Signal processing device, magnetic tape cartridge, magnetic tape reading apparatus, processing method of signal processing device, operation method of magnetic tape reading apparatus, and non-transitory computer-readable storage medium

Patent metadata

Field	Value
Publication number	US-11495248-B2
Application number	US-202117353820-A
Country	US
Kind code	B2
Filing date	Jun 22, 2021
Priority date	Jun 23, 2020
Publication date	Nov 8, 2022
Grant date	Nov 8, 2022

Abstract

A signal processing device includes a receiver that receives a plurality of playback signal sequence obtained by digitizing a plurality of reading results by a plurality of A/D converter, the plurality of reading results being obtained by reading data by a plurality of reading elements from a magnetic tape and a plurality of equalizers that perform waveform equalization of the plurality of playback signal sequence. The plurality of equalizers perform the waveform equalization by using a plurality of non-linear filters that have been learned to reduce distortion that occurs non-linearly in the plurality of playback signal sequence according to a condition under an environment in which the data is read from the magnetic tape. The plurality of non-linear filters being optimized to a suitable characteristic for the plurality of reading elements by optimization based on the plurality of reading results.

First claim

What is claimed is:

1. A signal processing device comprising:
a receiver that receives a plurality of playback signal sequence obtained by digitizing a plurality of reading results by a plurality of A/D converters, the plurality of reading results being obtained by reading a specific pattern recorded as the data in a specific region of a magnetic tape by reading with a plurality of reading elements installed in a reading head from the magnetic tape on which the data is recorded; and
a plurality of equalizers that perform waveform equalization of the plurality of playback signal sequence received by the receiver,
wherein
the plurality of equalizers perform the waveform equalization by using a plurality of non-linear filters that have been learned to reduce distortion that occurs non-linearly in the plurality of playback signal sequence according to a condition under an environment in which the data is read from the magnetic tape, and
the plurality of non-linear filters are optimized to a suitable characteristic for the plurality of reading elements, based on the plurality of reading results;
wherein the condition includes a condition caused by an individual difference in the reading head.

2. The signal processing device according to claim 1 ,
wherein the specific pattern is read by the plurality of data reading elements in parallel to an action of the specific pattern being recorded in the specific region by a plurality of recording elements arranged upstream of the plurality of reading elements in a forward direction of the magnetic tape.

3. The signal processing device according to claim 1 ,
wherein the condition includes a condition caused by an individual difference in the magnetic tape.

4. The signal processing device according to claim 1 ,
wherein the condition includes a speed condition regarding a speed at which the magnetic tape runs.

5. The signal processing device according to claim 4 ,
wherein the speed condition includes a condition regarding a running speed of the magnetic tape for a case in which recording is performed on the magnetic tape.

6. The signal processing device according to claim 1 ,
wherein the condition includes a condition caused by an individual difference in a processing circuit that affects the waveform equalization.

7. The signal processing device according to claim 1 ,
wherein the non-linear filter is a filter having a neural network on which the learning has been performed.

8. The signal processing device according to claim 7 , further comprising:
a plurality of storage elements that are each respectively provided to each of the reading elements and in which the playback signal sequence is stored in time-series,
wherein the neural network has a front-stage layer having a plurality of front-stage layer nodes corresponding to the plurality of storage elements and a back-stage layer,
each of the plurality of storage elements outputs the input playback signal sequence to a corresponding front-stage layer node among the plurality of front-stage layer nodes,
each of the plurality of front-stage layer nodes outputs the playback signal sequence input from a corresponding storage element among the plurality of storage elements to the back-stage layer,
the back-stage layer
converts a composite value obtained based on a product sum of the playback signal sequence input from the plurality of front-stage layer nodes and a back-stage layer coupling weight by an activation function, and
outputs a back-stage layer value based on the converted value obtained by converting the composite value with the activation function, and
the back-stage layer coupling weights are determined by learning performed on the neural network to minimize a deviation amount between the back-stage layer value and a predetermined target value as the learning.

9. The signal processing device according to claim 8 ,
wherein the neural network has an input layer as the front-stage layer and has a middle layer and an output layer as the back-stage layer,
the plurality of front-stage layer nodes are a plurality of input layer nodes,
the middle layer has a plurality of middle layer nodes,
each of the plurality of input layer nodes outputs the playback signal sequence input from a corresponding storage element among the plurality of storage elements to the middle layer,
the plurality of middle layer nodes convert an middle layer value obtained as the composite value based on a product sum of the playback signal sequence input from the plurality of input layer nodes and an middle layer coupling weight by the activation function to generate the converted value and output the converted value to the output layer,
the output layer outputs an output layer value obtained as the back-stage layer value based on a product sum of the converted value input from the middle layer and an output layer coupling weight, and
the middle layer coupling weight and the output layer coupling weight are determined by learning performed on the neural network to minimize a deviation amount between the output layer value and a predetermined target value as the learning.

10. The signal processing device according to claim 9 ,
wherein the middle layer value is a value based on the product sum of the playback signal sequence and the middle layer coupling weight and a first variable, and
the first variable is determined by the learning performed on the neural network.

11. The signal processing device according to claim 8 ,
wherein the neural network consists of two layers, the front-stage layer and the back-stage layer.

12. The signal processing device according to claim 8 ,
wherein the back-stage layer value is a value based on a product sum of the converted value and the back-stage layer coupling weight and a second variable, and
the second variable is determined by the learning performed on the neural network.

13. The signal processing device according to claim 8 ,
wherein the plurality of storage elements are a plurality of delay elements in which the playback signal sequence is input with a delay of a predetermined time, and
the back-stage layer value is a value regarding the playback signal sequence input first among the plurality of playback signal sequence stored in the plurality of delay elements.

14. The signal processing device according to claim 8 ,
wherein the target value is teaching data set in advance based on at least one of an ideal playback signal sequence regarding known data recorded on a learning magnetic tape in a recording pattern set in advance along a longitudinal direction of the learning magnetic tape or an ideal playback signal sequence derived by computer simulation.

15. A magnetic tape cartridge comprising:
a magnetic tape,
wherein the magnetic tape is recorded with a parameter related to the plurality of non-linear filters employed by the signal processing device of claim 1 .

16. A magnetic tape cartridge comprising:
a non-contact storage medium,
wherein the non-contact storage medium is stored with a parameter related to the plurality of non-linear filters employed by the signal processing device of claim 1 .

17. A magnetic tape reading apparatus comprising:
a reading head installed with a plurality of reading elements that read data from a magnetic tape on which the data is recorded;
a receiver that receives a plurality of playback signal sequence obtained by digitizing a plurality of reading results with a plurality of A/D converters, the plurality of reading results being obtained by reading a specific pattern recorded as t

Assignees

• Fujifilm Corp

Inventors

• Musha Atsushi
• Okamoto Yoshihiro
• Nakamura Yasuaki
• Nishikawa Madoka

Classifications

• G06N3/045

  Combinations of networks · CPC title

• G06N3/048

  Activation functions · CPC title

• G06N3/08

  Learning methods · CPC title

• G11B5/4893

  relative to moving tape · CPC title

• G11B5/00813 Primary

  magnetic tapes · CPC title