Information reproducing method and an information reproducing drive utilizing PRML to output binary values
AssigneeHitachi, Ltd.
InventorHiroyuki Minemura
Life of this patent
Estimating the term…
Abstract
Disclosed herewith is an information reproducing method for realizing significant expansion of disk capacity and processing signals having different minimum run lengths, as well as an optical disk drive that uses the method. To achieve the above objects, a PRML method is used. According to the method, a compensation value is added to an initial target value decided by a convolution operation of NN bits according to a bit array consisting of N bits (N>NN) to obtain a new target value, which is then compared with each of reproduced signals sequentially to select a bit array in which the error between the reproduced signal and the target signal is minimized most likely, then the selected bit array is binarized.
Claims10
An information reproducing method that employs a PRML method for comparing a target signal with each reproduced signal for a continuous N time to select the most likelihood one of state changes therein, thereby transforming said reproduced signal to a binary value;
- wherein, when said PRML method is represented as PR(α1, α2, . . . , αN), the leftmost M1 coefficients and the rightmost M2 coefficients in a coefficient array are all zero coefficients while integer values M1 and M2 satisfy a relationship of “M1≧0, M2≧0, M1+M2≧1, M1+M2<N”;
- and
- wherein, when integer MM=M1+M2 and integer NN=N−MM are satisfied, said method includes:
- a step of using a target value obtained by adding 2N or less compensated values V2, stored in a pattern compensation table, corresponding to a value of an N-bit digital bit array to an initial target level V1 obtained by a convolution operation of each of NN non-zero coefficient values and an NN-bit digital bit array, said pattern compensation table being located in a decode unit;
- a step of binarizing said reproduced signal to the most likelihood bit array while comparing said reproduced signal with said target value (V1+V2);
- and
- a step of outputting said binarized reproduced signal to an information reproducing device.
An information reproducing method that employs a PRML method for selecting the most likelihood one of state changes in a reproduced signal while comparing a target signal with each reproduced signal obtained for a continuous N time, thereby binarizing said reproduced signal to a binary value;
- wherein, when said PRML method is represented as PR(α1, α2, . . . , αN)ML, the logic of said state change excludes a state change logic of a reproduced signal whose minimum run length is R1 and under in accordance with the minimum run length R1 (R1≧1);
- wherein said method includes:
- a step of using a target value obtained by adding 2N or less compensation values V2 corresponding to a value of an N-bit digital bit array to an initial target level V1 obtained by a convolution operation of each of N coefficient values (α1, α2, . . . , αN) and an N-bit digital bit array;
- a step of binarizing said reproduced signal to the most likelihood bit array while comparing said reproduced signal with said target value (V1+V2);
- a step of setting a value larger than an amplitude of said reproduced signal as said compensation value V2 corresponding to said N-bit digital bit array if the run length of said N-bit digital bit array is R2 and less than the minimum run length R2 (R2>R1) signal is to be reproduced;
- and
- outputting the results of the step of binarizing and the step of setting to an information reproducing device.
An information reproducing drive for outputting a binary value obtained from a reproduced signal with use of a PRML method, said drive comprising:
- a PR target output unit for outputting a PR class target value corresponding to an N-bit bit array;
- a decode unit including a pattern compensation table for storing a compensation value corresponding to each M-bit bit array, where M is greater than N;
- a waveform equalizer for equalizing a reproduced signal;
- and
- a branch metric calculation unit for calculating a branch metric value for each M-bit bit array by employing a target value obtained by adding up a PR target value output from said PR target value output unit and a compensation value stored in said pattern compensation table with respect to an output from said waveform equalizer.
Dependencies are inferred from the claim text.