium, a group G A plurality of blocks Br,c belonging to Br,c are acquired from the storage medium, in which case each block B, belongs to row r, column so, r is in the range from 0 to k-1, and c is not 0 n-1 range, and each block Br, soc contains bit position bx, where and x is in the range from integer Y to integer Z, and Blocks Br,c and a plurality of m-1 subgroups that are correctly read from the storage medium For each subgroup i, the bit positions of blocks Br,c belonging to each subgroup i bx, thereby obtaining the sum of each of the bit positions bx, where i is 0 m-1, m = qxk, where q is an integer constant, in which case , i = (r + k (c mod q) + INT (c/q) mod m where INT ( c/q) is the largest integer not exceeding (c/q), and In rare cases, the bit position bx of block Br,c that cannot be read from the storage medium is using the sum for subgroup i to derive. 15. In a method of decoding information data obtained from a storage medium, a group G A plurality of blocks Bs, c belonging to the storage medium are acquired from the storage medium, and in this case, each block Br,c belongs to row r and column c, r is in the range from 0 to k-1, and c is not 0 n-1, and each block Br,c contains bit position bx, where x is in the range of an integer Y to a distortion number Z, and Blocks Br,c and multiple m-1 subgroups that are correctly read from the storage medium For each subgroup i, the bit positions of blocks Br,c belonging to each subgroup i bx, thereby obtaining a sum for each of the bit positions bx, which sums are equal Bit position bx for all blocks Br, c included in one subgroup i , where i is 0 m-1, m = qxk, where q is an integer constant, in which case The blocks Br and c included in each subgroup i have a predetermined organization and of block Br,c included in block group i and which cannot be read from the storage medium. Using the sum for subgroup i to derive bit position bx and a method including. 16. In an apparatus for decoding information data obtained from a storage medium, a group G means for acquiring a plurality of blocks Br,c belonging to from the storage medium, Each block Br, c belongs to row r and column c, where r is in the range from 0 to k-1. , c ranges from 0 to n-1, and each block Br,c has a bit position bx, where x is in the range of an integer Y to an integer Z; Blocks Br,c and a plurality of m-1 subgroups that are correctly read from the storage medium For each subgroup i, the bit positions of blocks Br,c belonging to each subgroup i By means of summing bx and thereby obtaining the sum of each of the bit positions bx , where i is in the range 0 to m-1, m = qxk, and q is an integer constant. In this case, i=(r+k(c mod q)+INT(c/q)mo d m where INT(c/q) is the largest integer not exceeding (c/q); a block Br included in the subgroup i and not readable from the storage medium; Use the sum for subgroup i to derive bit position bx of c and means for. 17. In an apparatus for decoding information data obtained from a storage medium, a group G means for acquiring a plurality of blocks Br,c belonging to from the storage medium, In this case, each block Br,c belongs to row r and column c, where r ranges from 0 to k-1. , c ranges from 0 to n-1, and each block Br,c has a bit position bx, where x is in the range of an integer Y to an integer Z; Blocks Br,c and a plurality of m-1 subgroups that are correctly read from the storage medium For each subgroup i, the bit positions of blocks Br,c belonging to each subgroup i bx, thereby obtaining the sum of each of the bit positions bx, thereby obtaining the bit means for obtaining a sum for each of the cut positions bx, said sum being Set the pit position bx for all blocks Br, c included in subgroup i. is obtained by performing an exclusive OR operation on the values of the bits that have m−1 range, m = qxk, where q is an integer constant, in which case each Blocks Br and c included in subgroup i are arranged by means of a predetermined organization and Bits of blocks Br, c included in group i and not readable from the storage medium to use the sum for subgroup i to derive the cut position bx. and means.