JPH09326704A - Coding and decoding method for mutli-value recording - Google Patents

Abstract

PROBLEM TO BE SOLVED: To obtain a multi-value record coding method suitable for a recording and reproducing method where a mutli-value signal is reproduced depending on the amplitude of a reproduction signal by coded multi-value signals under the restriction depending on a minimum length and a maximum lenght of the same consecutive mutli-value codes. SOLUTION: A data signal 53 is given to a 4-value record coder 54. The coder 54 encodes an input data series so that a relation between a minimum length dmlt of same consecutive multi-value codes and a maximum length Kmlt of the same consecutive multi-value codes is expressed as Kmlt /dmlt >1 and a value Kmlt /dmlt is a definite value. The 4-value code obtained by the coder 54 is given to an external magnetic field application circuit 55 to drive a magnetic head 52. According to the coding method, in the recording and reproduction to/from a mutli-value recording medium where a mutli-value signal is reproduced depending on the amplitude of the reproduction signal, the mutli-value record coding that takes interference between the multi-value codes and clock synchronization into account is attained. Furthermore, the coding method is applicable to the recording and reproducing method for a magneto-optical recording medium or a magnetic recording medium.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an encoding method for performing multi-valued recording using a magneto-optical recording medium or a magnetic recording medium, and more particularly to a multi-valued signal according to the strength of an external magnetic field for recording. The present invention relates to a multilevel recording encoding suitable for recording / reproduction of a magneto-optical recording medium or a magnetic recording medium in which is recorded and a multilevel signal is reproduced according to the amplitude of a reproduction signal, and a decoding method thereof.

[0002]

2. Description of the Related Art Multi-level recording has attracted attention in order to enable higher density recording on recording media such as magnetic disks and magneto-optical disks. For example, the applicant's international publication No. 8-129784, Shimazaki, Yoshihiro et al. "Magnetic mu
ltivalued magneto-optical disk "Magneto-Optical Re
cording Int. Symp. 1994, Post Dead Line Paper Tech
nical Digest, No.27-S-01, p4, 1994; Optical Data
In Storage, Technical Digest, pp. 59-60, 1994, a multilevel signal is recorded according to the strength of the external magnetic field applied to the magneto-optical recording medium, and the recorded multilevel signal is set to the amplitude of the reproduction signal. A magneto-optical recording medium that can be reproduced according to the above and a recording / reproducing method thereof are described. When performing multilevel recording and reproduction using such a magneto-optical recording medium, it is important to select a data signal modulation system or coding system in order to improve the SN ratio.

By the way, two-dimensional encoding has been attracting attention as an encoding method that can be expected to achieve high density in magnetic recording and magneto-optical recording. Such two-dimensional coding is based on MW Marc
ellin and HJWeber “Two-Dimensional Modulation C
odes ”as IEEE JOURNALON SELECTED AREAS IN COMM
First reported in UNICATIONS, vol.10, No.1 (January 1992). Swanson and Wolf are IEEE TRANSACTIONS ON
MAGNETICS. VOL. 28. No. 6 (November 1992), (d
_x, k _y, k _y) by using the two-dimensional code in consideration of the constraints, not only the improvement of the clock function, have reported that the drop-out error occurrence is improved. Here, d _x is the minimum run length in one of the multiple channels, and if d _x is too small, intersymbol interference is likely to occur in the reproduced signal. k _x is the maximum run length in one of the multiple channels, and if k _x is too large, the clock signal required for synchronization cannot be stably extracted. k _y (k _y <k _x ) is Sw
This is a new constraint introduced by anson and Wolf, which is a k constraint when multiple channels are integrated.

In addition, Tasaki et al.
"Characteristics of two-dimensional codes and code examples" are reported on pages 31-37 of 995-05). According to the method in the document, the basic state transition matrix T _L corresponding to multiple L channels is used.
After deriving, the matrix D ₂ is obtained by excluding unnecessary rows and columns considering the k _y constraint. After that, D ₂ is raised to the nth power to obtain the number of effective candidate words, and the state set from which the codeword group having the maximum minimum Hamming distance can be selected is extracted. Then, the trellis diagram used in Viterbi decoding at the time of reproduction is drawn by using a set of states in which the codeword group having the maximum minimum Hamming distance can be selected.

[0005] said _{(d x, k x, k} y) two-dimensional code in consideration of the constraints, be applied to a recording and reproducing method capable of interpreting channel for recording and reproducing in the binary code as a multi-channel and a plurality integral Is considered to be effective. As a recording / reproducing method in which a plurality of channels for recording / reproducing with a binary code can be integrated and interpreted as a multi-channel, for example, recording / reproducing on a magnetic tape using a multi-head can be mentioned.

[0006]

The above problem (d _x , k _x ,
The k _y) two-dimensional code in consideration of the constraints, multi-value recording medium, such as the magnitude of the amplitude of the reproduced signal representing a multi-level code, for example, the recording and reproducing to the multilevel recording magneto-optical recording medium It can be applied. However, the recording / reproducing on such a multi-level recording medium cannot be simply interpreted as a multi-channel in which a plurality of channels for recording / reproducing with binary codes are integrated for the reason described below, and (d _x , k _x , _ky ) The constraint is not always suitable for the characteristics of the multi-valued recording medium in which the amplitude of the reproduced signal expresses the multi-valued code. For example, 4
It is assumed that a data sequence of 0 → 2 → 0 → 2 → 0 → 2 → 0 → 2 is recorded on the recording medium at the time of value recording. 0 → 2 → 0 → 2
→ 0 → 2 → 0 → 2 in binary notation, 00 → 10 → 0
It becomes 0 → 10 → 00 → 10 → 00 → 10. Here, it is interpreted as a multi-channel in which two channels for recording / reproducing with a binary code are integrated, and the first channel ch1 is made to correspond to the high-order bits in the case where a 4-level signal is represented by a 2-bit binary code. It is assumed that the lower bit is associated with the channel ch2 of. The binary code sequence of ch1 is 0 → 1 → 0 → 1 →
While 0 → 1 → 0 → 1, the binary code sequence of ch2 is 0 → 0 → 0 → 0 → 0 → 0 → 0 → 0. When interpreted as a multi-channel, clock synchronization becomes difficult when the binary code sequence becomes 0 → 0 → 0 → 0 → 0 → 0 → 0 → 0 like ch2. Therefore, the k _x constraint is considered. It turns out that it is necessary. However, in recording / reproducing on a multilevel recording medium in which the amplitude of the reproduced signal directly expresses a multilevel code, the amplitude of the reproduced signal differs depending on 0 → 2 → 0 → 2 → 0 → 2 → 0 → 2. Therefore, there is no problem in clock synchronization. That is, it is meaningless to set the k _x constraint in recording / reproducing on / from a multilevel recording medium in which the amplitude of the reproduction signal expresses a multilevel code.

As can be seen from the above example, (d _x ,
The (k _x , k _y ) constraint cannot be said to be suitable when the amplitude of the reproduced signal directly represents a multi-level code. Therefore, as a coding method for a multilevel recording medium in which the amplitude of a reproduced signal expresses a multilevel code, a new coding method different from the coding method premised on multi-channel recording / reproducing is desired. .

An object of the present invention is to provide an encoding method for multilevel recording and a decoding method thereof suitable for a recording / reproducing method in which a multilevel signal is reproduced according to the amplitude of the reproduction signal. Another object of the present invention is to provide a magneto-optical recording / reproducing method or a magnetic recording / reproducing method suitable for a magneto-optical recording medium or a magnetic recording medium in which a multilevel signal is reproduced according to the magnitude of the amplitude of a reproduced signal. It is in.

[0009]

According to a first aspect of the present invention, a code of a multilevel code used in a multilevel recording / reproducing method in which a multilevel signal is reproduced according to the amplitude of the reproduction signal. In the coding method, the minimum length d _{mlt in} which the same multi-level code is continuous and the maximum length k _{mlt in} which the same multi-level code is continuous are k _mlt / d _mlt
An encoding method for multilevel recording is provided, which is characterized in that an input data sequence is encoded such that> 1 and k _mlt / d _mlt has a finite value.

In the multi-level recording encoding method of the present invention, when the minimum code length in which the same multi-level code continues in the multi-level code sequence is defined as d _mlt and the maximum code length is defined as k _mlt , the ratio thereof is defined. k _mlt / d _mlt is k _mlt / d _mlt > 1 and k
By executing the encoding of the multi-level code under the constraint that the _mlt / d _mlt is finite, it is possible to suppress the inter-code interference of the multi-level code and to achieve proper clock synchronization. d _mlt is an important constraint for controlling intersymbol interference in multilevel recording, and can be set as appropriate in consideration of the recording sensitivity and laser spot diameter of the multilevel recording medium. _kmlt can be set in consideration of clock synchronization when reproducing multi-valued information recorded on a multi-valued recording medium. If _kmlt is too large, clock synchronization becomes difficult. In particular, it is preferable that _kmlt / _dmlt <10 is satisfied in order to suppress intersymbol interference between adjacent multilevel codes and prevent clock synchronization from becoming difficult.

In order to execute encoding in which _kmlt and _dmlt are _kmlt / _dmlt > 1 and _kmlt / _dmlt is finite, for example, a state transition diagram satisfying this constraint (Fig. 1) is used.
(See FIG. 2), the basic state transition matrix corresponding to this state transition diagram is derived (see FIG. 2). Next, the state transition matrix is raised to the power of the code length n. It is possible to extract a state transition group having non-zero elements from the n-th power state transition matrix, and obtain a multi-level codeword group from those transitions.

In the encoding method for multilevel recording of the present invention,
Code so that the maximum length k _mlt is k _mlt / d _mlt> 1 a is and k _{ml t} / d _mlt is finite value the same multi-level code to the minimum length d _{ml t} the same multi-level code to successive continuous It is preferable to select a multivalued code sequence composed of codewords having the largest minimum Euclidean distance from the converted multivalued code sequence. In a multilevel recording / reproducing method in which the magnitude of the amplitude of the reproduced signal directly expresses the multilevel code, the probability that an error occurs in the reproduced signal is not the Hamming distance when the multilevel signal is represented by the binary code, but rather the Hamming distance. This is because it is represented by the Euclidean distance between the value codes.

In the encoding method for multilevel recording of the present invention,
It is preferable to perform convolutional coding on the binary coded sequence input as the input data sequence and then project it to the multilevel coded sequence. At this time, it is preferable that the convolutional coding rate is m / (m + 1) for the length m of the binary code sequence input as the input data sequence.

According to the second aspect of the present invention, the same multilevel code is used in the multilevel code encoding method used in the multilevel recording / reproducing method in which the multilevel code is reproduced according to the amplitude of the reproduction signal. The minimum length d _mlt of continuous codes is the same as the maximum length k _{mlt of} continuous multi-level codes is _kmlt / d _mlt > 1, and k
When decoding the multilevel code encoded by the multilevel recording encoding method, which encodes the input data series so that _mlt / d _mlt has a finite value, (EN) Provided are a multilevel recording coding and a decoding method thereof, which are characterized by performing Viterbi decoding by using a trellis diagram obtained by mapping a code composed of a codeword having a maximum minimum Euclidean distance.

In the multilevel recording encoding and decoding method of the present invention, since the Viterbi decoding is performed using the trellis diagram obtained by mapping the code composed of the codeword having the maximum minimum Euclidean distance, the multilevel recording is performed. Decoding is performed under the error correction function suitable for the code. When calculating the path metric in the Viterbi decoding, it is preferable that the absolute value of the difference between each multi-level code at the time of recording and the multi-level code at the time of reproduction is regarded as the path metric, and the path having the minimum path metric is set as the surviving path. .

In the encoding and decoding method for multilevel recording according to the present invention, the binary code sequence input as the input data sequence is convolutionally encoded and then projected into a multilevel encoded sequence, Minimum length d _mlt where the same multi-valued code continues
The maximum length k _{mlt in} which the same multi-valued code is continuous is k _mlt / d
To encode the input data sequence so that _mlt > 1 and k _mlt / d _mlt has a finite value, and further select a multi-valued code sequence composed of code words having the maximum minimum Euclidean distance. When a multi-level code is encoded by using the trellis diagram obtained by mapping the code composed of the code word having the maximum minimum Euclidean distance, Viterbi decoding is performed, and the multi-level code is When decoding into a value code, it is preferable to perform Viterbi decoding corresponding to the convolutional coding.

The encoding method and the decoding method of the present invention are
It can be applied to a magneto-optical recording / reproducing method for a magneto-optical recording medium or a magnetic recording / reproducing method for a magnetic recording medium.

[0018]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Embodiments and examples of the present invention will be described with reference to the drawings. First, FIG. 1 shows a state transition diagram used for implementing the encoding method for multilevel recording according to the present invention. FIG. 1 is an example of a state transition diagram for executing a transition rule for d _mlt = 1 and k _mlt = 2 according to the encoding method of the present invention on a 4-valued data sequence. S00, S01, etc. shown in each circle in FIG. 1 indicate internal states at the time of transition, and arrows indicate state transitions that can be transited. In this example, 12 internal states S00 to S32 were used. The subscripts on the arrows indicate the four-level code output when each state transition is executed. For example, in the state S00, only the state S01 can be transitioned to, and 0 is output when the state S00 transits to the state S01. Further state S0
If 0 is output to the state S02 and the state is transitioned to 0, it is prohibited to output 0 from the state S02 to any internal state because there is a constraint of _kmlt = 2 in the next transition. Is configured.

Next, FIG. 2 shows a basic state transition matrix corresponding to the state transition diagram of FIG. Each row and each column of the basic state transition matrix shown in FIG. 2 corresponds to the internal states shown in the circle in FIG. State transition matrix r (1 ≦ r ≦ 12) rows, c
For the elements in the (1 ≦ r ≦ 12) column, 1 is set when the state transition from r to c is possible, and 0 is set when the state transition is not possible. For example, since the direct transition from the state S00 to the state S32 is prohibited in the state transition diagram of FIG. 1, the element in the S00 row and the S32 column indicates 0.

Next, the basic state transition matrix is raised to the nth power in order to extract a codeword group having an appropriate code (symbol) length n and complying with the above-mentioned _kmlt and _dmlt constraints. 2 to FIGS.
The result of n-th power calculation of the state transition matrix of is shown. Here, the case of n ≦ 10 is shown. If the element in the r-th row and the c-th column of the state transition matrix after the power calculation is a non-zero value, it is possible to transition from the state r to the state c with the code length n. Therefore, by finding a group of state transitions in which the element at row r and column c is not zero and obtaining the quaternary code output by those state transitions, the above-mentioned _kmlt , d with code length n is obtained.
It is possible to extract a 4-level code sequence that satisfies the _mlt constraint. For example, when the code length n = 5, the above k
Extract codewords that satisfy _mlt and d _mlt . 3 to 6, the following four states can be extracted as the states in which the elements of the fifth-order state transition matrix are not zero. That is, S00, S01, S10, and S11, and the following matrix is obtained from these four states.

[0021]

[Equation 1]

Since there is no zero element in the above matrix, it can be seen that the transition satisfying the above-mentioned _kmlt and _dmlt is possible. Here, S00, S01, S10, S
If the states 11 are renamed σ0, σ1, σ2, σ3, it means that the transition of σ0, σ1, σ2, σ3 can be realized with the code length 5. For example, the transition from σ0 to σ2 is
4-value series 0 → 0 → 1 → 3 → 0, 0 → 1 → 3 → 3 → 0, 0
→ 0 → 3 → 2 → 0 can be realized. It can be seen that this modulated 4-value coded sequence satisfies, of course, d _mlt = 1 and _{km lt} = 2. Also, k _mlt / d _mlt = 2>
It is one.

As described above, it is possible to extract a quaternary sequence of code length n which satisfies the d _mlt and _km lt constraints.
Further, among these four-valued sequences extracted by the procedure shown below, a sequence more suitable for a magneto-optical recording / reproducing method for reproducing a multi-valued signal according to the magnitude of the amplitude of the reproduced signal which is the object of the present invention. A group can be selected.

The magneto-optical recording medium suitable for the encoding of the present invention is, for example, relative to the external magnetic field strength at the time of recording as shown in FIG. 7 at the time of reproduction according to the structure of the recording medium and the magnetic characteristics of the recording layer described later. It has characteristics of signal output. Therefore, 4-level codes 0 to 3 can be assigned to each level of the relative signal output. Here, the four-level code is assigned as "0", "1", "2", "3" according to the magnitude of the relative signal output. From these signal intensities, the probability of erroneous conversion of “1” of the four-level code into “2” when reproducing the four-level signal and “
The probability of erroneous conversion of 2 "into" 3 "is almost equal. On the other hand, the Hamming distance (inter-code distance) of two code strings represents the different number of bits when the code word is expressed in binary notation. The Hamming distance when "1" is displayed as "2" is
The difference between (01) ₂ and (10) ₂ is 2, whereas
When the quaternary code “2” is displayed as “3”, the Hamming distance becomes 1 due to the difference between (10) ₂ and (11) ₂ .
Therefore, when a codeword group having the maximum minimum Hamming distance is extracted and applied to a trellis diagram using Viterbi decoding at the time of decoding a four-level code, an appropriate error correction function cannot be obtained. In the present invention, when extracting 4 value sequence code length n which satisfies the d _mlt and k _mlt constraints obtained from the matrix, the minimum Euclidean distance is extracted 4 value sequence to maximize. By doing so, it becomes possible to perform error correction that reflects the output characteristics of the reproduction signal when reproducing the multilevel recording medium.

The coding of the present invention can have stronger error correction capability when used in combination with trellis coding. FIG. 8 shows a schematic configuration of an encoder for carrying out the above-described four-level encoding after trellis encoding the input binary code sequence. This encoder is configured by connecting a convolutional encoder 81 and a mapping unit 82. For an input binary code sequence 80 composed of m binary codes, a binary code sequence having a coding rate of m / (m + 1) is obtained and then input to the mapping unit 82. The mapping unit 82 obtains the four-valued code sequence extracted according to the above-mentioned constraints of d _mlt and k _mlt and the definition of the minimum Euclidean distance.

In the convolutional encoder 81 of FIG.
Convolution with a code rate of m / (m + 1) will be described with reference to FIG. FIG. 9 is a diagram for explaining the process of the convolution unit 81 when m = 3. The binary sequence input to the convolution unit 81 is converted to the input binary code bi0, bi1, bi.
2 and the binary system output from the convolution unit 81 is output binary code sequences bo0, bo1, bo2, bo3. For example, if there is an input binary sequence desired to be recorded, the time t
_{When 0} , the binary code is cut out three by three, and the input binary code bi
0, bi1, and bi2 are assigned. At the next time t ₁ ,
From the binary sequence desired to be recorded, the following three binary codes are cut out and assigned to the input binary codes bi0, bi1 and bi2. In this way, the input binary code is sequentially input every 3 bits. As shown in FIG. 9, regarding the output binary codes bo0, bo1, and bo2, the input binary code bi
0, bi1, and bi2 are used as they are.
The input binary code bi2 is subjected to exclusive OR with the contents of the delay device 92 by the exclusive OR circuit 91 and input to the delay device 93. Regarding the output binary code bo3, the delay device 9
The contents of 3 are output. By performing the convolutional coding as shown in FIG. 9, the three binary codes are converted into four binary codes and input to the mapping unit 83 in FIG.

Binary codes bo0, bo1, bo2, bo3 output from the convolutional coding unit 81 of FIG. 8 are converted into a four-value code sequence 83 by the mapping unit 82 as follows. For example, the output binary code bo0, bo1,
bo2 and bo3 are paired with a pair of binary codes bo0 and bo1 and 2
It is divided into sets of value codes bo2 and bo3, and a state is assigned to each set. More specifically, bo0 is the upper bit,
When two bits with bo1 as the lower bit are displayed in decimal, 0 indicates state σ0 and 1 indicates state σ1.
If 2, then state σ2 is assigned, and if 3, state σ3 is assigned. Similarly, when two bits with bo2 as the upper bit and bo3 as the lower bit are displayed in decimal, the state σ0 is 0, the state σ1 is 1, the state σ2 is 2, and the state σ2 is 3. If so, the state σ3 is assigned. In this way, a four-valued sequence can be obtained from the input binary sequence, and those four-valued codes can be associated with the states σ0, σ1, σ2, σ3. Then, in order to execute the transition between these states with a desired code length, the basic state transition matrix based on the above-mentioned d _mlt and _{km lt} constraints can be used. Thus, the multi-level recording coding which satisfies the d _mlt and k _mlt constraints according to the present invention and utilizes the large coding gain which is a characteristic of trellis coding can be realized.

The above examples of convolutional coding and mapping are merely examples, and it is also possible to use different convolutional coding and mapping in the present invention. For example, when m = 2, three output binary codes are obtained by convolutional coding with a coding rate of 2/3, and then transitions among eight (2 ³ ) states (σ0 to σ7) are performed. It may be realized by considering d _mlt and k _mlt with a desired code length. As described above, the convolutional coding with an arbitrary code length and the mapping to the multi-level code sequence in which d _mlt and _kmt are taken into consideration with a desired code length are used together to execute the coding method of the present invention. Is possible. In addition, in FIG. 1, d _mlt = 1 and k _mlt
= 2, the state transition diagram was written, but _kmlt / _dmlt
By setting various values of k _mlt and d _mlt so that> 1, a state transition diagram similar to that of FIG. 1 can be drawn, and the encoding conversion for multilevel recording to multilevel code is performed by the same procedure as above. It becomes possible to obtain a table. However, and with minimal intersymbol interference between adjacent multilevel code, and for clock synchronization are prevented from becoming difficult to select the d _mlt and k _mlt that satisfies k _{ml t} / d _mlt <10 It is even more preferable.

Next, a magneto-optical recording medium to which the multilevel recording encoding and decoding method of the present invention is preferably applied will be described. As this type of recording medium, for example, the magneto-optical recording medium shown in the above-mentioned International Publication No. 8-129784 can be used. For example, as shown in FIG. 7, the magneto-optical recording medium satisfying the recording magnetic field characteristics shown in FIG. 7, that is, the reproduction signal output strength with respect to the external magnetic field strength H0 <H1 <H2 <H3 when recording is 4 An example of a magneto-optical recording medium for four-value recording which changes in the order of “0” → “2” → “1” → “3” as represented by a value signal will be described with reference to FIG.

The sectional structure of the magneto-optical recording medium is shown on the left side of FIG. This magneto-optical recording medium has a structure in which a first enhancing film, a first recording layer, a second enhancing film, a second recording layer, a third enhancing film, a reflecting film and a protective film are sequentially laminated on a substrate. SiN is used for each enhancement film and Al is used for the reflection film. In the first recording layer, the sublattice magnetic moment of the rare earth atom is more dominant than the sublattice magnetic moment of the transition metal atom from room temperature to the Curie temperature.
It is a bFeCo (RE-rich TbFeCo) layer. First
The recording layer has the recording magnetic field characteristics as shown in the lower right of FIG. 10, and the carrier-to-noise ratio of the light modulation recording signal with respect to the external magnetic field has one peak (recording state). On the other hand, the second
The recording layer is made of TbFeCo (RE-rich TbFeCo) in which the sublattice magnetic moment of the rare earth atom is more dominant than the sublattice magnetic moment of the transition metal atom from room temperature to the Curie temperature.
It has a structure in which an auxiliary magnetic layer made of PtCo is laminated on the Co layer. The second recording layer exhibits the recording magnetic field characteristics as shown in the upper right of FIG. 10, and the carrier-to-noise ratio of the optical modulation recording signal with respect to the external magnetic field has two peaks (recording state). Here, in order to obtain the magneto-optical recording medium satisfying the recording magnetic field characteristics as shown in FIG. 7, the sublattice magnetic moment of the transition metal atom of the second recording layer is the sublattice magnetic moment of the transition metal atom of the first recording layer. It may be about twice the lattice magnetic moment. For that purpose, the reproduction signal output range Aa of the second recording layer becomes larger than the reproduction signal output range Ab of the first recording layer shown in the lower left part of FIG. 10, as shown in the upper right part of FIG. Thus, a desired magneto-optical recording medium can be obtained by adjusting the film thickness or film composition ratio of each recording layer. Assuming that the downward external magnetic field is the external magnetic field in the recording direction and the upward external magnetic field is the external magnetic field in the erasing direction for signal recording, the relationship between the external magnetic fields H0 to H3 and the recorded four-valued code is as follows. become.

(I) An external magnetic field H 0 having a magnitude such that the entire magnetization direction of the first recording layer is oriented in the erasing direction (external magnetic field in the area indicated by circled numeral 1 in the upper right and lower right diagrams of FIG. 10)
Is applied in the erasing direction, the sublattice magnetic moments of the transition metal atoms in the first recording layer and the second recording layer can both be directed in the recording direction. The magnitude of change in the Kerr rotation angle detected as a signal from the magneto-optical recording medium is the first
Since it is proportional to the sum of the sublattice magnetic moments of the respective transition metal atoms in the recording layer and the second recording layer, the reproduction signal output corresponds to the lowest "0" in the four-valued code.

(Ii) An external magnetic field H1 having a magnitude such that the entire magnetization direction of the first recording layer is oriented in the erasing direction (external magnetic field in the area indicated by circled number 2 in the upper right and lower right diagrams in FIG. 10).
Is applied in the erasing direction, the sublattice magnetic moment of the transition metal atoms in the first recording layer can be directed in the recording direction, and the sublattice magnetic moment of the transition metal atoms in the second recording layer can be directed in the erasing direction. Since the sublattice magnetic moment of the transition metal atoms in the second recording layer is about twice the sublattice magnetic moment of the transition metal atoms in the first recording layer, the sublattice magnetic moments of the transition metal atoms in the entire magneto-optical recording medium are summed up. The moment is twice the sublattice magnetic moment of the transition metal atoms summed over the entire magneto-optical recording medium of (iii) below. That is, the reproduction signal output from the magneto-optical recording medium is a 4-level code
Equivalent to 2 ".

(Iii) Recording an external magnetic field H2 (the external magnetic field in the area indicated by the circled numeral 3 in the upper right and lower right views in FIG. 10) of a magnitude that allows the direction of the entire magnetization of the first recording layer to be oriented in the recording direction. By applying in the direction, the sublattice magnetic moment of the transition metal atoms of the first recording layer can be directed in the erasing direction, and the sublattice magnetic moment of the transition metal atoms of the second recording layer can be directed in the recording direction. The reproduction signal output from the magneto-optical recording medium in this case is assumed to be a four-level code "1".

(Iv) External magnetic field H3 of such a magnitude that the entire magnetization direction of the first recording layer is oriented in the recording direction (external magnetic field in the area indicated by circled number 4 in the upper right and lower right diagrams in FIG. 10).
Is applied in the recording direction, both the sublattice magnetic moments of the transition metal atoms in the first recording layer and the second recording layer can be directed in the erasing direction. The sublattice magnetic moment of the transition metal atoms summed over the entire magneto-optical recording medium is (iii)
3 times the total sublattice magnetic moment of the transition metal atoms in the entire magneto-optical recording medium. That is, the reproduction signal output from the magneto-optical recording medium corresponds to the four-level code "3".

From the reproduction signal output characteristics with respect to the external magnetic field for recording shown in the upper right diagram and the lower right diagram of FIG. 10 and the explanations of (i) to (iv) above, the recording magnetic field characteristic shown in FIG. It turns out that a relationship is obtained. Note that other recording magnetic field characteristics, for example, the external magnetic field strength H0 at the time of recording
Regarding the magneto-optical recording medium in which the intensity of the reproduction signal output for <H1 <H2 <H3 increases in this order, the reproduction output signal for the recording external magnetic field is different from the characteristics shown in the upper right and lower right diagrams of FIG. 4 by the same principle as
The value code will be recorded by the recording external magnetic field.

Next, using the encoding method of the present invention, FIG.
A method of performing four-level recording on the magneto-optical recording medium having the recording / reproducing characteristics as shown in 10 and 10 will be described. FIG.
FIG. 3 is a configuration diagram of a recording / reproducing system of a magneto-optical recording medium. At the time of recording, the magneto-optical recording medium 50 is driven by a drive system (not shown) to position the optical head 51 and the magnetic head 52 on a predetermined track of the magneto-optical recording medium 50. The data signal 53 is input to the 4-level recording encoder 54 and 4-level encoded according to the encoding method of the present invention.

Here, the configuration of the four-value recording encoder 54 is shown in FIG. The four-value recording encoder 54 includes a code setting unit 7
1 and an encoding table 72. Before actually inputting the binary information, the code setting unit 71 creates the coding table 72 in advance. Then, when the binary information is input, the 4-level record information is output using the encoding table 72. The code setting unit 71 includes a transition matrix setting unit 73,
Matrix exponentiation unit 75, coding rate setting unit 76, code extraction unit 77
It is composed of The transition matrix setting unit 73 uses d _mlt and _{km lt.}
A basic state transition matrix as shown in FIG. 2 is set based on the constraints. The coding rate setting unit 76 sets the coding rate, and the matrix exponentiation unit 75 exponentiates the basic state transition matrix based on this coding rate. Finally, the code extraction unit 77 performs code extraction that maximizes the minimum Euclidean distance, and creates the coding table 72. The binary code input to such an encoding table is output as a 4-level code having a code length n under the above-mentioned restrictions of d _mlt and _kmt . In the configuration of the encoder 54 of FIG. 12, the calculation and the like in the code setting section 71 can be executed outside the encoder 54 or the recording / reproducing system. In this case, the encoder 54 is executed outside. It may be configured by only the encoding table 72 storing the calculated result.

In FIG. 11, the four-level code encoded by the four-level encoder 54 is input to the external magnetic field applying circuit 55 to drive the magnetic head 52. At this time, the magnetic field strength is changed from H0 to H
Intensity modulation to 4 values of 3 and this intensity modulated 4
The external magnetic field having a value is applied to the optical information recording medium 50 in synchronization with the recording clock 56. Then, after the external magnetic field is switched to a predetermined value, a light pulse is irradiated from the optical head 51 driven by the laser driving circuit 57 to heat each recording layer of the light pulse irradiation unit to a temperature at which the magnetization can be reversed by the external magnetic field. To do. As a result, a magnetization domain corresponding to the magnitude of the external magnetic field is formed at the irradiation portion of each light pulse. During recording, a laser beam having a constant intensity may be continuously emitted instead of the pulsed light.

At the time of reproduction, the optical head 51 detects the reflected light from the recording track to obtain a reproduction signal corresponding to the Kerr rotation angle. After the reproduced signal was converted by the IV converter 58, optical intersymbol interference was removed by passing it through the equalizer 59, and the analog signal was A / D converted by the A / D converter 60. The digital transversal filter 61 absorbs the difference between the inner and outer circumference characteristics, and then the four-value judgment unit 62
Then, the 4-valued signal is determined, and then it is supplied to the Viterbi decoder 63 to binarize the 4-valued signal to obtain the data signal 64. The recovered clock is generated by the clock generator 6 from the signal after A / D conversion.
5, a clock signal is generated and supplied to the four-value determination unit 62 and the Viterbi decoder 63. In addition to the configuration example of the recording / reproducing system shown in FIG. 11, it is possible to implement the encoding and decoding method of the present invention by using a recording / reproducing system having an equivalent function.

In the above-described embodiments and examples of the present invention, the magnetic field modulation recording system for recording four-valued information according to the strength of the external magnetic field applied to the recording medium has been described as an example. Is 4 depending on the intensity of the light irradiating the recording medium.
It can also be applied to an optical modulation recording method for recording value information. For example, when four-valued information is recorded on the magneto-optical recording medium by the optical modulation method, the same two values as used in the above embodiment are used.
Using a magneto-optical recording medium having two recording layers, pulsed light that is signal-modulated (frequency-modulated) and light-intensity-modulated in accordance with 4-level information (or 2-level information) is applied while applying a magnetic field of constant intensity. It can be performed by irradiating.

In the above-described embodiment and examples, 4 as an example of multilevel encoding and decoding and multilevel recording and reproduction.
Although the case of values has been described, the present invention is not limited to this, and multi-level encoding and decoding such as three-value, five-value, six-value, seven-value, and eight-value, and multi-value recording and reproducing are also possible. Needless to say, it is included. Also, with respect to the code length n, it is possible to select various lengths and execute encoding.

In the above embodiments and examples, the case where the magneto-optical recording medium is recorded / reproduced as the recording medium has been described as an example, but the amplitude of the reproduced signal is a magnetic type that can express a multilevel code. The encoding and decoding method of the present invention can be applied to the case of recording / reproducing a recording medium, for example, a perpendicular magnetic recording medium.

[0043]

According to the encoding method of the present invention, inter-code interference and clock synchronization of a multi-level code are taken into consideration in recording / reproducing on a multi-level recording medium in which a multi-level code is represented by the amplitude of a reproduction signal. Multi-level recording encoding is possible. That is, instead of considering the minimum length d _x where 0s continue in one channel of a multi-channel like the two-dimensional code of Swanson and Wolf, or Tasaki et al., The constraint by the minimum length d _mlt where the same multilevel code continues Since the above is adopted, it becomes possible to perform multi-level recording encoding which can suppress inter-code interference of multi-level codes to be small. Similarly, the maximum length k _y 0 in the case of integrating the maximum length k _x and multi-channel in one channel of the multichannel 0 are continuous as Swanson and Wolf or Tasaki et al two-dimensional code, is continuous Rather than thinking about it, the maximum length k _mlt where the same multi-valued code continues
Since the restriction due to is considered, multi-level recording encoding suitable for clock synchronization becomes possible.

Furthermore, in addition to the constraints of k _mlt and d _mlt, a multi-valued code sequence composed of code words having the maximum minimum Euclidean distance is selected, so that when decoding by Viterbi decoding, It is possible to provide error correction suitable for reproduction of a reproduction signal in which the magnitude of amplitude represents a multilevel code. Furthermore, in the encoding method and the decoding method of the present invention, 2
After performing convolutional coding on the value code sequence, it is projected to a multi-level code sequence, and it is possible to perform the error correction function more strongly by performing Viterbi decoding corresponding to such convolution. Become.

[Brief description of drawings]

FIG. 1 is an example of a state transition diagram used in an encoding method for multilevel recording according to the present invention.

FIG. 2 is a diagram showing a basic state transition matrix corresponding to the state transition diagram of FIG.

FIG. 3 is a diagram showing a result of power calculation of the state transition matrix of FIG. 2 performed 1 to 3 times.

FIG. 4 is a diagram showing a result of power calculation of the state transition matrix of FIG. 2 performed 4 to 6 times.

5 is a diagram showing a result of performing power calculation of the state transition matrix of FIG. 2 7 to 9 times.

FIG. 6 is a diagram showing a result of 10-times power calculation of the state transition matrix of FIG.

FIG. 7 is a graph showing reproduction signal output characteristics with respect to magnetic field strength at the time of recording of a magneto-optical recording medium suitable for recording by the multilevel recording encoding method according to the present invention.

FIG. 8 shows a trellis-coded input binary code sequence,
It is a figure which shows the structure of the encoder for implementing four-value encoding.

9 is a diagram for explaining a convolution process with a coding rate of m / (m + 1) in the convolutional coding unit of FIG. 8;

10 is a diagram showing the structure of a magneto-optical recording medium having the reproduction signal output characteristics with respect to the magnetic field strength of FIG. 7 and the magnetic characteristics of its recording layer.

FIG. 11 is a diagram showing a configuration of a recording / reproducing apparatus for recording / reproducing a magneto-optical recording medium by using the encoding and decoding method for multilevel recording of the present invention.

FIG. 12 is a diagram showing the configuration of the encoder shown in FIG. 11.

[Explanation of symbols]

 50 magneto-optical recording medium 51 optical head 52 magnetic head 53 data signal 54 four-level encoder 56 recording clock 58 I / V converter 62 four-value determination unit 60 A / D converter 63 Viterbi decoder 71 code setting unit 72 code Conversion table 81 convolutional coding unit 82 mapping unit

Claims (11)

[Claims] 1. A multi-level code encoding method used in a multi-level recording / reproducing method in which a multi-level signal is reproduced according to the amplitude of a reproduced signal, wherein a minimum length d _mlt of the same multi-level code is continuous. The multi-valued coded input data sequence is such that the maximum length k _{mlt in} which the same multi-valued code as that of k _mlt / d _mlt > 1 is and the _kmlt / d _mlt has a finite value. Encoding method for recording. 2. The minimum length d _{mlt in} which the same multi-level code is continuous and the maximum length k _{mlt in} which the same multi-level code is continuous are k _mlt / d _mlt.
Select a multi-valued code sequence composed of code words having the largest minimum Euclidean distance from the multi-valued code sequences coded such that> 1 and k _mlt / d _mlt have a finite value. The encoding method for multilevel recording according to claim 1, wherein 3. The _kmlt and _dmlt are _kmlt / d
3. The encoding method for multilevel recording according to claim 1, wherein _mlt <10 is satisfied. 4. The input 2 as the input data series
The encoding method for multilevel recording according to any one of claims 1 to 3, wherein the value code sequence is convolutionally encoded and then projected onto the multilevel code sequence. 5. The input 2 as the input data series
The encoding method for multilevel recording according to claim 4, wherein the convolutional coding rate is m / (m + 1) with respect to the length m of the value code sequence. 6. When decoding a multi-level code encoded by the multi-level recording encoding method according to claim 2, a code composed of a code word that maximizes the minimum Euclidean distance is generated. A multilevel recording coding and decoding method characterized by performing Viterbi decoding using a mapped trellis diagram. 7. In the path metric calculation in the Viterbi decoding, the absolute value of the difference between each multi-level code at the time of recording and the multi-level code at the time of reproduction is regarded as the path metric, and the path having the minimum path metric is the surviving path. The multilevel recording encoding and decoding method according to claim 6, wherein 8. A Viterbi decoding corresponding to the convolutional coding is performed when a multilevel code encoded by the multilevel recording encoding method according to claim 4 or 5 is decoded into a binary code. An encoding and decoding method for multilevel recording characterized by the above. 9. A method for encoding and decoding a multilevel code used in a multilevel recording / reproducing method in which a multilevel signal is reproduced according to the magnitude of the amplitude of the reproduced signal, wherein the two input data sequences are input. After convolutionally coding the value code sequence, it is projected into a multilevel coded sequence, and the minimum length d _mlt where the same multilevel code continues and the maximum length k _mlt where the same multilevel code continues are _kmlt. / D _mlt > 1, and encodes the input data sequence so that k _mlt / d _mlt has a finite value, and selects a multi-valued code sequence composed of code words having the maximum minimum Euclidean distance. When the multi-valued code is decoded by performing the Viterbi decoding using the trellis diagram obtained by mapping the code composed of the code word having the maximum minimum Euclidean distance, When decoding to binary code Multilevel recording encoding and decoding method and performing Viterbi decoding corresponding to the convolutional coding the. 10. A magneto-optical recording medium in which a multilevel signal is reproduced according to the amplitude of a reproduction signal by using the encoding method for multilevel recording according to claim 1. Magneto-optical recording / reproducing method. 11. A magnetic recording medium in which a multilevel signal is reproduced according to the amplitude of a reproduction signal by using the encoding method for multilevel recording according to claim 1. Magnetic recording / reproducing method.