Magnetic reproducing processing device, magnetic recording / reproducing device, and magnetic reproducing method

Abstract

Provided are a magnetic reproduction processing device capable of suppressing errors, a magnetic recording and reproduction device, and a magnetic reproduction method. According to an embodiment, a magnetic reproduction processing device includes a decoder. The decoder includes a convolution layer including a plurality of filters, and an attention layer capable of deriving a degree of contribution related to the plurality of filters. The decoder is capable of outputting a decoding result obtained by synthesizing results obtained by processing an input signal in the plurality of filters according to the degree of contribution.

Description

[0001] This application is based on Japanese Patent Application 2022-138505 (filed on August 31, 2022), under which it enjoys priority benefits. This application incorporates the entire contents of that application by reference. Technical Field

[0002] The embodiments of the present invention relate to a magnetic reproduction processing apparatus, a magnetic recording reproduction apparatus, and a magnetic reproduction method. Background Technology

[0003] For example, fewer errors are desired in magnetic reproduction processing devices. Summary of the Invention

[0004] Embodiments of the present invention provide a magnetic reproduction processing apparatus, a magnetic recording reproduction apparatus, and a magnetic reproduction method capable of suppressing errors.

[0005] Technical solutions for solving the problem

[0006] According to an embodiment of the present invention, a magnetic reproduction processing apparatus includes a decoder. The decoder includes a convolutional layer comprising a plurality of filters; and an attention layer capable of deriving contribution values ​​related to the plurality of filters. The decoder is capable of outputting a decoding result, which is obtained by synthesizing the results of processing the input signal in the plurality of filters based on the contribution values.

[0007] The magnetic reproduction processing apparatus configured as described above can provide a magnetic reproduction processing apparatus, a magnetic recording reproduction apparatus, and a magnetic reproduction method that can suppress errors. Attached Figure Description

[0008] Figure 1 This is a schematic diagram illustrating the magnetic reproduction processing apparatus and magnetic recording reproduction apparatus according to the first embodiment.

[0009] Figure 2 This is a schematic diagram illustrating the magnetic reproduction processing apparatus and magnetic recording reproduction apparatus according to the first embodiment.

[0010] Figure 3 This is a schematic diagram illustrating the magnetic reproduction processing apparatus and magnetic recording reproduction apparatus according to the first embodiment.

[0011] Figure 4 This is a schematic diagram illustrating the magnetic reproduction processing apparatus and magnetic recording reproduction apparatus according to the first embodiment.

[0012] Figure 5 This is a graph illustrating the characteristics of a magnetic reproduction processing device.

[0013] Figure 6 This is a schematic diagram illustrating a magnetic reproduction processing apparatus according to an exemplary embodiment.

[0014] Label Explanation

[0015] 10: Convolutional layer; 10I: Input layer; 10M: Storage unit; 10N: Input node; 10O: Output layer; 11: Filter; 15: Attention layer; 20: Connection layer; 25: Other layers; 71: Decoder; 72: Error correction decoder; 73: Contribution adjuster; 74: Waveform adjuster; 75: Combiner; 76d: Display device; 76f: Interface circuit; 76i: Input device; 76m: Storage circuit; 76p: Processing circuit; 78 : Input interface; 80: Magnetic recording medium; 80D: Magnetic recording and reproduction unit; 80H: Magnetic head; 80R: Recording and reproduction unit; 110-113, 118, 119: Magnetic reproduction processing device; 210-213: Magnetic recording and reproduction device; CR: Contribution; LH1: Output information; RR1: Decoding result; Sig-c: Decoding signal; Sig-i: Input signal; Sig-r: Reproduced signal; cF: Filter number; nx: Number of repetitions Detailed Implementation

[0016] Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings.

[0017] The accompanying drawings are schematic or conceptual, and the relationship between the thickness and width of each part, the ratio of the size between parts, etc., are not required to be the same as in reality. Even when representing the same parts, the dimensions and ratios between them may sometimes be different depending on the accompanying drawings.

[0018] In this application specification and figures, the same reference numerals are used for elements that have been described with respect to the preceding figures, and detailed descriptions are omitted where appropriate.

[0019] (First Embodiment)

[0020] Figure 1 This is a schematic diagram illustrating the magnetic reproduction processing apparatus and magnetic recording reproduction apparatus according to the first embodiment.

[0021] like Figure 1 As shown, the magnetic recording and reproduction apparatus 210 according to the embodiment includes the magnetic reproduction processing apparatus 110 according to the embodiment. The magnetic recording and reproduction apparatus 210 may also include a magnetic recording and reproduction unit 80D.

[0022] The magnetic recording and reproduction unit 80D includes a magnetic recording medium 80. The magnetic recording medium 80 may include, for example, a hard disk drive (HDD). The magnetic recording and reproduction unit 80D may also include, for example, an SSD (Solid State Drive). The magnetic recording and reproduction unit 80D may also include a recording and reproduction unit 80R. The recording and reproduction unit 80R may include, for example, a magnetic head 80H. Information is recorded on the magnetic recording medium 80 by the recording and reproduction unit 80R (magnetic head 80H). The information recorded on the magnetic recording medium 80 is reproduced by the recording and reproduction unit 80R (magnetic head 80H). A reproduction signal Sig-r is obtained from the recording and reproduction unit 80R.

[0023] The reproduced signal Sig-r obtained by the magnetic recording and reproduction unit 80D (recording and reproduction unit 80R) is provided to the magnetic reproduction processing apparatus 110. The reproduced signal Sig-r is processed (decoded) in the magnetic reproduction processing apparatus 110. The result of the processing (decoding) by the magnetic reproduction processing apparatus 110 is output from the magnetic reproduction processing apparatus 110 as a decoded signal Sig-c. The decoded signal Sig-c is, for example, a binary signal "1, 0".

[0024] The magnetic reproduction processing apparatus 110 includes a decoder 71. The magnetic reproduction processing apparatus 110 includes, for example, an input interface 78. The reproduced signal Sig-r, or a signal based on the reproduced signal Sig-r, is provided to the decoder 71 via the input interface 78. An input signal Sig-i, including the reproduced signal Sig-r, is provided to the decoder 71. The input signal Sig-i (e.g., the reproduced signal Sig-r) can be a continuous or discontinuous signal in a time series.

[0025] like Figure 1 As shown, the decoder 71 includes a convolutional layer 10 and an attention layer 15. As described later, the decoder 71 may include an input layer 10I, a connection layer 20, and an output layer 10O.

[0026] Convolutional layer 10 includes multiple filters 11. Attention layer 15 is able to derive the contribution CR associated with the multiple filters 11.

[0027] Decoder 71 can output decoding result RR1, which is obtained by combining the input signal Sig-i processed in multiple filters 11 based on the contribution CR. Thus, higher precision decoding is possible.

[0028] For example, consider a first reference example of Viterbi decoding to PRML (Partial Response Maximum Likelihood). As will be described later, the decoding accuracy is sometimes insufficient in the first reference example. On the other hand, consider a second reference example, for example, where the input signal Sig-i is processed by a filter in decoder 71. In the second reference example, a filter is used to properly process the entire input signal Sig-i. In the second reference example, the decoding accuracy is sometimes insufficient depending on the state of the input signal Sig-i.

[0029] In contrast, in this embodiment, the result obtained by processing the input signal Sig-i in multiple filters 11 is used for decoding. This enables high-precision decoding. According to this embodiment, a magnetic reproduction processing apparatus capable of suppressing errors can be provided.

[0030] like Figure 1 As shown, the magnetic reproduction processing apparatus 110 may include a storage unit 10M. The storage unit 10M is capable of storing coefficients related to a plurality of filters 11. The coefficients are, for example, "weights". The attention layer 15 is capable of using the coefficients stored in the storage unit 10M to derive the contribution CR. The attention layer 15 is capable of estimating the contribution CR, for example.

[0031] like Figure 1 As shown, the attention layer 15 estimates the contribution CR corresponding to the number cF of the multiple filters 11 based on the coefficients stored in the storage unit 10M.

[0032] In this implementation, the coefficients stored in the storage unit 10M can be determined, for example, through machine learning.

[0033] Multiple filters 11 can be rationalized, for example, through machine learning. Decoder 71 includes, for example, an NN (neural network) construction. For example, attention layer 15 may include an NN (neural network). Convolutional layer 10 may include, for example, a CNN (convolutional neural network).

[0034] like Figure 1 As shown, the decoder 71 may also include an input layer 10I. The input signal Sig-i is input to the input layer 10I. The input signal Sig-i input to the input layer 10I is provided to the convolutional layer 10 and the attention layer 15.

[0035] like Figure 1As shown, the decoder 71 may also include a connection layer 20. Connection layer 20 combines the processing results of convolutional layer 10 and attention layer 15. For example, it synthesizes the results obtained from processing the input signal Sig-i in multiple filters 11 based on the contribution CR. For example, it calculates the product of the result obtained from processing the input signal Sig-i in multiple filters 11 and the contribution CR. The sum of the calculated products is then calculated. The processing result of connection layer 20 (decoding result RR1) can be output. Decoding result RR1 may include likelihood information.

[0036] like Figure 1 As shown, the processing result (decoding result RR1) of connection layer 20 can also be provided to other layers 25. In other layers 25, for example, at least a portion of multiple processing results can be combined. The processing results of other layers 25 can be provided to output layer 100.

[0037] like Figure 1 As shown, decoder 71 may also include output layer 100. Output layer 100 is capable of outputting output information LH1 based on the decoding result RR1. Output information LH1 may include, for example, a likelihood ratio.

[0038] In this implementation, the contribution CR estimated by the attention layer 15 can vary depending on the state of the input signal Sig-i. The state of the input signal Sig-i can be, for example, the signal waveform. The state of the input signal Sig-i can also be, for example, the length of the signal.

[0039] For example, the input signal Sig-i (reproduced signal Sig-r) includes N k The reproduced signal is T. "T" is the smallest recording unit (smallest recording period) in the reproduction (and recording). "k" is an integer greater than or equal to 1. For example, the reproduced signal Sig-r includes signals such as "1T", "2T", ..., "10T", ...

[0040] For example, the input signal Sig-i includes N i The reproduced signal of T and N j The reproduced signal of T. "i" is an integer greater than or equal to 1. "j" is an integer greater than or equal to 1. "j" is different from "i". For example, with N... i The contribution of T to the reproduced signal CR is different from that of N. j The contribution CR of the reproduced signal of T. For example, with N i The coefficients related to the reproduced signal of T are different from those related to N. j The coefficients related to the reproduced signal of T.

[0041] like Figure 1As shown, the magnetic reproduction processing apparatus 110 may further include an error correction decoder 72. As described above, the decoder 71 may also include an output layer 100. The output layer 100 is capable of outputting output information LH1 based on the decoding result RR1. The output information LH1 can be provided to the error correction decoder 72. Errors are corrected in the error correction decoder 72. The error correction decoder 72 may, for example, include an LDPC (Low-Density Parity-Check Codes) decoder.

[0042] like Figure 1 As shown, the output of the error correction decoder 72 can be provided to the decoder 71 as part of the input signal Sig-i. Through repeated processing, the error is further suppressed.

[0043] In an implementation, the number of input nodes 10N in the input layer 10I can be, for example, more than 5 and less than 300,000.

[0044] In an implementation, the number of multiple filters 11 can be, for example, more than 2 and less than 1000.

[0045] like Figure 1 As shown, the attention layer 15 can acquire at least a portion of the input signal Sig-i in parallel with the convolutional layer 10.

[0046] Figure 2 This is a schematic diagram illustrating the magnetic reproduction processing apparatus and magnetic recording reproduction apparatus according to the first embodiment.

[0047] like Figure 2 As shown, in the magnetic recording and reproduction apparatus 211 according to the embodiment, the magnetic reproduction processing apparatus 111 further includes a contribution adjuster 73. Otherwise, the structure of the magnetic reproduction processing apparatus 111 can be the same as that of the magnetic reproduction processing apparatus 110.

[0048] The contribution adjuster 73 is capable of adjusting the contribution CR associated with the plurality of filters 11. The contribution adjuster 73 adjusts the contribution CR associated with at least a portion of the plurality of filters 11 based on at least a portion of the processing result of the error correction decoder 72. In this example, at least a portion of the processing result of the error correction decoder 72 is provided to the contribution adjuster 73. In embodiments, the method for adjusting the contribution CR in the contribution adjuster 73 can be varied. For example, the contribution CR can be adjusted according to the characteristics of the magnetic recording reproduction unit 80D to which it is targeted. More appropriate decoding can be implemented.

[0049] Figure 3 This is a schematic diagram illustrating the magnetic reproduction processing apparatus and magnetic recording reproduction apparatus according to the first embodiment.

[0050] like Figure 3 As shown, in the magnetic recording and reproducing apparatus 212 according to the embodiment, the magnetic reproduction processing apparatus 112 further includes a waveform controller 74 (WC). Otherwise, the structure of the magnetic reproduction processing apparatus 112 can be the same as that of the magnetic reproduction processing apparatus 110.

[0051] Decoder 71 can acquire the input signal Sig-i adjusted by waveform adjuster 74. For example, waveform adjuster 74 acquires the reproduced signal Sig-r, adjusts the waveform of the reproduced signal Sig-r, and outputs it as the input signal Sig-i. Decoder 71 (e.g., input layer 10I) can acquire the input signal Sig-i adjusted by waveform adjuster 74. For example, it can adjust the waveform to match the characteristics of the magnetic recording reproduction unit 80D that is targeted. More appropriate decoding is possible. Waveform adjuster 74 may include, for example, FIR (Finite Impulse Response).

[0052] Figure 4 This is a schematic diagram illustrating the magnetic reproduction processing apparatus and magnetic recording reproduction apparatus according to the first embodiment.

[0053] like Figure 4 As shown, in the magnetic recording and reproducing apparatus 213 according to the embodiment, the magnetic reproduction processing apparatus 113 includes a plurality of decoders 71. Otherwise, the structure of the magnetic reproduction processing apparatus 113 may be the same as that of the magnetic reproduction processing apparatus 110.

[0054] In the magnetic reproduction processing apparatus 113, multiple decoders 71 can perform parallel processing. The processing results of the multiple decoders 71 are combined by a combiner 75. The output of the combiner 75 is provided to the error correction decoder 72. Thus, for example, the combiner 75 can combine the decoding result RR1 obtained from one of the multiple decoders 71 with the decoding result RR1 obtained from another of the multiple decoders 71. For example, the combiner 75 can combine the output information LH1 (e.g., likelihood ratio) obtained from one of the multiple decoders 71 with the output information LH1 (e.g., likelihood ratio) obtained from another of the multiple decoders 71. The result obtained by combining through the combiner 75 is provided to the error correction decoder 72. The processing result of the error correction decoder 72 can be provided to the multiple decoders 71, enabling more appropriate decoding. For example, faster processing is possible.

[0055] For example, the first learning condition of one of the multiple decoders 71 differs from the second learning condition of another of the multiple decoders 71. In one example, the first error function under the first learning condition differs from the second error function under the second learning condition. In another example, the arrangement of multiple learning data values ​​under the first learning condition differs from the arrangement of multiple learning data values ​​under the second learning condition. For example, under the first and second learning conditions, the arrangement of the learning data values ​​is in reverse order. For example, in learning, the reconstructed signal Sig-r is used as learning data. The reconstructed signal Sig-r is represented as multiple signal strength values ​​at multiple times. The learning data includes values ​​from the first to the kth. "k" is an integer greater than 2. "k" corresponds to a time. Under the first learning condition, learning is performed in the direction from the first value to the kth value. Under the second learning condition, learning is performed in the direction from the kth value to the first value. By processing the learning data under different learning conditions of the multiple decoders 71, more appropriate decoding can be performed.

[0056] Figure 5 This is a graph illustrating the characteristics of a magnetic reproduction processing device.

[0057] exist Figure 5 The characteristics related to the magnetic reproduction processing apparatus 110-113 are illustrated below. Furthermore, in... Figure 5 The characteristics related to the magnetic reproduction processing apparatus 118 of the first reference example and the characteristics related to the magnetic reproduction processing apparatus 119 of the second reference example are also illustrated. In the magnetic reproduction processing apparatus 118, Viterbi decoding using a PR filter is performed. Corresponding to the decoding in the first reference example, the number of filters is 1. In the magnetic reproduction processing apparatus 119 of the second reference example, the reproduced signal is processed by a filter. The first reference example corresponds to a structure in which the decoder 71 is replaced with a Viterbi decoder. The second reference example corresponds to a structure in which the decoder 71 is processed by a filter. Apart from the above, the structures of the magnetic reproduction processing apparatus 118 and the magnetic reproduction processing apparatus 119 are the same as the structure of the magnetic reproduction processing apparatus 110. Figure 5 The horizontal axis represents the number of iterations, nx. The number of iterations, nx, is represented using standardization. Figure 5 The vertical axis represents BER (Bit Error Rate).

[0058] exist Figure 5In this example, the input layer 10I has 11 nodes and 1 channel. The convolutional layer 10 includes 15 filters 11. In other layers 25 (e.g., fully connected layers), there are 5 layers and 10 nodes. The output layer 10O has 1 output node. In this example, the attention layer 15 is a fully connected layer with 5 layers, 11 input nodes, and 15 output nodes. The output of the attention layer 15 is a softmax output.

[0059] like Figure 5 As shown, in the magnetic reproduction processing apparatuses 110 to 113 according to the embodiments, a lower BER is obtained than that in the magnetic reproduction processing apparatus 118 of the first reference example and the magnetic reproduction processing apparatus 119 of the second reference example. The BER in magnetic reproduction processing apparatus 111 is lower than that in magnetic reproduction processing apparatus 110. The BER in magnetic reproduction processing apparatus 112 is lower than that in magnetic reproduction processing apparatus 111. The BER in magnetic reproduction processing apparatus 113 is lower than that in magnetic reproduction processing apparatus 112.

[0060] The magnetic reproduction processing apparatus involved in the implementation may include a computer.

[0061] Figure 6 This is a schematic diagram illustrating a magnetic reproduction processing apparatus according to an exemplary embodiment.

[0062] like Figure 6 As shown, the magnetic reproduction processing apparatus (e.g., magnetic reproduction processing apparatus 110) according to the embodiment may include a processing circuit 76p, a storage circuit 76m, and an interface circuit 76f, etc. The processing circuit 76p is, for example, an electronic circuit. The storage circuit 76m may also include, for example, at least one of ROM (Read Only Memory) and RAM (Random Access Memory). As the storage circuit 76m, for example, part of the magnetic recording reproduction unit 80D may also be used.

[0063] The magnetic reproduction processing apparatus (e.g., magnetic reproduction processing apparatus 110) involved in the embodiments may also include a display device 76d and an input device 76i, etc. The display device 76d may include various displays. The input device 76i may include, for example, a device with operating functions (e.g., a keyboard, mouse, touch input panel, or voice recognition input device, etc.).

[0064] Among the multiple elements included in the magnetic reproduction processing apparatus (e.g., magnetic reproduction processing apparatus 110) according to the embodiments, they can communicate with each other via at least one of wireless and wired methods. The locations where the multiple elements included in the magnetic reproduction processing apparatus 110 are installed can also be different from each other. For example, a general-purpose computer can be used as the magnetic reproduction processing apparatus 110. For example, multiple computers interconnected can be used as the magnetic reproduction processing apparatus 110. Dedicated circuits can also be used as at least a part of the magnetic reproduction processing apparatus 110. For example, multiple interconnected circuits can be used as the magnetic reproduction processing apparatus 110.

[0065] The implementation may also include a program. The program causes the computer (magnetic reproduction processing device 110) to perform the above-described actions. The implementation may also include a storage medium storing the above-described program.

[0066] (Second Implementation)

[0067] The second embodiment relates to a magnetic reproduction method. The magnetic reproduction method of this embodiment utilizes the magnetic reproduction processing apparatus (e.g., 110-113) and variations thereof described in the first embodiment. This provides a magnetic reproduction method capable of suppressing errors.

[0068] The implementation methods may include the following technical solutions.

[0069] (Technical Solution 1)

[0070] A magnetic reproduction processing device,

[0071] Equipped with a decoder

[0072] The decoder includes:

[0073] Convolutional layers, which include multiple filters; and

[0074] An attention layer that can derive the contribution values ​​associated with the multiple filters.

[0075] The decoder is capable of outputting a decoding result, which is obtained by combining the results of processing the input signal in the plurality of filters based on the contribution.

[0076] (Technical Solution 2)

[0077] According to the magnetic reproduction processing device described in technical solution 1

[0078] It also includes a storage unit capable of storing coefficients related to the plurality of filters.

[0079] The attention layer can use the coefficients to derive the contribution.

[0080] (Technical Solution 3)

[0081] The magnetic reproduction processing device according to technical solution 1 or 2

[0082] The decoding result includes likelihood information.

[0083] (Technical Solution 4)

[0084] The magnetic reproduction processing apparatus according to any one of technical solutions 1 to 3

[0085] The input signal includes the reproduced signal obtained from the magnetic recording reproducibility unit.

[0086] (Technical Solution 5)

[0087] The magnetic reproduction processing apparatus according to any one of technical solutions 1 to 4

[0088] The input signal includes N i The reproduced signal of T and N j The reproduced signal of T,

[0089] T is the smallest recording unit in the reproduction process.

[0090] The i is an integer greater than or equal to 1.

[0091] j is an integer greater than or equal to 1.

[0092] The j is different from the i.

[0093] With the N i The contribution of T to the reproduced signal and the contribution of N j The contribution of T to the reproduced signal is different.

[0094] (Technical Solution 6)

[0095] The magnetic reproduction processing apparatus according to any one of technical solutions 1 to 5

[0096] The decoder also includes an output layer.

[0097] The output layer is capable of outputting information based on the decoding result.

[0098] The output information includes the likelihood ratio.

[0099] (Technical Solution 7)

[0100] The magnetic reproduction processing apparatus according to any one of technical solutions 1 to 5

[0101] It also has an error correction decoder.

[0102] The decoder also includes an output layer.

[0103] The output layer is capable of outputting information based on the decoding result.

[0104] The output information can be provided to the error correction decoder.

[0105] (Technical Solution 8)

[0106] The magnetic reproduction processing device according to technical solution 7

[0107] The output of the error correction decoder can be provided to the decoder as part of the input signal.

[0108] (Technical Solution 9)

[0109] The magnetic reproduction processing apparatus according to any one of technical solutions 1 to 8

[0110] The number of the plurality of filters is more than 2 and less than 1000.

[0111] (Technical Solution 10)

[0112] The magnetic reproduction processing apparatus according to any one of technical solutions 1 to 9

[0113] The attention layer can acquire at least a portion of the input signal in parallel with the convolutional layer.

[0114] (Technical Solution 11)

[0115] The magnetic reproduction processing apparatus according to any one of technical solutions 1 to 10,

[0116] The decoder also includes an input layer.

[0117] The input signal is input to the input layer.

[0118] The input signal that has been fed into the input layer is provided to the convolutional layer and the attention layer.

[0119] The number of input nodes in the input layer is more than 5 and less than 300,000.

[0120] (Technical Solution 12)

[0121] The magnetic reproduction processing apparatus according to any one of technical solutions 1 to 10,

[0122] It also has a waveform adjuster.

[0123] The decoder is able to obtain the input signal after it has been adjusted in the waveform adjuster.

[0124] (Technical Solution 13)

[0125] The magnetic reproduction processing apparatus according to any one of technical solutions 1 to 12,

[0126] It also has a contribution adjuster.

[0127] The contribution adjuster can adjust the contribution associated with the plurality of filters.

[0128] (Technical Solution 14)

[0129] The magnetic reproduction processing apparatus according to any one of technical solutions 1 to 13

[0130] It also has a connector.

[0131] Multiple decoders are provided.

[0132] The combiner is capable of combining the decoding result obtained from one of the plurality of decoders with the decoding result obtained from another of the plurality of decoders.

[0133] (Technical Solution 15)

[0134] The magnetic reproduction processing apparatus according to any one of technical solutions 1 to 13

[0135] Multiple decoders are provided.

[0136] The first learning condition of one of the plurality of decoders is different from the second learning condition of another of the plurality of decoders.

[0137] (Technical Solution 16)

[0138] The magnetic reproduction processing device according to technical solution 15

[0139] The first error function under the first learning condition is different from the second error function under the second learning condition.

[0140] (Technical Solution 17)

[0141] The magnetic reproduction processing device according to technical solution 15

[0142] The arrangement of the multiple learning data values ​​under the first learning condition is different from the arrangement of the multiple learning data values ​​under the second learning condition.

[0143] (Technical Solution 18)

[0144] The magnetic reproduction processing apparatus according to any one of technical solutions 1 to 15

[0145] The decoder includes a neural network construction.

[0146] (Technical Solution 19)

[0147] A magnetic recording and reproducing device, comprising:

[0148] The magnetic reproduction processing apparatus described in technical solution 4; and

[0149] The magnetic recording reproduction unit.

[0150] (Technical Solution 20)

[0151] A magnetic reproduction method using the magnetic reproduction processing apparatus described in any one of technical solutions 1 to 18.

[0152] According to the embodiments, it is possible to provide a magnetic reproduction processing apparatus, a magnetic recording reproduction apparatus, and a magnetic reproduction method that can suppress errors.

[0153] The embodiments of the present invention have been described above with reference to examples. However, the present invention is not limited to these examples. For example, the specific structure of each element of the magnetic reproduction processing apparatus, such as the decoder, error correction decoder, waveform adjuster, and contribution adjuster, is included within the scope of the present invention as long as those skilled in the art can appropriately select from the well-known range to similarly implement the present invention and obtain the same effect.

[0154] Any technical solution obtained by combining any two or more elements of the various examples within the scope of technical feasibility, as long as it contains the spirit of the present invention, is also included within the scope of the present invention.

[0155] All magnetic reproduction processing apparatuses, magnetic recording reproduction apparatuses, and magnetic reproduction methods that can be implemented by appropriate design modifications by those skilled in the art, based on the magnetic reproduction processing apparatus, magnetic recording reproduction apparatus, and magnetic reproduction method described above as embodiments of the present invention, as long as they contain the spirit of the present invention, are also within the scope of the present invention.

[0156] Within the scope of the present invention, those skilled in the art will be able to conceive of various modifications and alterations, which are also considered to fall within the scope of the present invention.

[0157] Several embodiments of the present invention have been described above, but these embodiments are merely illustrative and not intended to limit the scope of the invention. These new embodiments can be implemented in a wide variety of other ways, and various omissions, substitutions, and modifications can be made without departing from the spirit of the invention. These embodiments and / or variations thereof are included within the scope and spirit of the invention, and are included within the scope of the invention as described in the claims and its equivalents.

Claims

1. A magnetic reproduction processing device, Equipped with a decoder The decoder includes: A convolutional layer, which includes multiple filters; and An attention layer that can derive the contribution values ​​associated with the multiple filters. The decoder is capable of outputting a decoding result, which is obtained by combining the results of processing the input signal in the plurality of filters based on the contribution degree. The input signal includes N i The reproduced signal of T and N j The reproduced signal of T, T is the smallest recording unit in the reproduction process. The i is an integer greater than or equal to 1. j is an integer greater than or equal to 1. The j is different from the i. With the N i The contribution of T to the reproduced signal and the contribution of N j The contribution of T to the reproduced signal is different.

2. The magnetic reproduction processing apparatus according to claim 1, It also includes a storage unit capable of storing coefficients related to the plurality of filters. The attention layer can use the coefficients to derive the contribution.

3. The magnetic reproduction processing apparatus according to claim 1, The input signal includes the reproduced signal obtained from the magnetic recording reproducibility unit.

4. The magnetic reproduction processing apparatus according to claim 1, The decoder also includes an output layer. The output layer is capable of outputting information based on the decoding result. The output information includes the likelihood ratio.

5. The magnetic reproduction processing apparatus according to claim 1, It also has an error correction decoder. The decoder also includes an output layer. The output layer is capable of outputting information based on the decoding result. The output information can be provided to the error correction decoder.

6. The magnetic reproduction processing apparatus according to claim 5, The output of the error correction decoder can be provided to the decoder as part of the input signal.

7. The magnetic reproduction processing apparatus according to claim 1, It also has a waveform adjuster. The decoder is able to obtain the input signal after it has been adjusted in the waveform adjuster.

8. A magnetic recording and reproducing apparatus, comprising: The magnetic reproduction processing apparatus according to claim 3; and The magnetic recording reproduction unit.

9. A magnetic reproduction method, Use the magnetic reproduction processing apparatus according to claim 1.

Images