JP3000938B2 - Data detection / reproduction method and apparatus - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method and an apparatus for detecting and reproducing digitally recorded data such as a digital recording disk device and a digital recording VTR.

[0002]

2. Description of the Related Art Conventionally, in a digital recording disk device, a digital recording VTR, or the like, data is not recorded as it is, but is recorded and encoded before recording. Conventionally, 1,7 codes and 2,7 codes are known as typical recording codes.

Here, a code conversion table of 1,7 codes is shown in FIG.
Shown in After converting 2 data bits into 3 channel bits or converting 4 data bits into 6 channel bits, the NR
Record according to ZI rules. The NRZI rule is inverted with "1",
This is a rule that recording is performed by performing non-inversion with “0”. 1,
As a feature of the seven codes, there is a feature that one or more and seven or less “0” exist between “1” and “1” after conversion.

Next, a code conversion table for 2,7 codes is shown in FIG. Either convert 2 data bits to 4 channel bits, convert 3 data bits to 6 channel bits, or convert 4 data bits to 8 channel bits Thereafter, recording is performed according to the NRZI rule. As a large feature of the 2,7 code, there is a feature that after conversion, two or more "0" s exist between "1" and "1".

Conventionally, there has been proposed a method of performing data detection by Viterbi decoding after controlling a center level of a recording / reproducing signal. This is described in, for example,
No. 5504, entitled "Optical Disk Apparatus", which records predetermined reference data in advance, detects the maximum and minimum values of the reproduced signal, sets the average level to the center level, and crosses the center level. After adding an offset value to the data so that the average value of the reproduced data becomes 0, or accumulating data that crosses the data and adding an offset value so that the average value becomes 0, the offset of the reproduced data is reduced, and then the Viterbi signal is reduced. It has been proposed to perform the decoding.

[0006]

At present, there is a phase change disk and a device as one of recordable digital disks. This disc has a “crystalline state” and an “amorphous” state. For example, digital data is recorded by associating digital data “H” with “crystalline state” and “L” with “amorphous state”. It is.

However, when data is recorded / reproduced on / from this digital disk, the reproduced data has a "crystal state".
That is, when transitioning from “H” to “amorphous state”, ie, “L”, data that should be in “crystal state”, ie, “H” shifts to “amorphous state”, ie, “L” side, and thereafter frequently occurs in Viterbi decoding output. There is a problem that an error occurs. Similar errors in different directions may occur on disks other than phase change disks or on tape media.

In such an error, the maximum value and the minimum value do not change, and the "H" data shifts to the "L" side only at the transition point from "H" to "L". Accordingly, the center level is determined based on the maximum value and the minimum value.
The effect cannot be exerted by the decoding method disclosed in Japanese Patent Laid-Open Publication No. H10-197572.

Here, an object of the present invention is to solve the above-mentioned problem, and when a waveform shift occurs in a specific direction due to a medium characteristic, decoding is performed in consideration of a waveform shift generated due to the medium characteristic in advance. To provide a data detection method and apparatus that improves the overall bit error rate.

[0010]

The present invention solves the above-mentioned problems. In the present invention, the reproduction data changes from "H" level to "L" at a transition point from "H" level to "L" level. In a data detection / reproduction method used in a recording / reproduction system in which an error to the “level” is likely to occur, a transition from the “H” level to the “L” level is detected, and the reproduction data before and after the transition is determined by a predetermined positive offset value. Then, by performing data detection determination, the effect of the subsequent Viterbi decoding can be greatly exerted, and the bit error rate can be minimized.

[0011]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Here, an embodiment of the present invention will be described with reference to the drawings.

FIG. 1 shows an error occurrence pattern of reproduced data. FIG. 1A shows “H” for the reproduction data.
FIG. 1 shows a situation in which a shift from the “H” level to the “L” level occurs at a transition point where the level shifts to the “L” level.
FIG. 1B shows a state in which the reproduction data shifts from the “L” level to the “H” level at a transition point where the “H” level shifts to the “L” level, and FIG.
FIG. 1D shows a state in which a shift from the “H” level to the “L” level occurs at the transition point where the playback data shifts from the “L” level to the “H” level, and FIG. On the other hand, a situation in which a shift from the “L” level to the “H” level occurs at a transition point where the “L” level shifts to the “H” level is shown.

In the present embodiment, an example will be described regarding a partial response PR (1, 1) equalization method as a reproduction equalization method.

FIG. 2 shows a block diagram for performing recording and reproduction in the partial response PR (1, 1) equalization method.
The recording-encoded input signal a (t) becomes the recording signal b
(T) is recorded on the disc. Playback signal c
(T) is equalized output d (t) by the partial response PR (1, 1) equalizing means, and is passed through the waveform shaping means and Viterbi decoding disclosed by the present invention, and is output to the Viterbi decoded output e.
Reproduced as (t).

FIG. 3 is a timing chart of each signal in the block diagram of FIG. 2 showing the recording and reproduction.

When the input signal a (t) is the signal shown in FIG. 3, the waveforms of the recording signal b (t), the reproduced signal c (t), and the equalized output d (t) are shown in FIG. And the Viterbi decoded output e (t) becomes the signal shown in FIG.

The equalized output d (t) is equal to 3 as shown in FIG.
Value signal. If there is no error in the recording / reproducing system, d (t) = "0" is obtained for the polar change of the input signal, and d (t) = "+ 1" or d (t) for the polar non-change of the input signal. "-1" is obtained. Here, the equalization output d (t) is subjected to the following processing in the waveform shaping means shown in FIG.

When a shift from the "H" level to the "L" level occurs at a change point where the "H" level shifts to the "L" level, for example, "1" → "0" → "-1" The transition is made with a shift of “0.8” → “−0.2” → “−1”. Then, the deviation increases, and the first data (in this case, 0.45) and the second data, for example, “0.45” → “−0.55” → “−1”,
(In this case, -0.55), the Viterbi decoded output e (t) described later becomes an error.

In such a case, a predetermined value is determined in advance, and the value is added to the reproduction data to correct the reproduction data (here, the reproduction data is a signal obtained by reproducing and equalizing the reproduction signal. Point). For example, when the value shifts from a positive value to a negative value, for example, 0.1 is added to both data (first data and second data). Therefore,
The previous data of “0.45” → “−0.55” → “−1” is “0.55” → “−0.45” → “−1”.
A Viterbi decoding output described later has no error.

When a shift from the "L" level to the "H" level occurs at a transition point from the "H" level to the "L" level, for example, "1" → "0" → "-" 1 "
"1" .fwdarw. "0.2" .fwdarw. "-0.
8 ”and shifts. This shift increases, and the second data (0.55 in this case) becomes“ 1 ”→“ 0.55 ”→“ −0.45 ”. ) And the magnitude of the absolute value of the third data (in this case, -0.45) is reversed, the Viterbi decoding described later results in an error.

In such a case, a predetermined value is determined in advance, and the reproduction data is corrected by adding the value to the reproduction data. For example, when the value shifts from a positive value to a negative value, −0.1 is added (subtracted by 0.1) to both data (second data and third data). Therefore, the previous data “1” → “0.55” → “−0.45” becomes “1” → “0.45” → “−0.55”,
A Viterbi decoding output described later has no error.

When a shift from the "H" level to the "L" level occurs at a transition point from the "L" level to the "H" level, for example, "-1" → "0" → " 1 "
"-1" → "-0.2" → "0.
8 ”and shifts. This shift increases, and the second data (in this case, −0 in this case) becomes“ −1 ”→“ −0.55 ”→“ 0.45 ”. .55) and the magnitude of the absolute value of the third data (0.45 in this case) is reversed, an error occurs in Viterbi decoding described later.

In such a case, a predetermined value is determined in advance, and the reproduction data is corrected by adding the value to the reproduction data. For example, when shifting from a negative value to a positive value, 0.1 is added to both data (the second data and the third data). Therefore, the above “−1” → “−”
The data of “0.55” → “0.45” is “−1” → “−
0.45 "→" 0.55 ", and a Viterbi decoded output described later has no error.

When a shift from the "L" level to the "H" level occurs at the transition point from the "L" level to the "H" level, for example, "-1" → "0" → " 1 "
"-0.8" → "0.2" →
A shift occurs with a shift of “1”. Then, this deviation increases, and the first data (in this case, -0.45) and the second data (in this case, -0.45) → “0.55” → “1” When the magnitude of the absolute value of (0.55) is reversed, Viterbi decoding described later results in an error.

In such a case, a predetermined value is determined in advance, and the reproduction data is corrected by adding the value to the reproduction data. For example, when the value shifts from a positive value to a negative value, −0.1 is added to both data (first data and second data) (0.1 is subtracted). Therefore, the previous data of “−0.45” → “0.55” → “1” becomes “−0.55” → “0.55” → “1”, and
A Viterbi decoding output described later has no error.

Next, the Viterbi decoding will be described.

FIG. 4 shows a state transition diagram in Viterbi decoding, and FIG. 5 shows a Torres diagram.

The reproduction state is represented by S0, S1, S2, and S3.
State, and when -1 is input at S0, the state transits to S0 and the output data becomes 0. When 0 is input at S0, the state transits to S1 and the output data becomes 1, and when 1 is input at S1, the state transits to S2. When output data is set to 0 and 1 is input in S2, S
2, the output data is set to 0, and when 0 is input in S2, the process shifts to S3 and the output data is set to 1. When -1 is input in S3, the process shifts to S0 and the output data is set to 0. When there is an input that violates the above, the state of the violation is detected, the bit error is corrected by judging the original state that is not a violation, and the error rate for random errors is improved.

In FIGS. 4 and 5, the denominator is the input of the Viterbi decoding means and the numerator is the output of the Viterbi decoding means.

Here, in the state shown in FIG. 7, a probability distribution in which -1 to 1 are inputted when 0 should be inputted is shown. In this example, the distribution is a normal distribution.

From FIG. 7, in the transition from S0 to S1, the probability of detecting Δy is:

[0032]

(Equation 1)

[0033] In S2 → S3 transition, the probability of detecting the Δy, calculated P ₂₃ = P ₀₁ Similarly, the probability of detecting the Δy in S1 → S2 transition,

[0034]

(Equation 2)

In the transition from S2 to S2, the probability of detecting Δy is: P ₂₂ = P _{12 In the} transition of S0 → S0, the probability of detecting Δy is:

[0036]

(Equation 3)

In the transition from S3 to S0, the probability of detecting Δy is P ₃₀ = P ₀₀ . Here, if we define the metric as the negative logarithm of the probability,

[0038]

(Equation 4)

[0039]

(Equation 5)

[0040]

(Equation 6)

In the future, the metric will discuss not the absolute value but the relative value of the length, and add a fixed value to the sum of these logarithmic values, multiply the sum by a fixed value, and then compare. Each This value, and 1 _00, 1 _01, 1 _12, 1 _22, 1 _23, 1 _30, is defined as a branch metric. Each branch _{metric, 1 00 = y + 0.5,1 01} = 0,1 12 = -
y + 0.5,1 ₂₂ = −y + 0.5,1 ₂₃ = 0,1 ₃₀ = y
+0.5.

Here, at time n, states S3 and S
2, the path metrics of S1 and S0 are _{expressed as mn} (S3), m
_{n (S2), m n (} S1), when defined as _{m n (S0), m n} (S3) = m n-1 (S2) +1 23 = m n-1 (S2) m n (S2) = min _{[m n-1 (S2)} +1 22, m n-1 (S1) +1 12] = min [m n-1 (S2) -y + 0.5, m n-1 (S1) -y + 0. _{5] m n (S1) =} m n-1 (S0) +1 01 = m n-1 (S0) m n (S0) = min [m n-1 (S0) +1 00, m n-1 (S3) +1 ₃₀ ] = min [ _mn-1 (S2) -y + 0.5, _mn-1 (S1) + y + 0. 5], and these expressions can be expanded as follows.

Merge 0 m _n-1 (S2) <m _n-1 (S1) and m _n-1 (S0) <
When _mn-1 (S3), _mn (S3) = _mn-1 (S2) _mn (S2) = _mn-1 (S2) -y + 0.5 _mn (S1) = _mn-1 (S0) _mn (S0) = _mn-1 (S0) + y + 0.5, and the trilles diagram is as shown in FIG.

Merge 1 m _n-1 (S2) <m _n-1 (S1) and m _n-1 (S0) ≧
When _mn-1 (S3), _mn (S3) = _mn-1 (S2) _mn (S2) = _mn-1 (S2) -y + 0.5 _mn (S1) = _mn-1 (S0) _mn (S0) = _mn-1 (S3) + y + 0.5, and the trilles diagram is as shown in FIG.

Merge 2 m _n-1 (S2) ≧ m _n-1 (S1) and m _n-1 (S0) <
When _mn-1 (S3), _mn (S3) = _mn-1 (S2) _mn (S2) = _mn-1 (S1) -y + 0.5 _mn (S1) = _mn-1 (S0) _mn (S0) = _mn-1 (S0) + y + 0.5, and the trilles diagram is as shown in FIG.

Merge 3 _mn-1 (S2) ≥ _mn-1 (S1) and _mn-1 (S0) ≥
When _mn-1 (S3), _mn (S3) = _mn-1 (S2) _mn (S2) = _mn-1 (S1) -y + 0.5 _mn (S1) = _mn-1 (S0) _mn (S0) = _mn-1 (S3) + y + 0.5, and the trilles diagram is as shown in FIG.

Assuming that the reproduced data is y, y + 0.5 and -y + 0.5 are calculated from the input data, and a merge is determined.

As the merge 0, the path metric S2 is smaller than the path metric S1 and the path metric S
When 0 is smaller than the path metric S3, the path metric S3 is set to the path metric S2, the path metric S2 is set to the path metric S2-y + 0.5, the path metric S1 is set to the path metric S0, and the path metric S0 is set to the path metric S0 + y + 0.5.

As a merge 1, the path metric S
2 is smaller than the path metric S1 and the path metric S0 is larger than or equal to the path metric S3, the path metric S3 is the path metric S2, the path metric S2 is the path metric S2-y + 0.5, and the path metric S1 is the path metric S0. , Path metric S0 is set to path metric S3 + y + 0.5.

Further, as a merge 2, the path metric S2 is greater than or equal to the path metric S1, and
When the path metric S0 is smaller than the path metric S3, the path metric S3 is set to the path metric S2, the path metric S2 is set to the path metric S1-y + 0.5, the path metric S1 is set to the path metric S0, and the path metric S0 is set to the path metric S0 + y + 0.5. .

Further, as a merge 3, the path metric S2 is greater than or equal to the path metric S1, and
When the path metric S0 is larger than or equal to the path metric S3, the path metric S3 is changed to the path metric S3.
2. Path metric S2 is converted to path metric S1-y +
0.5, path metric S1 is path metric S0, and path metric S0 is path metric S3 + y + 0.5.

Thereafter, when the following state occurs, the paths are merged, that is, the paths are unified, and a corresponding output sequence is obtained. FIG. 6 shows an example in which a path merge occurs.

The paths merge when the next state occurs. 1. Three consecutive states are (Merge 0 or Merge 2) →
At the time of merge 2 → merge 2, the path merges with S0. 2. 3 consecutive states are (Merge 0 or Merge 1) →
When merge 1 → merge 1, the path merges with S2.

In FIG. 6, the path merges with S0 at the position of *. By path merging, the paths in the past are unified, and according to the transition of the paths, output 0 is output for states S0 and S2 and output 1 is output for states S1 and S3 to obtain an output data sequence. In FIG. 6, “0100
Output data of 10 ″ is obtained.

Next, an example of the configuration of the waveform shaping means of the present invention will be described with reference to FIG.

FIG. 12 shows an example of the circuit configuration of the waveform shaping means. The ternary input 81 is a D-flip-flop 85
And is input to one of the selection circuits 88. At the same time, the offset value input 82 (AO
F3 to AOF0) and input to the other side of the selection circuit 88. The selection circuit 88 selects whether to output the input signal as it is as the output signal 90 or to output the addition result of the addition circuit 86 as the output signal 90 by using the selection signal 89 output from the selection signal calculation circuit 87 and outputs it. In the selection signal calculation circuit 87, 83 (CGP0) and 84 (CG
The selection signal 89 is calculated by changing the calculation method according to P1).

That is, when 83 is 0 and 84 is 0,
The selection signal is always 0, and the selection circuit 88 always outputs D-
The output of the flip-flop 85 is output as it is.

When 83 is 0 and 84 is 1, the MS of 82
When B continues to 1,0, that is, 82 is indicated by a 2 'complement, and when the polarities continue to-and +, the addition result of the addition circuit 86 is output in two consecutive samples. The signal is output as a signal 90, otherwise, D-
The output of flip-flop 85 is output as output signal 90 as it is.

When 83 is 1 and 84 is 0, the MS of 82
When B continues to 0 and 1, that is, when the polarities continue to + and-, the addition result of the adding circuit 86 is output as an output signal 90 in two consecutive samples, and otherwise the D-flip-flop is used. Output signal 9 as it is
Output as 0.

When 83 is 1 and 84 is 1, MS of 82
When B continues to 1,0 or 0,1, that is, when the polarity is inverted, the addition result of the addition circuit 86 is output as an output signal 90 in two consecutive samples. The output of flip-flop 85 is output as output signal 90 as it is.

Thus, the above-described offset addition can be realized. FIG. 13 shows 83 (CGP
0), 84 (CGP1), 82 (AOF3 to AOF0)
Is shown.

The above is the transition point from the "H" level to the "L" level, or from the "L" level to the "H" level, from the "H" level to the "L" level or from the "L" level to the "H" level. If a shift to the level occurs and the amount of the shift is known in advance, the input data is corrected based on predetermined data.

On the other hand, taking advantage of the characteristics of the recording code,
It is also possible to correct the data by determining the deviation from the reproduction signal.

That is, in the case of the 1,7 code shown in FIG. 14, the shortest continuous bit length is 2, and the longest continuous bit length is 8. Therefore, when PR (1,1) ternary detection is performed,
“+1” to “−1” are consecutive one bit or more and consecutive 7
Must be less than or equal to a bit. Therefore, when "+1" or "-1" appears in the reproduced data for 0 bits, or when 8 or more consecutive bits appear, the offset value is determined by determining this. Here, the appearance of 0 bits means that
For example, when "-1, -1, -1,0" is input,
"0, -1, -1" without shifting to "+1, 0, -1"
It is to move to.

When the 0 bit appears, when it appears on the “H” level side, it is determined that an offset has occurred from the “H” level side to the “L” level side, and a predetermined positive Add the offset value. Further, when it appears on the “L” level side, it is determined that an offset has occurred from the “L” level side to the “H” level side, and a predetermined negative offset value is added to all the reproduced data. Thereafter, data detection determination is performed by Viterbi decoding.

When 8 bits or more appear, it is set to “H”.
When appearing on the level side, it is determined that an offset has occurred from the "L" level side to the "H" level side, and a predetermined negative offset value is added to all the reproduced data. Also,
When appearing on the “L” level side, it is determined that an offset has occurred from the “H” level side to the “L” level side, and a predetermined positive offset value is added to all the reproduced data. Thereafter, data detection determination is performed by Viterbi decoding.

Further, in the present embodiment, it is possible to calculate the amount of deviation and adaptively switch the offset value.

"+1" or "-1" is 0 in the reproduction data.
The case where a bit appears is shown. For example, PR (1,
1) "-1, -1, -0.6, +0, 4,
It is assumed that reproduced data of −0.4, −1, −1 ″ is obtained. At this time, the ternary judgment threshold level is set to −
Assuming 0.5 and +0.5, the ternary determination result is “−1,
−1, −1, 0, 0, −1, −1 ”, which means that 0 bits of“ +1 ”appear in the reproduced data described above, and therefore an offset is added to the + side.
What is necessary is just to add “+0.2”. At this time, there is a method of adding to all reproduction data and a method of adding to only change point data. The method of adding only to the change point data includes a method of adding only to the rising change point data, a method of adding only to the falling change point data, and a method of adding to both the rising / falling change point data.

In the method of adding to all data, “−0.8,
−0.8, −0.4, +0.6, −0.2, −0.8,
−0.8 ”, and the ternary judgment result is“ −1, −1,
0, 1, 0, -1, -1 ".
In the method of adding to the falling transition point data, as a result of the addition, it is added only to the samples before and after the polarity inversion, and "-1, -1, -0.4, +0.6, -0.2,-
1, -1 ", and the same ternary determination result is obtained. Now, for the offset data of 0.2, the minimum value is calculated for each sample so that 0 bit does not appear in the reproduced data. Alternatively, if the value is added for each sample, the average value of the minimum value that eliminates the appearance of 0 bits may be calculated and added.

Next, "+1" or "-"
1 "appears when 8 bits appear. For example, P
"-0.8, -0.4, +
0.6, +1, +1, +1, +1, +1, +1, +1,
+0.4, -0.7 ". At this time, the ternary judgment threshold level is set to -0.5,
Assuming +0.5, the ternary judgment result is “−1, 0, +
1, + 1, + 1, + 1, + 1, + 1, + 1, + 1,0,
In this case, an offset is added to the-side, and in this example, "-0.2" may be added. There is a method of adding to the reproduction data, and a method of adding only to the data of the change point.The method of adding only to the change point data includes a method of adding only to the rising change point data, a method of adding only to the falling change point data, There is a method of adding to both the falling transition point data.

In the method of adding to all data, "-1, 0,
−0.6, +0.4, +0.8, +0.8, +0.8,
+ 0.8 + 0.8, +0.8, +0.8, +0.2,-
0.9 ", and the ternary judgment result is" -1, -1, 0,
1, 1, 1, 1, 1, 1, 1, 0, -1 ". In the method of adding to the rising / falling transition point data," -0.8, -0.6, +0.4 , + 1, + 1,
+1, +1, +1, +1, +1, +0.2, -0.9 "
And the same ternary determination result is obtained. Well, this-
Although the offset data is 0.2, the minimum value may be calculated and added to each sample so that 8 bits do not appear in the reproduced data, or if the value is added to each sample, the 8 bits appear. The average of the minimum value that disappears may be calculated and added.

In this way, by adding an appropriate value as an offset, a good bit error rate can be obtained when combined with Viterbi decoding.

The present invention is, of course, applicable to the circuit shown in FIG.
The same applies to codes. In that case, the shortest continuous bit length is 3, and the longest continuous bit length is 8. Therefore, when PR (1,1) ternary detection is performed, “+1” or “−1” must be two or more consecutive bits and seven or less consecutive bits. Therefore, "+1" is added to the reproduction data.
Alternatively, when "-1" appears 1 bit or less, or 8 or more consecutive bits appear, it is determined and an offset value is obtained. Here, as for the portion of "1 bit or less", "0 bit or less" in the above-mentioned 1,7 code may be extended to "1 bit or less", and for the "continuous 8 bit or more" portion, A decoding method similar to the 1,7 code can be applied.

Further, not only 1,7 code and 2,7 code,
The present invention can be applied to all codes having the shortest continuous bit length and the longest continuous bit length, that is, run-length limited codes.

In the NRZ code, after the code conversion, the minimum value −1 of the sequence of “0” to “1” is d and the maximum value −1 is k, and in the NRZI code, the code is “1” after the code conversion. “1”
The minimum value of “0” between d is d and the maximum value is k. Here, when the PR (1,1) ternary detection is performed, "+1"
Or "-1" must be greater than or equal to consecutive d bits and less than or equal to consecutive k bits. Therefore, "+1" is added to the reproduction data.
When "-1" appears in (d-1) bits or less or consecutive (k + 1) bits or more, it is determined to determine the offset value. Here, "(d-1)
As for the portion "less than bits", "0 bit or less" in the above described 1,7 code may be extended to "(d-1) bits or less", and the "continuous (k + 1) bits or more" portion is described above. "8 bits or more" in the 1,7 code may be extended to "(k + 1) bits or more".

Further, the equalization method is not limited to PR (1, 1). For the recording codes d and k, for example, PR
(1) Perform binary detection and set the equalized output signal to "+1",
When "-1" is set, "+1" to "-1" must be continuous (d + 1) bits or more and continuous (k + 1) bits or less. Therefore, when "+1" to "-1" appear in the reproduction data for d bits or less or consecutive (k + 2) bits or more, it is determined to determine the offset value. Here, with respect to the portion "d bit or less", the "0 bit or less" in the above described 1,7 code may be extended to "d bit or less", and the "continuous (k + 2) bit or more" portion is described above. 1,
"8 bits or more" in the 7 code may be extended to "(k + 2) bits or more". The same applies to other equalization schemes.

[0077]

As described above, according to the present invention, an error from the "H" level to the "L" level occurs at the transition point from the "H" level to the "L" level in the reproduced data. In a recording / reproducing system which is easy to perform, a transition from the "H" level to the "L" level is detected, a predetermined positive offset value is added to data before and after the transition, and then data detection determination is performed. Alternatively, in a recording / reproducing system in which an error easily transitions from an “L” level to an “H” level or from an “L” level to an “H” level, an error is likely to occur in data before and after the transition. By performing data detection determination after adding a predetermined offset value in the direction and in the opposite direction, the effect of the subsequent Viterbi decoding can be greatly exerted, and the bit error rate can be minimized.

[Brief description of the drawings]

FIG. 1 is a diagram illustrating an error occurrence pattern improved by the present invention.

FIG. 2 is a block diagram showing an embodiment of a configuration of a recording / reproducing system to which the present invention is applied.

FIG. 3 is a diagram showing an example of a timing chart of the recording / reproducing system of FIG. 2;

FIG. 4 is a diagram showing a state transition diagram in partial response, PR (1,1) +3 value detection.

FIG. 5 is a diagram showing a trilles diagram in partial response, PR (1,1) +3 value detection.

FIG. 6 is a diagram showing an example of path merge in partial response, PR (1,1) +3 value detection.

FIG. 7 is a diagram showing a detection probability in a partial response PR (1, 1).

FIG. 8 is a diagram showing a trilles diagram at the time of merge 0 in the present embodiment.

FIG. 9 is a diagram showing a trilles diagram at the time of merge 1 in the present embodiment.

FIG. 10 is a diagram showing a trilles diagram at the time of merge 2 in the present embodiment.

FIG. 11 is a diagram showing a trilles diagram at the time of merge 3 in the present embodiment.

FIG. 12 is a block diagram showing an embodiment of the circuit configuration of the waveform shaping means of the present invention.

FIG. 13 is a diagram illustrating the content of a signal input to the waveform shaping means of the present invention.

FIG. 14 is a diagram showing a code conversion table of 1,7 codes.

FIG. 15 is a diagram showing a code conversion table of 2,7 codes.

FIG. 16 is a diagram illustrating an example of a waveform shaping waveform.

[Explanation of symbols]

 81 Three-value input 82 Offset value input 83 CGP0 84 CGP1 85 D-flip-flop 86 Addition circuit 87 Selection signal calculation circuit 88 Selection circuit 89 Selection signal 90 Output signal

──────────────────────────────────────────────────続 き Continued on front page (58) Field surveyed (Int. Cl. ⁷ , DB name) G11B 20/14 341

Claims (7)

(57) [Claims] [Claim 1] using the characteristic of the recording code, it determines that the inversion interval below the shortest recording wavelength in the reproduced data appeared, when the inversion interval appeared to "L" level side
From "H" level side to the "L" level side to determine the offset occurs by adding the offset value of a predetermined positive in all playback data, when the inversion interval appeared to "L" level side "L" after the offset from the level side to "H" level side by adding a negative offset value given to the entire reproduction data determined to have occurred, the data detection and performing data detection determination Playback method. 2. A shortest recording wavelength determined from a recording code in a data detection method used in a recording / reproducing system in which an error tends to occur in the same direction at a level transition conversion point in the same direction with respect to reproduced data. Is detected, and an offset value is added in the reverse direction of the level direction, which is likely to occur at a level transition change point where this error is likely to occur, so that such an inversion interval does not appear,
This offset value is retained, and then the same offset value is similarly added at the level transition change point, and even if an inversion interval shorter than the shortest recording wavelength determined from the recording code still appears, the inversion below the shortest recording wavelength detects that the interval has emerged as inversion interval below the shortest recording wavelength does not appear, to calculate the offset value again, claim 1 Symbol, characterized in that performing data detection determination by adding the offset value A data detection / reproduction method characterized in that control is performed using the data detection / reproduction method described above. Wherein using the characteristic of the recording code, it determines that the inversion interval exceeding the longest wavelength in the reproduced data appeared, when the inversion interval appeared to "H" level side
"L" level to the side it is determined that "H" offset to the level side has occurred by adding a negative offset value given to the entire reproduction data when the inversion interval appeared to "L" level side "H" after the offset from the level side to the "L" level side by adding the determination to the offset value of the predetermined positive to all the reproduction data to be generated, the data detection and performing data detection determination Playback method. Against 4. A reproduced data, the data detection method of error at the level transition transformation point in the same direction to the same direction of the level is used for easily recording and reproducing system occurs, longest record is determined from the recorded code It detects that an inversion interval exceeding the wavelength has appeared, and adds an offset value in the opposite direction of the level direction that is likely to occur at a level transition change point where this error is likely to occur so that such an inversion interval does not appear,
This offset value is retained, and thereafter, the same offset value is similarly added at the level transition change point, and even if an inversion interval exceeding the longest recording wavelength determined from the recording code still appears, the inversion interval exceeding the longest recording wavelength as but do not appear to calculate the offset value again, and controls using the data detection reproducing method according to claim 3 Symbol mounting and performing data detection determined by adding the offset value Data detection and reproduction method. 5. Reproducing digitally recorded data,
In a data detection / reproduction device for outputting a reproduced signal through an equalizing means and a decoding means and outputting an output signal, the reproduced data output of the equalizing means is changed from "H" level to "L" level or from "L" level to "H". A waveform shaping means for detecting a change point at which the level shifts to the level, adding a predetermined offset value to the reproduction data before and after the level shift, and outputting the added reproduction data to the decoding means before and after the level. Characteristic data detection and reproduction device. 6. Using the characteristics of the recording code, it determines that the inversion interval below the shortest recording wavelength in the reproduced data appeared, when the inversion interval appeared to "H" level side
From "H" level side to the "L" level side to determine the offset occurs by adding the offset value of a predetermined positive in all playback data, when the inversion interval appeared to "L" level side Bei means for performing "L" level to side "H" added to the data detection determining a predetermined negative offset value to all the reproduced data to determine the offset occurs to the level side
Data detection reproducing apparatus characterized by was e. 7. Using the characteristics of the recording code, it determines that the inversion interval exceeding the longest wavelength in the reproduced data appeared, when the inversion interval appeared to "H" level side
"L" level to the side it is determined that "H" offset to the level side has occurred by adding a negative offset value given to the entire reproduction data when the inversion interval appeared to "L" level side "H" "from level side by adding the" L "offset value of a predetermined positive it is determined that the offset occurs in the entire reproduction data to the level side means for performing data detection determination
Data detection reproducing apparatus characterized by comprising.