JP2008287839A - Digital data reproducing unit - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a digital data reproducing unit with suppressed contention between an adaptive equalization technique and an adaptive maximum likelihood decoding technique. <P>SOLUTION: The digital data reproducing unit includes: an equalization means for outputting equalized signals by applying waveform equalization processing on each digital signal, based on a predetermined coefficient; a maximum likelihood decoding means for outputting decoded signals by applying a maximum likelihood decoding processing on the above equalized signals being outputted from the equalization means, based on reference levels; an ideal waveform generation means for outputting one of the above reference levels according to the decoded signals outputted from the maximum likelihood decoding means; an optimization means for calculating an error between the ideal waveform signal outputted from the ideal waveform generation means and each equalized signal outputted from the equalization means, and for optimizing the predetermined coefficient of the equalization means so as to minimize the error; and an optimization means for optimizing the reference levels of the maximum likelihood decoding means. The above reference level optimization means performs optimization of the reference levels by changing at least one reference level among the reference levels, without changing at least two reference levels. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

The present invention relates to a digital data reproducing apparatus, and more particularly to PRML (Partial Response and M
The present invention relates to a digital data reproducing apparatus that performs a decoding process using an aximum likelihood method.

In recent years, information recording / reproducing apparatuses that perform recording / reproducing processing on a recording medium such as an optical disc such as a DVD (Digital Versatile Disc) have become widespread, and higher recording density has been demanded in various systems. On the other hand, for example, there is a PRML method as a method for recording and reproducing information on an optical disk or the like.

In this PRML method, an ideal waveform is obtained by performing convolution integration between a decoding result and a PR (Partial Response) class, an equalization error between the ideal waveform and the equalizer output is calculated, and the equalization error is minimized. As described above, the processing parameters of the reproduction signal (for example, the tap coefficient, gain, offset, etc. of the equalizer) are changed using an LMS (Least Mean Square) algorithm or the like.

On the other hand, regarding the ML (maximum likelihood decoding technique) of the PRML method, Patent Document 1 describes the concept of adaptive Viterbi detection. However, a technique relating to a combination of the PR, for example, adaptive equalization processing and ML control is not disclosed.

The difficulty of this combination is described in Patent Document 2 regarding the combination technique of an adaptive equalizer and an adaptive Viterbi detector. As a problem, optimization of the gain by the AGC circuit 16 according to the target waveform signal T is performed simultaneously with optimization of the tap coefficients C0 to C6 of the FIR filter 17 and optimization of the reference levels R0 to RN of the Viterbi decoder. In this case, changes in tap coefficients C0 to C6 and reference levels R0 to RN of the Viterbi decoder may cause unnecessary changes in the gain that should have been optimized. It operates stably, and in the worst case, there is a problem that the control system diverges due to competition.
Japanese Patent No. 2960436 Japanese Patent No. 3529767 (page 9, paragraph 0081)

An object of the present invention is to provide a simple and high-performance digital data reproducing apparatus that suppresses competition between adaptive equalization technology and adaptive maximum likelihood decoding technology.

In order to solve the above problems, the digital data reproducing apparatus of the present invention is a partial response a
nd Maximum Likelihood detection digital data playback device,
An equalization unit that performs waveform equalization processing based on a predetermined coefficient and outputs an equalized signal; and a maximum likelihood decoding process based on a reference level for the equalized signal output by the equalization unit, and a decoded signal A maximum likelihood decoding means for outputting, an ideal waveform generating means for outputting one of the reference levels in accordance with the decoded signal output from the maximum likelihood decoding means, and an ideal waveform signal output by the ideal waveform generating means And an optimization means for calculating an error from the equalization signal output by the equalization means and optimizing the predetermined coefficient of the equalization means so as to minimize the error, and a reference level of the maximum likelihood decoding means, The reference level optimizing unit optimizes the reference level without changing at least one reference level and changing at least two reference levels. That And butterflies.

ADVANTAGE OF THE INVENTION According to this invention, the digital data reproducing | regenerating apparatus which suppressed the competition with an adaptive equalization technique and an adaptive maximum likelihood decoding technique is obtained.

  Examples of the present invention will be described below.

A first embodiment of the present invention will be described with reference to FIGS. FIG. 1 is a block diagram showing an example of a main part of an optical disk reproducing apparatus according to an embodiment of the present invention. FIG. 2 is a block diagram showing an example of an optical disk reproducing apparatus according to the embodiment of the present invention. FIG. 3 is a diagram illustrating configurations of the adaptive equalizer and the adaptive learning device. FIG. 4 is a diagram showing the configuration of the Viterbi decoder according to the embodiment of the present invention. FIG. 5 shows (1, 7) of the maximum likelihood decoding processing according to the present invention used in the embodiment.
It is a figure which shows the state transition in a RLL modulation system and PR (1221) ML decoding system. FIG. 6 is a trellis diagram corresponding to PR (1221) having a minimum run length of 1. FIG. 7 is a diagram illustrating a configuration of the reference value generator. FIG. 8 is a diagram showing the configuration of the equalization error generator. Moreover, FIG. 9 is a figure which shows the effect of a present Example.

[Embodiment of the Invention]
Hereinafter, a digital data reproducing apparatus according to the present invention will be described in detail with reference to the drawings, taking an optical disk recording / reproducing apparatus as an example. In this embodiment, an example of an optical disk recording / reproducing apparatus has been described. However, the target recording medium is not limited to an optical disk, and for example, if a recording medium such as a magneto-optical disk is used, This produces a working effect.

[Configuration of optical disk playback device]
An embodiment of an optical disk reproducing apparatus according to the present invention will be described with reference to the accompanying drawings.
FIG. 1 is a diagram showing a configuration example of an optical disc reproducing apparatus according to an embodiment of the present invention.

An optical disk reproducing apparatus for reproducing data recorded on the optical disk medium 1 includes a preamplifier unit 2, an A / D converter 3, an adaptive equalizer 4, a maximum likelihood decoder 5, an adaptive learner 6, and an equalization error generator 7. ,
A reference value generator 8 and a demodulator 9 are provided. The output of the demodulator 9 is output to an error correction circuit (not shown).

A reproduction signal from the optical disk medium 1 is amplified by the preamplifier unit 2 and converted into a multi-value digital signal by the A / D converter 3. The digitized multilevel reproduction signal is converted into an adaptive equalizer 4.
In FIG. 4, waveform equalization processing is performed. The multilevel reproduction signal after waveform equalization is
It is reproduced as binary data of 1 'or' 0 '. The binary data output from the maximum likelihood decoder 5 is demodulated by the demodulator 9. In the demodulator 9, for example, ETM (Eight
to Twelve Modulation) is performed. The adaptive learning device 6 adjusts the equalization characteristics in the adaptive equalizer 4 according to the adaptive processing described later. The reference value generator 8 adjusts the reference value used by the maximum likelihood decoder 5 from the binary data string obtained by the maximum likelihood decoder 5 and the output of the equalization error generator 7. The equalization error generator 7 generates an equalization error from the output of the reference value generator 8 and the output of the adaptive equalizer 4.

<Configuration of Optical Disc Device According to the Present Invention>
FIG. 1 is a block diagram illustrating an example of a main part of the information reproducing apparatus according to the present invention, and FIG. 2 is a block diagram illustrating an example of the overall configuration.
(Basic Configuration of Optical Disc Device) In FIG. 2, an optical disc device A according to the present invention performs data recording or data reproduction on an optical disc D. The optical disc apparatus A has a tray 32 for transporting an optical disc D accommodated in a disc cartridge, a motor 33 for driving the tray, a clamper 34 for holding the optical disc D, and the optical disc D held thereby by a predetermined rotational speed. And a spindle motor 35 that is rotated at the same time. Further, a CPU 46 that controls the entire operation as a control unit, a ROM 47 that stores a basic program of the control operation, and a RAM 48 that stores each control program and application data in a rewritable manner via a control bus. Connected. Further, the feed motor 36 that transports the pickup PU, the focus / tracking actuator driver / feed motor driver 40 that controls the focus and tracking of the pickup, and the spindle motor 35 are connected to the control unit such as the CPU 46, respectively. A spindle motor driver 41 for driving the tray motor and a tray motor driver 42 for driving the tray motor are provided.

Furthermore, a preamplifier 12 connected to the pickup head PUH and amplifying the detection signal;
The servo amplifier 38 further includes a servo seek control unit 39 that supplies a seek signal for performing a seek operation to the driver. Pickup head PUH and preamplifier 1
2. A data processing unit 1 connected to the servo seek control unit 39 and the like for processing the detection signal and the recording signal, and a RAM 43 for storing data used for these various processes are provided. In order to transmit / receive a signal from the data processing unit 1 to / from an external device, an interface control unit 45 is provided with a RAM 44.

In such an optical disc apparatus, in the present invention, an optimum equalization signal required for the Viterbi decoding process is supplied by using the data processing unit 1 including the configuration shown in FIG. That is, the target waveform signal T is generated from the ideal signal I, and thereby the FIR filter

The parameters for the reproduction processing such as the tap coefficient of the data are optimized. As a result, by performing equalization close to the 2T signal level of the ideal waveform beyond the 2T signal level necessary for the Viterbi decoding process in the conventional apparatus, noise is unnecessarily amplified and a decoding error occurs. It solves the problem.

(Basic operation of optical disk device)
The optical disk apparatus provided in the embodiment of the present invention having such a configuration performs the reproducing process and the recording process of the optical disk as follows. That is, when the optical disc D is loaded into the optical disc apparatus A, the control of the optical disc D recorded in the control data zone in the emboss data zone of the lead-in area of the optical disc D is performed using the pickup head PUH and the data processing unit 1. Information is read and supplied to the CPU 46.

In the optical disk apparatus A of the present invention, laser control (not shown) is performed under the control of the CPU 46 based on operation information by a user operation, control information of the optical disk D recorded in the control data zone in the optical disk, current status, and the like. Energized by the unit to generate a laser beam.
The generated laser beam is converged by the objective lens 31 and irradiated onto the recording area of the disc. As a result, data is recorded in the storage area of the optical disc D (mark string generation:
The data is recorded on the optical disk D by the interval between the variable length marks and the length of each variable length mark), or the intensity of light corresponding to the stored data is reflected and detected, This data is reproduced.

In FIG. 2, the setting of the laser control unit included in the pickup head PUH is set by the data processing unit 1, but the setting erases the reproduction power for obtaining the reproduction signal RF, the recording power for recording data, and the data. Different in erasing power. The laser beam has three levels of power, ie, reproduction power, recording power and erasing power, and the semiconductor laser unit is energized by the laser control unit so that the laser beam of each power is generated. .

The laser control unit includes a resistor and a transistor (not shown), and a power supply voltage is applied to the resistor, the transistor, and a semiconductor laser as a semiconductor laser unit. As a result, the amplification factor varies depending on the base current of the transistor, different currents flow through the semiconductor laser oscillator, and laser beams with different intensities are generated. Here, recording waveform compensation is performed according to the characteristics of each optical disk, laser power is generated according to the recording waveform pulse W output from the recording waveform generation circuit 11, and recording processing onto the optical disk is performed. It has become.

In addition, the optical disk D is stored directly or in a disk cartridge and transported into the apparatus by a tray 32 so that the optical disk D is disposed to face the objective lens 31. A tray motor 33 for driving the tray 32 is provided in the apparatus. Also,
The loaded optical disk D is rotatably held by a clamper 34 on a spindle motor 35 and is rotated at a predetermined rotational speed by the spindle motor 35.

The pickup head PUH has a photodetector (not shown) for detecting a laser beam therein.
have. This photodetector detects the laser beam reflected by the optical disc D and returned via the objective lens 31. The detection signal (current signal) from the photodetector is converted into current / voltage.

The signal is converted into a voltage signal by the I / V unit, and this signal is supplied to the preamplifier 12 and the servo amplifier 34.
To be supplied. From the preamplifier 12, a signal for reproducing data in the header portion and a signal for reproducing data in the recording area are output to the data processing unit 1. Servo signal from servo amplifier 34 (
Track error signal and focus error signal) are output to the servo seek control unit 39.

Here, as a method of optically detecting the focus shift amount, for example, there are the following astigmatism method and knife edge method.
Astigmatism method, that is, an optical element (not shown) for generating astigmatism is arranged in the detection optical path of the laser beam reflected by the light reflecting film layer or the light reflecting recording film of the optical disc D, and the photodetector This is a method of detecting a change in the shape of the laser beam irradiated on the top. The light detection area is divided into four diagonal lines.

Has been. A focus error detection signal (focus signal) is obtained by taking the difference between the diagonal sums in the servo seek control unit 39 with respect to the detection signal obtained from each detection region.

This is a knife edge method, that is, a method of arranging a knife edge that shields a part of the laser light reflected by the optical disk D asymmetrically. The light detection area is divided into two, and a focus error detection signal is obtained by taking a difference between detection signals obtained from the respective detection areas.

Usually, either the astigmatism method or the knife edge method is employed. Optical disc D
Has a spiral or concentric track, and information is recorded on the track. Information is reproduced or recorded / erased by tracing the focused spot along the track. In order to stably trace the focused spot along the track, it is necessary to optically detect the relative positional deviation between the track and the focused spot.

Generally, there are the following phase difference detection method, push-pull method, twin spot method, etc. as the track deviation detection method.
A differential phase detection method, that is, a change in intensity distribution on the photodetector of the laser beam reflected by the light reflecting film layer or the light reflecting recording film of the optical disc D is detected. The light detection area is divided into four diagonal lines. For the detection signal obtained from each detection area,
In the servo seek control unit 39, the phase difference between the diagonal sums is taken and the track error detection signal (
Tracking signal).

In the push-pull method, that is, in this method, a change in intensity distribution on the photodetector of the laser beam reflected by the optical disc D is detected. The light detection area is divided into two, and a track error detection signal is obtained by taking a difference between detection signals obtained from the respective detection areas.

Twin-spot method, that is, reflection of ± first-order diffracted light irradiated on the optical disk D by arranging a diffraction element in the light transmission system between the semiconductor laser element and the optical disk D to divide the light into a plurality of wavefronts. Detect changes in light intensity. Separately from the light detection area for detecting the reproduction signal, a light detection area for individually detecting the reflected light amount of the + 1st order diffracted light and the reflected light amount of the −1st order diffracted light is arranged, and a track error detection signal is obtained by taking a difference between the respective detection signals. Get.

By such focus control and track control, the focus signal, tracking signal, and feed signal are sent from the servo seek control unit 39 to the focus and tracking actuator driver and feed motor driver 40, and the objective lens 31 is controlled by the driver 40 with focus servo control. And tracking servo control. Further, an energizing signal is supplied from the driver 40 to the feed motor 36 in response to the access signal, and the pickup head PUH is transported.

The servo seek control unit 39 is controlled by the data processing unit 1. For example, an access signal is supplied from the data processing unit 1 to the servo seek control unit 39 to generate a feed signal.

Further, the spindle motor driver 41 and the tray motor driver 42 are controlled by a control signal from the data processing unit 1, the spindle motor 35 and the tray motor 33 are energized, the spindle motor 35 is rotated at a predetermined rotational speed, and the tray motor 33 is driven. Will properly control the tray.

The reproduction signal RF corresponding to the header data supplied to the data processing unit 1 is CP.
Supplied to U46. As a result, the CPU 46 determines the sector number as an address of the header portion based on the reproduction signal RF, and compares it with the sector number as an address to access (record data or reproduce recorded data). It has become.

The reproduction signal RF corresponding to the data in the recording area supplied to the data processing unit 1 is RA
The necessary data is stored in M48, the reproduction signal RF is processed by the data processing unit 1 and supplied to the interface control unit 45, and the reproduction processing signal is supplied to an external device such as a personal computer.
<Viterbi Decoding Processing with Optimization of Playback Processing Parameters According to the Present Invention>
Next, the process of optimizing the playback process parameters, which is a feature of the present invention, will be described in detail with reference to the drawings.
[Adaptive equalization using LMS (Least Mean Square) method]
The operation of an adaptive equalizer widely used in the PRML technology will be described with reference to FIG. FIG. 3 is a block diagram showing details of the adaptive equalizer 4 and the adaptive learning circuit 6. As shown in FIG. 3, the adaptive equalizer 4 includes delay circuits 201 and 202, multiplication circuits 203, 204, and 20.
5, a delay circuit 206, and addition circuits 207 and 208 are provided. The adaptive learning circuit 6 includes registers 209, 210 and 211, coefficient update circuits 212, 213 and 214, and a delay circuit 215.
, 216. The delay circuit 206 delays the input signal by a predetermined time (clock) and outputs it. The delay circuits 201 and 202 delay the input signal by one clock and output it. Multipliers 203, 204, and 205 output the product of two input values. Adder circuit 207,
208 outputs the sum of two input values. In this FIG. 2, 3 taps using three multipliers.
Although an example of the digital filter is shown, the basic operation is the same even if the number of multipliers is changed. Here, only the case of 3 tap will be described.

If the input signal of the adaptive equalizer 4 at time k is x (k) and the multipliers input to the multipliers 203, 204 and 205 are c1, c2 and c3, respectively, the output Y (k) of the adaptive equalizer 4
Can be expressed as:

Y (k) = x (k) * c1 + x (k-1) * c2 + x (k-2) * c3
... (1)

For an adaptive equalizer output Y (k) at time k, an equalizer output that does not include noise is Z (k), and an equalization error output by an equalization error generator 7 described later is E (k). And These have the following relationship.

E (k) = Y (k) −Z (k)
... (2)

The adaptive learning circuit 6 updates the coefficient of each multiplier according to the following equation.
c1 (k + 1) = c1 (k) −α * x (k) * E (k)
... (3)

c2 (k + 1) = c2 (k) −α * x (k−1) * E (k)
(4)

c3 (k + 1) = c3 (k) −α * x (k−2) * E (k)
... (5)

Α in the expressions (3) to (5) is an update coefficient, and a positive small value (for example, 0.01) is set.
The delay circuits 215 and 216 output the input signal with a delay of one clock. Coefficient update circuit 2
12, the calculation shown in Expression (3) is performed to update the input multiplier to the multiplier 203. The update result is stored in the register 209. The coefficient update circuit 213 updates the input multiplier to the multiplier 204 by performing the calculation shown in Expression (4). The update result is stored in the register 210. The coefficient update circuit 214 updates the input multiplier to the multiplier 205 by performing the calculation shown in Equation (5). The update result is stored in the register 211. As described above, adaptive learning of the adaptive equalizer 4 is performed.

Although the adaptive equalization by the LMS method has been described above, this is an example of the adaptive equalization method, and the present invention is not limited to the LMS method, and can be applied to the RLS method and other adaptive equalization methods. .

[Table Lookup (TLU) type non-linear response Viterbi detection]
Next, a general configuration of the maximum likelihood decoder 5 based on the Viterbi algorithm will be described. FIG. 4 is a diagram showing a configuration of the maximum likelihood decoder 5. The maximum likelihood decoder 5 includes a branch metric unit 121, a comparison / selection unit 122, a metric register 124, and a path memory 123. The branch metric unit 121 generates branch metrics BM_0 to BM_F from the output Y (k) of the adaptive equalizer 4 and the output Z (k) of the reference value generator 8. The comparison / selection unit 122 calculates a cumulative metric from the output of the branch metric unit 121 and the output of the metric register 124 and performs comparison / selection processing. The selected cumulative metric is temporarily held in the metric register 124 and used for the comparison / selection process at the next time. In the path memory 123, the past comparison selection result is held,
The final binary data A (k) is output.

FIG. 5 is a diagram showing state transitions in the (1, 7) RLL modulation method and the PR (1221) ML decoding method. FIG. 6 is an example of a trellis diagram corresponding to PR (1221) with a minimum run length of 1 showing the state transition of FIG. 5 in time series. If the minimum run length is 1 and PR (1221), the number of internal states is 6, which are S0, S1, S3, S4, S6, and S7, respectively. The state at time k and the state at time k + 1 are branches.
And connect each branch to Z_0, Z_1, Z_3, Z_6, Z_7, Z_8, Z_9, Z_C,
Let Z_E and Z_F. As shown in FIG. 6, these are the states S0 at the time k + 1 from the state S0 at the time k.
The branch connected to is Z_0, and similarly the branch connected from S0 to S1 is Z_1, and S1 to S3
The branch connected to is Z_3, the branch connected from S3 to S6 is Z_6, the branch connected from S3 to S7 is Z_7, the branch connected from S4 to S0 is Z_8, and the branch connected from S4 to S1 is Z_9 Yes, the branch from S6 to S4 is Z_C, and the branch from S7 to S6
Z_E, and the branch connecting S7 to S7 is Z_F.

In the case of a partial response of class (1221), the ideal channel response Z can be obtained by convolution of input binary data string A and PR class. That is,

Z = [1221] * A
(6)

However, (*) is an operator representing convolution. Since the input binary data string A is composed of continuous 4-bit binary data and the minimum run length is 1, there are the following 10 combinations in transposition.

A = [0000], [0001], [0011], [0110], [0111], [1000], [1001], [1100], [1110], [11
11] (7)

Since there are 10 types of input binary data strings A, the ideal channel response Z is originally 10 types, but since the results of the convolution operation are duplicated, 7 types of values are taken.
The branch metric unit 121 calculates the distance difference between the channel output Y (k) and the ideal channel response Z as the branch metric BM at each time. That is,
BM_x (k) = (Y (k) −Zx) ²
(8)

The result of calculating the branch metric BM from the equation (8) for each branch of the equation (6) is the output of the branch metric unit 121.
Here, consider a case where linearity does not hold (non-linearity) in the calculation of Z in Equation (6). Even in this case, it is assumed that an ideal channel response Z can be obtained by an input binary data string A having a finite length. For the sake of brevity, it is assumed that an ideal channel response Z is determined by a continuous 4-bit binary data sequence A as in equation (7). Then, the following formula is established.

Z = TLU (A)
(9)

Here, TLU means a table lookup type function, and is generally realized by a memory or the like. As in equation (6), there are 10 input binary data strings A, so Z in equation (9) is also 10
Take, for example, a 7-bit scalar value. Each branch metric BM is calculated for Z in Equation (9) according to Equation (8). In this way, an appropriate branch metric can be defined even for a channel response that cannot be expressed by a convolution operation.

[Generation of equalization error signal and adaptive control of reference level (table lookup function)]
In order to perform the table lookup type maximum likelihood decoding described in the previous section, it is necessary to obtain a table lookup function TLU corresponding to the actual recording / reproduction characteristics of the optical disc medium 1. Therefore, the table lookup function TLU is obtained from the actual reproduction signal by adaptive control.

First, the table lookup function of equation (9) is defined from the convolution expression of equation (6).
It is assumed that the binary data sequence A output from the maximum likelihood decoder 5 is [0000] at a certain time k. Then, an ideal channel response Z_0 (k) at time k is obtained according to equation (9).

Z_0 = TLU (A [0000])
If the output of the adaptive equalizer 4 at this time is Y (k), the equalization error E (k) is defined by the above-described equation (2).
E (k) = Y (k)-Z_0 (k)
…(Ten)

In order to perform adaptive control of the table lookup function TLU, the following processing is performed.
Z_0 (k + 1) = Z_0 (k) + α · E (k)
(11)

Here, α is an update coefficient as in the equations (3) to (5), and is a small positive value (for example, 0.0
1) is set. The subscripts k and k + 1 represent time, and represent that the value of Z_0 changes as time passes.

Similarly, at a certain time k, the binary data sequence A output by the maximum likelihood decoder 5 is [0001].
If it is, the same process as in equation (11) is performed to update the value of Z_1. Thereafter, the value of Z_x (k) corresponding to the value of the obtained binary data string A is updated. In this way, the table lookup function TLU gradually matches the actual recording / reproduction characteristics of the optical disc medium 1.

[Restriction of adaptive control of reference level (table lookup function)]
Next, reference level adaptive control limitation, which is a feature of the present invention, will be described. When the adaptive control of the filter described so far and the adaptive control of the reference level of the maximum likelihood decoder are performed simultaneously, there is a problem that competition occurs between the two adaptive controls. Therefore, in the adaptive control of the table lookup function according to the present invention, there is a restriction that adaptive control of at least two elements of the table lookup function is not performed. In particular, when the binary data string A is [0000] or [1111], the calculation shown in Expression (11) is not performed, and the values of Z_0 and Z_F are not updated. By performing such a restriction, the table lookup function can be converged to an optimum value with a simple configuration.

[Reference value generator and equalization error generator]
The configuration of the reference value generator 8 having the above characteristics will be described with reference to FIG. In this figure, the pattern determiner 303 takes the output A (K) of the maximum likelihood decoder 5 as an input, and the pattern of the past 4 bits is expressed by the formula (
Judge which pattern shown in 7) matches. Reference value registers 320 to 325 are expressions
This register holds the output of the table lookup function TLU shown in (9). Originally, ten reference value registers Z_0 to Z_F are necessary, but in FIG. 7, only six registers Z_0, Z_1, Z_3, Z_6, Z_E, and Z_F are used in order to omit redundant descriptions. A summary of the outputs of the reference value registers 320 to 325 is connected to the maximum likelihood decoder 5 as a reference value table Z_x, and at the same time, one value is selected by the selector 304 according to the output of the pattern determiner 303 and Z ( is output as k). The reference value updaters 310 to 313 perform reference value update processing shown in Expression (11). Here, as in the case of the reference value register, originally eight reference value updaters are necessary, but in FIG. 7, only four are used to omit redundant description. Each reference value updater is connected to the equalization error signal E (k), the reference value register value Z_x, and the output of the pattern determiner 303, and when selected by the pattern determiner 303, the expression (11) The reference value update process shown in (1) is performed. FIG. 8 is a diagram illustrating the configuration of the equalization error generator 7. The equalization error generator 7 is composed of a delayer 301 and a subtracter 302,
By delaying the output Y (k) of the adaptive equalizer 4 by a delay unit 301 by a predetermined time and subtracting the ideal channel output Z (k) from the output of the delay unit 301, the equation (10) is obtained. The indicated equalization error is obtained.

With the above configuration, Viterbi detection that performs adaptive equalization by the LMS algorithm and adaptive control of the reference level can be stably and simultaneously processed with a simplified configuration.
[Application to other classes]
In the above description, the case of PRML in 6 states has been described, but the present invention is not limited to the case of 6 states and can be applied to the case of an arbitrary number of states.
FIG. 9 shows the result of comparing the error rate when the conventional PRML detection is performed on the data recorded on the same double-layer HD DVD-R media and the error rate by the digital data reproducing apparatus according to the present invention. . The horizontal axis is the relative value of Radial Tilt that reflects the warp of the medium, and is directly facing the objective lens 31 at 0.2. At this time, the Byte Error Rate on the vertical axis can be compared. As can be seen from this result, the digital data reproducing apparatus according to the present invention can reduce the error rate to ½ or less than the conventional case (when only adaptive learning of the adaptive equalizer is performed). That is, even when a medium having a lower quality than before is used, data can be reproduced with a good error rate, and the cost of the medium can be reduced.

The present embodiment is a digital data reproducing apparatus that includes an adaptive equalizer and a maximum likelihood decoder, and performs adaptive learning of the adaptive equalizer and automatic adjustment of a reference level used by the maximum likelihood decoder. The level adjustment is performed on a combination excluding at least two elements among all combinations of reference levels.

As an effect, it is possible to reduce a reproduction error rate in a recording medium having a large nonlinear distortion.
The present invention is not limited to the above-described embodiments, and various modifications can be made without departing from the scope of the invention.
Various inventions can be formed by appropriately combining a plurality of constituent elements disclosed in the above-described embodiments. For example, some components may be deleted from all the components shown in the embodiment. Furthermore, constituent elements according to different embodiments may be appropriately combined.

The block block diagram which shows an example of the principal part of the optical disk reproducing | regenerating apparatus of the Example which concerns on this invention. 1 is a block diagram showing an example of an optical disc playback apparatus according to an embodiment of the present invention. The figure which shows the structure of an adaptive equalizer and an adaptive learning device. The figure which shows the structure of the Viterbi decoder of the Example which concerns on this invention. The figure which shows the state transition by the (1,7) RLL modulation system and PR (1,2,2,1) ML decoding system of the maximum likelihood decoding process based on this invention used for the Example. A trellis diagram corresponding to PR (1221) with a minimum run length of 1. The figure which shows the structure of a reference value generator. The figure which shows the structure of an equalization error generator. The figure which shows the effect of a present Example.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 ... Optical disk medium, 2 ... Preamplifier part, 3 ... A / D converter, 4 ... Adaptive equalizer, 5 ... Maximum likelihood decoder, 6 ... Adaptive learner, 7 ... Equalization error generator, 8 ... Reference value Generator, 9 ... demodulator, 12 ...
Preamplifier, 14 ... AD converter, 15 ... parameter control unit, 16 ... ACG circuit, 17
... FIR filter, 18 ... Viterbi decoder, 19 ... ideal waveform generator, 20 ... target waveform generator.

Claims (4)

A digital data playback device that performs Partial Response and Maximum Likelihood detection,
An equalization means for performing waveform equalization processing on a digital signal based on a predetermined coefficient and outputting an equalized signal;
Maximum likelihood decoding means for performing a maximum likelihood decoding process on the equalized signal output by the equalizing means based on a reference level and outputting a decoded signal;
Ideal waveform generating means for outputting one of the reference levels according to the decoded signal output from the maximum likelihood decoding means;
Optimization for calculating the error between the ideal waveform signal output from the ideal waveform generation means and the equalization signal output from the equalization means and optimizing the predetermined coefficient of the equalization means so as to minimize the error And an optimization means for optimizing a reference level of the maximum likelihood decoding means, wherein the optimization means for the reference level changes at least one reference level in the reference level and changes at least two reference levels. A digital data reproduction apparatus characterized by optimizing the reference level without doing so. 2. The digital data reproducing apparatus according to claim 1, wherein the ideal waveform generating means outputs an ideal waveform by table lookup according to the composite signal. The optimization means for the predetermined coefficient includes the ideal waveform signal output from the ideal waveform generation means,
The at least one of the predetermined coefficients of the equalization means is optimized so that an error from the equalization signal output by the equalization means is minimized by LMS algorithm processing. 2. The digital data reproducing apparatus according to 2. The optimization unit uses the equalization unit to calculate an error between the ideal waveform signal output from the ideal waveform generation unit and the equalization signal output from the equalization unit and minimize the error. 3. The digital data reproducing apparatus according to claim 2, wherein the predetermined coefficient and a plurality of tap coefficients are optimized.

Images