JP5170819B2 - Information reproducing apparatus, information recording / reproducing apparatus, and information reproducing method - Google Patents

Description

The present invention relates to an information reproducing apparatus, information recording and reproducing apparatus, and relates to an information reproducing how, more specifically, the information reproducing apparatus for reproducing information recorded on an optical disk by using a PRML method, the information recording and reproducing apparatus, and about the information reproducing how.

  In recent years, with the advancement of digital technology and the improvement of data compression technology, it has been required to increase the capacity of optical disks. As a main measure for this requirement, it is conceivable to reduce the spot diameter of laser light used for reproducing information and improve the resolution of the optical system.

  For example, in an optical disc apparatus that reproduces and records information on an optical disc such as a BD (blu-ray disc) having a larger recording capacity than a DVD (digital versatile disc) or the like, a laser beam having a wavelength of about 390 nm to 420 nm is By condensing with an objective lens having a numerical aperture of about 0.70 to 0.90, the spot diameter of the laser beam formed on the recording layer of the optical disk is reduced to about 0.48 μm, for example, 0.160 μm to 0 .Reading and writing of information by a recording mark having a diameter of 138 μm or less is possible.

  However, due to problems such as the transparency of the polycarbonate material used for optical discs, it is difficult to further shorten the wavelength of laser light and increase the NA of the objective lens. An optical disc (hereinafter also referred to as “super-resolution optical disc”) capable of reproducing information whose recording mark pitch is smaller than the diffraction limit (hereinafter referred to as “super-resolution playback”) has been proposed (for example, a patent). Reference 1 to Patent Document 4).

  This super-resolution optical disc has, for example, a super-resolution layer containing a material whose optical constant changes when irradiated with laser light. The reproduction laser beam is condensed on the super-resolution layer. The super-resolution reproduction is enabled by forming a minute mask region or minute aperture region in the light spot due to the change in the optical constant at the time.

  However, when the information recorded on the super-resolution optical disk is reproduced using the optical disk apparatus disclosed in Patent Documents 1 to 4, nonlinear distortion occurs in the RF signal obtained from the reflected light of the optical disk, resulting in a reproduction error. Therefore, it is difficult to significantly increase the capacity.

  On the other hand, in recent years, it has been widely used to adopt a partial response (PR) method in a digital information reproducing apparatus of an optical disc apparatus. This is because, as the recording density of a recording medium for recording digital information increases, it becomes difficult to read a read signal of one bit of digital information without the influence (intersymbol interference) due to the adjacent bits. is there.

In the partial response method, the waveform equalization processing unit actively creates a known linear waveform interference to prevent deterioration of signal performance due to equalization decoding processing. Recently, in this partial response method, A PRML (Partial Response Maximum Likelihood) method that enables high-accuracy signal processing by combining an ML (Maximum Likelihood) method, which is a maximum likelihood estimation method, has been put into practical use.

  However, an optical disk reading system has nonlinearity in principle by reading out a signal using light diffraction, and an RF signal has asymmetry due to the nonlinearity. The RF signal also includes a non-linear component due to mark position fluctuations depending on the recording pattern due to thermal interference during recording. Asymmetry, nonlinear components, and the like included in these RF signals become a problem when attempting to increase the density of information recorded on the optical disk.

  Therefore, as shown in Non-Patent Document 1, there has been proposed a method of performing maximum likelihood estimation using a nonlinear compensation table in a state in which a nonlinear component of intersymbol interference is taken into consideration. However, in the method described in Non-Patent Document 1, since the maximum likelihood estimation bit length and the nonlinear compensation table length are equal, when the nonlinear compensation table length is shorter than the optical spot length on the recording medium, In the case of density recording, the nonlinear compensation effect is limited. This is especially true when reproducing information from a super-resolution reproduction medium where high-density recording that exceeds the diffraction limit of the optical system is performed, or when aberrations such as coma and astigmatism that increase the light spot diameter occur. The compensation effect is not sufficient.

  Therefore, for example, as shown in Patent Document 5, in addition to linear intersymbol interference prediction with a certain bit length, a method for performing nonlinear compensation and maximum likelihood estimation with a nonlinear compensation table at a longer bit length is disclosed. Has been. However, in this method, since it is necessary to perform maximum likelihood estimation in the same range as the nonlinear compensation bit length, there is a problem that the circuit scale increases dramatically as the nonlinear compensation range is increased. The circuit for realizing maximum likelihood estimation generally doubles the number of states each time one bit is added, and therefore the circuit scale also doubles. Therefore, for example, when pattern compensation bits are added by 3 bits, the circuit scale becomes 10 times or more, and the reproducing apparatus becomes expensive.

  Further, for example, Patent Document 6 discloses a method of performing nonlinear compensation on a long-range bit string without greatly increasing the circuit scale by using past provisional determination results. However, the increase in recording density due to these nonlinear compensations is as small as 10 to 20%, and it is difficult to significantly improve the capacity.

JP-A-6-183152 JP-A-5-205314 JP-A-11-250493 JP 2001-250274 A JP 2004-326839 A JP 2001-126394 A Naoki Ide, “Adaptive Partial-Response Maximum-Likelihood Detection in Optical Recording Media” ISOM2002

  The present invention has been made under such circumstances, and a first object thereof is to provide an information reproducing apparatus capable of reproducing information recorded on an optical disc with a mark smaller than the diffraction limit with high accuracy. is there.

  A second object of the present invention is to provide an information reproducing method capable of accurately reproducing information recorded on an optical disc with a mark smaller than the diffraction limit.

  A third object of the present invention is to provide an information recording / reproducing apparatus that can accurately reproduce information recorded on an optical disc with a mark smaller than the diffraction limit.

According to a first aspect of the present invention, there is provided an information reproducing apparatus for reproducing information using the PRML method, and an optical system including a light source and an objective lens for condensing the light emitted from the light source onto an optical disc; An optical head having a photodetector for receiving reflected light from the optical disc; a signal generating means for generating an RF signal from the output signal of the photodetector; and a recording mark whose information is smaller in pitch than the diffraction limit. A waveform equalizer for waveform equalizing the RF signal to a preselected partial response characteristic of the PRML method on the recorded optical disc; and a plurality of compensation values for each bit pattern of the recording mark are recorded. A compensated pattern compensation memory; a path memory having a function of outputting past determination results for a plurality of states according to the PRML method; and the plurality of compensation values The information reproducing apparatus comprising; using a compensation value the corresponding past determination results of out, a branch metric calculator for calculating a likelihood for each received signal of said waveform equalized RF signal.

According to this, the likelihood corresponding to each state of the waveform equalized RF signal is calculated using a compensation value according to the past determination result. Therefore, nonlinearity is compensated for in the likelihood calculation process, and as a result, information recorded with a mark smaller than the diffraction limit on the optical disc can be reproduced with high accuracy.

According to a second aspect of the present invention, there is provided an information reproducing method for reproducing information using the PRML method, wherein the information is read from an optical disc recorded with recording marks having a pitch smaller than a diffraction limit. When, the RF signal, a waveform equalization process for waveform equalization to partial response characteristic of the pre-selected the PRML method; compensation value calculation for calculating a compensation value for each bit pattern of the waveform equalized RF signal A branch metric for calculating a likelihood for each received signal of the waveform equalized RF signal using the compensation value according to a past determination result for a plurality of states determined according to the partial response characteristics And a value calculating step.

According to this, the likelihood corresponding to each state of the waveform equalized RF signal is calculated using a compensation value according to the past determination result. Therefore, nonlinearity is compensated for in the likelihood calculation process, and as a result, information recorded with a mark smaller than the diffraction limit on the optical disc can be reproduced with high accuracy.

  According to a fourth aspect of the present invention, there is provided an information recording / reproducing apparatus for recording and reproducing information with respect to an optical disc, the information reproducing device of the present invention; and a recording mark having a pitch smaller than a diffraction limit for the optical disc. And an information recording / reproducing apparatus comprising:

According to this, the information recording and reproducing apparatus includes information reproducing apparatus of the present invention. Therefore, it is possible to improve the calculation accuracy of likelihood, and as a result, information recorded with a mark smaller than the diffraction limit on the optical disc can be accurately reproduced.

  Hereinafter, an embodiment of the present invention will be described with reference to FIGS. FIG. 1 shows a schematic configuration of an optical disc apparatus 20 according to an embodiment of the present invention.

  The optical disk device 20 shown in FIG. 1 includes a spindle motor 22 for rotating the optical disk 15, an optical pickup device 23, a seek motor 21 for driving the optical pickup device 23 in the radial direction of the optical disk 15, and laser control. A circuit 24, a drive control circuit 26, a reproduction signal processing circuit 28, a buffer RAM 34, a buffer manager 37, an interface 38, a flash memory 39, a CPU 40, a RAM 41, and the like are provided. Note that the arrows in FIG. 1 indicate the flow of typical signals and information, and do not represent the entire connection relationship of each block. In the present embodiment, as shown in FIG. 2 as an example, the optical disc 15 is sandwiched between a pair of transparent substrates 15a and 15e, the recording layer 15c on which information is recorded, and the reflective layer that reflects light. It is assumed that the super-resolution optical disc has a layer 15b and a super-resolution layer 15d including a material whose optical constant changes depending on temperature and is capable of super-resolution reproduction.

  The optical pickup device 23 is a device for irradiating the optical disc 15 with laser light and receiving reflected light from the optical disc 15. The optical pickup device 23 includes a light source that emits laser light having a wavelength corresponding to the optical disk 15 and an objective lens that condenses the light from the light source on the optical disk 15, and is reflected by the optical disk 15 and returned via the objective lens. An optical system that guides light to a predetermined position, a light receiver that has a plurality of light receiving regions that are arranged at the predetermined position and receive the return light, a drive system that drives the objective lens minutely (all not shown), and the like. Yes. Each light receiving region of the light receiver outputs a signal (photoelectric conversion signal) corresponding to the amount of received light to the reproduction signal processing circuit 28. The drive system also includes a focusing actuator for driving the objective lens in the focus direction and a tracking actuator for driving the objective lens in the tracking direction. Here, as an example, the wavelength of the laser light emitted from the light source is 635 nm, and the numerical aperture (NA) of the objective lens is 0.6. In this case, the diffraction limit is about 530 nm (laser wavelength / 2NA).

  By the way, the relationship between the CNR (carrier / noise ratio) and the reproduction power Pr when reproducing the recording marks (recording mark length = 200 nm) formed on the optical disk 15 using the optical pickup device 23 is taken as an example. It is shown in FIG. In FIG. 3, it can be seen that when the reproduction power Pr is 2 mW or more, the CNR exceeds 30 dB, and super-resolution reproduction is possible. Hereinafter, the reproduction power capable of super-resolution reproduction is also referred to as “super-resolution reproduction power”.

  When laser light with super-resolution reproduction power is focused on the optical disk 15, the temperature of that portion rises, and in the light spot, as shown in FIG. In the formed beam spot BS, a minute aperture area HA or, as an example, a minute mask area MA is generated as shown in FIG. Each of the minute opening area HA and the minute mask area MA has a shape with a tail in the direction opposite to the traveling direction of the light spot. The black circles in the figure are recording marks formed on the optical disk 15, and which of the micro-opening area HA and the micro-mask area MA occurs in the super-resolution layer 15d depends on the material and the layer configuration of the super-resolution layer 15d. It depends.

  When the minute opening area HA occurs, the amount of return light varies greatly depending on whether or not the recording mark is included in the minute opening area HA. On the other hand, when the minute mask area MA occurs, the amount of the return light greatly changes depending on whether or not the recording mark is masked by the minute mask area MA. For example, when the mask area MA is generated, as shown in FIG. 5 as an example, the optical constant of the super-resolution layer changes due to heating by irradiation with the laser beam 23a having the optical axis AX. To do. As a result, a mask MA is formed at the spot rear portion, and the reflectance from the reflective layer 15b is lowered. Considering the light intensity distribution on the incident-side surface PL of the super-resolution layer 15d, the incident light indicated by the curve L1 in FIG. 5 has a symmetric light spot shape, but the reflected light indicated by the curve L2 is asymmetric. A light spot shape (state). Therefore, information from the recording layer 15c is read out with an asymmetrical light spot, and phase distortion occurs in the RF signal.

  Further, the reproduction power dependence of the light intensity distribution of the reflected light on the incident-side surface PL of the super-resolution layer is shown as an example in FIG. In FIG. 6, the center position of the light spot is used as the reference (origin). According to this, when the reproduction power Pr is increased to the super-resolution reproduction power, the rear of the light spot is masked, and the light intensity distribution of the reflected light is changed to a shape that is shaved from the rear part. As shown in FIG. 6, the shape of the light spot formed on the recording layer of the super-resolution optical disc is such that the NA of the objective lens for condensing the laser light on the recording layer is increased when the wavelength of the laser light is reduced. In this case, when compared with the light spot formed on the recording layer of the optical disc, the tip of the light spot intensity distribution is sharp while maintaining the width of the skirt.

  Therefore, when the recording density is increased in the super-resolution optical disk, the bottom of the light spot of the laser light is smaller than the case where the wavelength of the laser light and the objective lens NA are increased to reduce the spot system of the light spot. Intersymbol interference occurs in a wide range due to the size of the. FIG. 7A is a conceptual diagram showing the clock timing when information recorded at a normal recording density on an optical disc such as a DVD is reproduced and the light intensity of reflected light from the optical disc. FIG. 7B is a conceptual diagram showing clock timing when information recorded at a density twice the normal recording density is reproduced and reflected light from the optical disc. FIG. FIG. 5 is a conceptual diagram showing clock timing and reflected light from an optical disc when information recorded at double density is collected by a high NA objective lens or reproduced by a short wavelength laser beam.

  When information recorded at a normal recording density is reproduced, the range of the area PA where the intersymbol interference of the light spot occurs is about 3 clocks as shown in FIG. When the information recorded at the recording density is reproduced, the range of the area PA is about 5 clocks as shown in FIG. 7B. An area where intersymbol interference of a light spot occurs when information recorded at twice the normal recording density is collected by a high NA objective lens or reproduced by a short wavelength laser beam. The range of PA is about 3 clocks as shown in FIG. Therefore, when reproducing information recorded on a super-resolution optical disc at a density of, for example, twice or more, the range of the area PA where intersymbol interference occurs as shown in FIG. Compared to the case shown in (C), it becomes longer, and there is a need to suppress intersymbol interference over a longer range.

  Also, as shown in FIG. 6, the width of the bottom of the light spot of the laser beam does not change even when recording on the super-resolution optical disk, so the light spot is not simply reduced. Therefore, in high-density recording on a super-resolution optical disc, it is very difficult to form a recording mark on the recording layer of the optical disc with high accuracy, and a recording error caused by thermal interference caused by an adjacent or adjacent recording mark is difficult. May occur. Such a recording error adversely affects the reproduction signal as nonlinear intersymbol interference.

  In addition, even in a super-resolution optical disc for reproduction only in which a pit pattern is formed by a pit portion and a space portion, there is a difference in heat capacity due to a difference in structure between the pit portion and the space portion, and a light spot is formed. The heat distribution of the super-resolution layer changes according to the pit pattern of the portion that is present. For this reason, the light spot shape changes according to the pit pattern, and it is considered that the reproduced signal is adversely affected as non-linear intersymbol interference.

  As described above, it is important to suppress nonlinear intersymbol interference over a long range in order to reproduce information recorded on a super-resolution optical disk with high density.

  Returning to FIG. 1, the reproduction signal processing circuit 28 includes an amplifier 28a, a servo signal generation circuit 28b, a wobble signal generation circuit 28c, an RF signal generation circuit 28d, a decoder 28e, and the like.

  The amplifier 28a converts each of a plurality of photoelectric conversion signals from the light receiver of the optical pickup device 23 into a voltage signal and amplifies it with a predetermined gain.

  The servo signal generation circuit 28b generates a servo signal (focus error signal, track error signal, etc.) based on each output signal of the amplifier 28a. The servo signal generated here is output to the drive control circuit 26.

  The wobble signal generation circuit 28c generates a wobble signal based on each output signal of the amplifier 28a.

  The RF signal generation circuit 28d generates an RF signal based on each output signal of the amplifier 28a.

  The decoder 28e extracts address information and a synchronization signal from the wobble signal. The address information extracted here is output to the CPU 40, and the synchronization signal is output to the drive control circuit 26.

  The decoder 28e performs a decoding process and an error detection process on the RF signal. When an error is detected, the decoder 28e performs an error correction process, and then stores the reproduced data in the buffer RAM 34 via the buffer manager 37. To do.

FIG. 8 is a block diagram of the decoder 28e. As shown in FIG. 8, the decoder 28e includes a high pass filter (HPF) 60, an equalization circuit 61, an AD converter (ADC) 62, an asymmetric FIR filter 63, an interpolator 64, an equalization circuit 65, a decision feedback Viterbi decoding. A device 66 and a PLL 67 are provided.

The equalization circuit 61 is arranged at the subsequent stage of the HPF 60, and is a high frequency band attenuated by a decrease in MTF (modulation transfer function) by the optical system of the RF signal input from the RF signal generation circuit 28d and from which low frequency noise has been removed by the HPF 60. Emphasize components to reduce intersymbol interference. Note that the equalization circuit 61 also serves as a low-pass filter (LPF) that cuts high-frequency components so that aliasing noise does not occur during AD conversion in the ADC 62.

ADC62 is disposed downstream of the equalization circuit 61 converts (AD conversion) to the output signal of the equalizer circuit 61 into a digital signal.

The asymmetric FIR filter 63 is an FIR filter having an asymmetric equalization coefficient with the tap center as the origin. The asymmetric FIR filter 63 performs a filtering process on the output signal of the ADC 62, and asymmetry of the light spot formed on the above-described super-resolution optical disk. The phase distortion of the RF signal due to the above is corrected.

  The interpolator 64 receives the output signal of the ADC 62 or the output signal of the asymmetric FIR filter 63, and interpolates the sample value at the clock timing from the sample values of two or more times before and after.

  The PLL 67 reproduces a clock of a signal recorded on the optical disc 15 from the output signal of the interpolator 64 (hereinafter also referred to as “reproduction clock”), and instructs the interpolator 64 of the clock timing. That is, sampling in synchronization with the reproduction clock is performed by a combination of the interpolator 64 and the PLL 67. The phase comparison part of the reproduction clock and the signal clock in the PLL 67 may be one that performs phase comparison on a signal binarized using a slicer, and the level value of the reproduction signal and the phase error between the reproduction clock and the signal clock are A method using a method of detecting a phase difference from a difference from an ideal signal level value when no signal is present may be used.

The equalization circuit 65 is arranged at the subsequent stage of the interpolator 64 and performs waveform equalization with respect to the output signal of the interpolator 64 so that a response corresponding to a desired PR (Partial Response) characteristic is obtained. The desired PR characteristic is, for example, PR (1, 2, 2, 2, 1).

Decision feedback Viterbi decoder 66, as will be described in detail later, is disposed downstream of the equalization circuit 65, the output signal of the equalizer circuit 65, a Viterbi a maximum likelihood (maximum likelihood) decoding method (Viterbi This is a decision feedback type non-linear Viterbi decoder that performs decoding processing by a decoding method and outputs binarized data. That is, here, the equalization circuit 65 and the decision feedback Viterbi decoder 66 perform PRML (Partial Response Maximum Likelihood) method signal processing that combines the partial response method and the maximum likelihood decoding method.

  Returning to FIG. 1, the drive control circuit 26 generates a drive signal for the drive system of the optical pickup device 23 based on the servo signal from the reproduction signal processing circuit 28 in order to correct the displacement of the objective lens 60. Thereby, tracking control and focus control are performed. Furthermore, the drive control circuit 26 generates a drive signal for driving the seek motor 21 and a drive signal for driving the spindle motor 22 based on an instruction from the CPU 40. The drive signal of each motor is output to the seek motor 21 and the spindle motor 22, respectively.

  The buffer RAM 34 temporarily stores data reproduced from the optical disc 15 (reproduced data) and the like. Data input / output to / from the buffer RAM 34 is managed by the buffer manager 37.

  The laser control circuit 24 controls the light emission power of the light source of the optical pickup device 23.

  The interface 38 is a bidirectional communication interface with a host device 90 (for example, a personal computer), and is a standard interface such as ATAPI (AT Attachment Packet Interface), SCSI (Small Computer System Interface), and USB (Universal Serial Bus). It is compliant.

The flash memory 39 stores various programs described by codes readable by the CPU 40, light emission characteristics of the light source of the optical pickup device 23, equalization coefficient information to be described later, and the like.

  The CPU 40 controls the operation of each unit in accordance with the program stored in the flash memory 39 and stores data necessary for control in the RAM 41 and the buffer RAM 34.

<Details of Decision Feedback Viterbi Decoder>
Here, the details of the decision feedback Viterbi decoder 66 will be described. The description of the decision feedback Viterbi decoder 66 will be described on the assumption that the PR class of the decision feedback Viterbi decoder 66 is PR (1, 2, 2, 2, 1) and the minimum inversion interval is 2T. However, the PR class of the decision feedback Viterbi decoder 66 is not limited to PR (1, 2, 2, 2, 1). For example, PR (1, 2, 2, 1) or PR (1, 1) Any PR class such as Further, the provisional determination described later of the decision feedback Viterbi decoder 66 is performed using a 3-bit bit string, but the present invention is not limited to this, and a bit string of 3 bits or more may be used.

  First, for convenience of explanation, an explanation of a normal Viterbi decoder will be given prior to explanation of a decision feedback Viterbi decoder. FIG. 9 shows a block diagram of a normal Viterbi decoder 66 '. As shown in FIG. 9, the Viterbi decoder 66 ′ includes a branch metric calculator 70, an ACS calculator 71, a path memory 72, an output selector 73, and a path metric memory 74. In addition, since the Viterbi decoder 66 ′ configured as described above has a PR class of PR (1, 2, 2, 2, 1), its state transition diagram is 4 bits as shown in FIG. Ten states (states) S0000 to S1111 each represented by a bit string and 16 branches (branches) B (n) (n corresponding to the states S0000 to S1111 as indicated by arrows in FIG. = 1, 2,..., 16). The state transition between the states S0000 to S1111 can be represented by a trellis diagram shown in FIG.

The branch metric calculator 70 calculates the Euclidean distance (branch metric BM _t ) between the target value of the branch from each state at the present time (time t−1 in FIG. 11) to each state at the time t and the input signal. Do. More specifically, the target value in each branch B (n) is a value defined by the PR class and the bit string corresponding to each branch, and the branch metric BM _t (n) of each branch B (n) is expressed by the following equation: It is represented by (1).

BM _t (n) = (PP (n) × PR−RF) ² (1)
However, n is an integer of 1 to 16, PR is a matrix [12221] ^T , and RF is a value of an RF signal waveform-equalized to a PR characteristic as an input signal. PP (n) is a 5-bit bit string corresponding to each branch B (n). For example, the bit string PP (5) corresponding to the branch B (5) in FIG. 10 is [00110], and the bit string PP (6) corresponding to the branch B (6) is [01111]. Is a 5-bit bit string obtained by adding the first bit of the bit string after the transition to the 4-bit bit string indicating the state before the transition corresponding to the branch B (n). The branch metric calculator 70 calculates 16 branch metrics BM _t (n) based on the equation (1). In the equation (1), x indicates matrix multiplication.

Returning to Figure 9, ACS (Add-Compare- Select) arithmetic unit 71, the path metric _PM t-1 (0000) for each state S0000~S1111 at time t-1 stored in the path metric memory 74 Pm _t reads _-1 (1111), the branch metric _BM t (n) from the time t-1 which is calculated by the branch metric calculator 70 to the time t, the corresponding path metric _PM t-1 (0000) Pm _{t −1} (1111) is added to calculate an added value PM ′ (n). For example, the path metric PM _t-1 (0011) is added to the branch metric BM _t (5) and the branch metric BM _t (4), and the path metric PM _t-1 (1110) is added to the branch metric BM _t (9). Is added. That is, the path metric PM _t−1 of each state at the transition source is added to the branch metric BM (n) corresponding to the branch B (n) indicating the transition.

  Then, when there are two paths joining the states S0000 to S1111 at time t in the trellis diagram of FIG. 11, the path metrics corresponding to the two paths are compared, and the smaller path metric is compared. Is determined as the surviving path of the state at time t. If there is only one path joining each of the states S0000 to S1111, the path is unconditionally determined as the surviving path of the state at time t. Each determination result is recorded in the path memory 72. As a result, the determination result by the ACS computing unit 71 is sequentially recorded in the path memory 72, and the past surviving paths are recorded as a history. In addition, the ACS computing unit 71 uses the added value PM ′ (n) corresponding to the surviving path in parallel as a new path metric value at time t, so that the path metrics PM (0000) to PM (1111) in the path metric memory 74 Update the value of.

FIG. 12 is a block diagram of the path memory 72. The path memory 72 includes the same number of shift registers as the number of states determined by the PR class. In this embodiment, since the PR class is PR (1, 2, 2, 2, 1), it has path memory cells 90 _{0 to} 90 ₁₆ each including 10 shift registers. The shift registers of the respective path memory cells 90 _{0 to} 90 ₁₆ hold the determination result in the ACS computing unit 71 while shifting to the next path memory cell every time. that time. According to the estimation result output from the ACS computing unit 71, the estimation result corresponding to the state of the selected surviving path one hour before is copied. Thereby, in the subsequent stage of the path memory cells, survivor path by selecting the path with the judgment becomes less and less, the estimation result remaining in the shift register of the path memory cell 90 ₁₆ of the final stage is the approximately the same results. That is, the path merge is completed.

Returning to FIG. 9, the output selector 73 selects, from the path metric memory 74, one surviving path corresponding to the minimum path metric value, that is, most likely. Of the output from the path memory cell 90 ₁₆ of the final stage of the path memory 72, and outputs a value corresponding to the selected path as binary data (decision value). When the number of path memory cells in the path memory 72 is sufficient and the quality of the RF signal is good, the results remaining in the final stage of the plurality of shift registers in the path memory are almost the same. In this case, the output selector 73 is not necessarily required.

  Next, the decision feedback Viterbi decoder 66 according to this embodiment will be described. FIG. 13 is a block diagram of the decision feedback Viterbi decoder 66. Compared with the above-described conventional Viterbi decoder 66 ', the decision feedback Viterbi decoder 66 replaces the branch metric calculator 70 with a branch metric calculator 77 with a compensation function, and replaces the path memory 72 with a temporary determination function path memory. The difference is that a pattern compensation memory 75 is added instead of 76. Hereinafter, the decision feedback Viterbi decoder 66 will be described focusing on differences from the conventional Viterbi decoder 66 '.

  FIG. 14 is a block diagram of the path memory 76 with a provisional decision result output function of the decision feedback Viterbi decoder 66. The path memory with provisional determination result output function 76 outputs the contents of the path memory cell to the pattern compensation memory 75 as a provisional determination result corresponding to each state. Since the provisional determination result has bit information immediately after the bit under maximum likelihood estimation, it is reflected in the maximum likelihood estimation as the determination information of the backward bit, so that a wider range (long bit length) on the medium. Nonlinear compensation is possible.

  FIG. 15 is a diagram for explaining acquisition of a provisional determination result. Hereinafter, a method for acquiring a provisional determination result will be described with reference to FIG. The state corresponding to the time t to the time t-4 in FIG. 15 is a state transition part during maximum likelihood estimation. The state corresponding to the time t-5 to the time t-9 is a provisional determination result portion. In FIG. 15, when the time is traced backward from the respective states at t−1 along the survival paths indicated by the bold arrows, the provisional determination bit string corresponding to each survival path is determined to be one. That is, the route that follows the surviving path from each state S0000 to S1111 at time t-1 to state S0000 at time t-9 when path merging ends is uniquely determined, and each state S0000 to S1111 at time t-1 is determined. The temporary determination bits from time t-5 to time t-9 are determined as shown in FIG. Then, as can be understood from FIG. 14 and FIG. 15, the path memory 76 with a temporary determination result output function has a 3-bit bit string B (0000) corresponding to the states S0000 to S1111 from time t-5 to time t-7. ) To B (1111) are output to the pattern compensation memory 75 as temporary determination bit strings.

FIG. 16 is a block diagram of the pattern compensation memory 75. As shown in FIG. 16, the pattern compensation memory 75 has 16 B (n) compensation value set storage memories 51 _{1 to} 51 ₁₆ corresponding to each branch B (n), and has a provisional determination result output function. provisional decision bit string B from the path memory 76 (0000) ~B (1111) is input to the corresponding B (n) the compensation value set storage memory _{51 1-51 16} respectively. The B (n) compensation value set storage memories 51 _{1 to} 51 ₁₆ output compensation values CV (n) corresponding to the input temporary determination bit strings B (0000) to B (1111). As an example, in FIG. 17 (A) shows a table indicating the correspondence of B (1) compensation set storage memory 51 ₁ of the bit string B (0000) and the compensation value CV (1), in FIG. 17 (B) B (6) the compensation set storage table showing the correspondence of the bit string B in the memory 51 ₆ (0110) and the compensation value CV (n) is shown, B (12) in FIG. 17 (C) bit string of the compensation set storing memory _{51 12} A table showing the correspondence between B (1100) and the compensation value CV (12) is shown. For example, if the survivor paths are determined as shown in the trellis diagram of FIG. 15 is a B (1) bit sequence B which is input to the compensation value set storage memory 51 ₁ (0000) is (110) is there. Therefore, B (1) In the compensation value set storage memory 51 _1, the value of the compensation value CV (1) based on the table shown in FIG. 17 (A) is determined to -0.11. Similarly, B (6) the compensation value set storage memory 51 _6-bit string is input to the B (0110) is a (000), B (12) the compensation value set stored bit string B (1100) which is input to the memory _{51 12} Is (111). Thus, (6) the value of the compensation value set storage memory 51 ₆ and B (12) in the compensation value set storing memory _{51 12,} FIG. 17 (B) and the compensation value CV (6) based on the table of FIG. 17 (C) Is determined to be −0.01, and the value of the compensation value CV (12) is determined to be −0.08. As an example, the compensation values CV (1), CV (6), and CV (12) determined as described above are output to the branch metric calculator 77 with a compensation function.

Returning to FIG. 13, the branch metric calculator 77 with a compensation function adds a compensation value corresponding to each branch B (n) output from the pattern compensation memory 75 to a linear intersymbol interference prediction value having a bit length determined by the PR class. A value obtained by adding CV (n) is set as a target value, and a Euclidean distance (branch metric) between this target value and the value of the waveform-equalized RF signal as the input signal is calculated. The following formula (2) is a calculation formula of the branch metric by the branch metric calculator 77 with compensation function, and the branch metric calculator 77 with compensation function calculates the branch metric BM _t (n) according to formula (2). Equation (2) differs from Equation (1) in that it has a compensation value CV (n) term corresponding to each branch B (n).

BM _t (n) = (PP (n) × PR + CV (n) −RF) ² (2)

When the calculation of each branch metric BM _t (n) is completed by the above equation (2), the decision feedback Viterbi decoder 66 generates a binary signal in the same manner as the conventional Viterbi decoder 66 ′ described above. The

As described above, according to the optical disc apparatus 20 according to the present embodiment, the branch metric BM (n) corresponding to each branch B (n) is determined by the branch metric calculator 77 with a compensation function according to the PR class. A value obtained by adding a compensation value CV (n) corresponding to each branch B (n) output from the pattern compensation memory 75 to a linear intersymbol interference prediction value of It is calculated based on the value of the waveform equalized RF signal. Therefore, the nonlinear component included in the RF signal is compensated in the maximum likelihood estimation process, and as a result, information recorded on the optical disc 15 with a mark smaller than the diffraction limit can be reproduced with high accuracy.

  Further, the branch metric BM (n) is calculated in consideration of the compensation value CV (n) determined by the 3-bit bit string B (0000) to B (1111) corresponding to each branch B (n). Since the PR class of the embodiment is PR (1, 2, 2, 2, 1), this is equivalent to the number of nonlinear correction bits being 8 bits. Therefore, a decoder having a circuit for 5 bits excluding 3 bits of the provisional decision bit string can perform decoding substantially equivalent to a decoder having a circuit for 8 bits, thereby reducing the cost of the apparatus. It becomes.

  18A is a diagram showing an eye pattern of an output signal of the ADC 62, FIG. 18B is a diagram showing an eye pattern of an output signal of the asymmetric FIR filter 63, and FIG. 19A is a decision feedback. FIG. 19B is a diagram showing an eye pattern of an output signal from a decision feedback nonlinear viterbi decoder 69 with a compensation signal output, which will be described later, when there is no bit, that is, when the nonlinear correction bit length is 5 bits. It is a figure which shows the eye pattern of the output signal of the decision feedback type | mold nonlinear Viterbi decoder 69 with a compensation signal output mentioned later when a decision feedback bit is 3 bits, ie, when a nonlinear correction bit length is 8 bits. In the eye pattern shown in FIG. 18 (A), the eye is not open at all, and in the eye pattern shown in FIG. 18 (B), the eye is slightly opened, but it is far from the sufficient state and the asymmetry is also large. . In the eye pattern shown in FIG. 19A, asymmetry is lost and the eye is slightly opened, but it is not sufficient. On the other hand, in the eye pattern shown in FIG. 19B, it can be seen that the eye is clearly opened and sufficient non-linear compensation is performed. That is, as described above, when reproducing the optical disk 15, nonlinear compensation in a long range is necessary, and as a means for realizing it, decision feedback is optimal from the viewpoint of circuit scale. Note that the non-linear intersymbol interference removal method using the Volterra filter, which is a non-linear filter, is also highly effective in reducing the bit error rate for super-resolution optical discs. The method used is the best.

FIG. 20 is a diagram showing a comparison of the effects of the prior art and the present invention. If the limit of the bit error rate is 1 × 10 ⁻⁴ , the capacity increasing effect of the super-resolution optical disk represented by Patent Documents 1 to 4 is very small, about 1%. Further, in the method of Patent Document 6, a capacity increase of about 4% can be expected, but the effect is not sufficient. In comparison, the shortest mark length that can be realized in the present invention is shortened to 170 nm or less, and a capacity increase of 1.45 times or more (45% or more) can be expected.

Table 1 below shows the relationship between the number of decision feedback bits and the error rate when the shortest recording mark length is about 162 nm. When non-linear compensation is not performed, the bit error rate is 2.01 × 10 ⁻² . When there is non-linear compensation and there is no provisional decision bit, the bit error rate is about half, which is 4.44 × 10 ⁻³ . Furthermore, when the provisional decision bit of this method is 3 bits or more, the decision feedback has an effect of reducing the bit error rate of one digit or more, which is 2.67 × 10 ⁻⁴ . Further, it is sufficient that the nonlinear compensation length is 8 bits, and the bit error rate has the same number of digits for 8 bits, 9 bits, and 10 bits.

FIG. 21 is a graph showing the relationship between the number of nonlinear correction bits and the error rate when the shortest recording mark length is about 162 nm. It can be seen that there is a significant improvement in the bit error rate when the nonlinear correction bit number is between 7 bits and 8 bits, and the nonlinear correction bit number needs to be at least 8 bits. It can also be seen that the bit error rate is in the 1 × 10 ⁻⁴ unit where there is no practical problem with 8 bits or more, and it is sufficient to have a compensation value for a bit pattern of 8 bits or more.

  In the above embodiment, the compensation value is obtained by sequentially calculating the linear intersymbol interference prediction value calculated from the past determination result from the determination feedback Viterbi decoder 66 and the error value of the actual signal corresponding to the time. An averaged average value can be used. In FIG. 22, as an example, a block diagram of the compensation value learning mechanism 30 is shown together with a block diagram of the decision feedback Viterbi decoder 66.

The compensation value learning mechanism 30 converts the input value signal into the compensation value calculator 9, the delay unit 12 that delays the input signal by the decoding time in the decision feedback Viterbi decoder 66, and the binarized signal from the decision feedback Viterbi decoder 66. A PR intersymbol interference generator 10 that provides intersymbol interference corresponding to the PR class is provided. In the compensator learning mechanism 30, a binary signal from the decision feedback Viterbi decoder 66 and a signal output from the equalization circuit 65 via the delay unit 12 are binarized via the PR intersymbol interference generator 10. An error value (a non-linear component of the input signal) obtained by subtracting the digitized signal is input to the compensation value calculator 9, and the compensation value calculator 9 is based on the binarized signal via the PR intersymbol interference generator 10. A bit pattern compensation value corresponding to the binarized data string is read from the pattern compensation memory 75, a difference between the input error value and the compensation value is derived, and a value obtained by multiplying the difference by a predetermined constant is added to the compensation value. Thus, the value of the pattern compensation memory 75 is rewritten. The compensation learning mechanism 30 makes it possible to obtain an optimum compensation value by repeating the above operation.

  When the compensation value rewriting cycle is short, for example, every clock, that is, when the compensation value rewriting frequency, which is the reciprocal of the compensation value rewriting cycle, is large, interference may occur with the feedback loop for obtaining the recovered clock in the PLL 67. For this reason, it is desirable from the viewpoint of reproduction stability that the rewrite frequency of the compensation value is set to a value equal to or lower than the open loop crossing frequency of the feedback loop in the PLL 67.

In addition, when reproducing high-density recording from the beginning, compensation value information for nonlinear correction and an appropriate value of the waveform equalization coefficient in the FIR filter 63 are not known, so that the PLL 67 uses a reproduction clock from the signal. There is a possibility that it cannot be extracted and cannot be reproduced at all. In that case, as shown in the flowchart of FIG. 27, first, the compensation value and the waveform equalization coefficient are learned for the low density recording density, and then the reproduction is performed at a slightly high density recording density. This is a method for obtaining compensation values and waveform equalization coefficients for high-density reproduction step by step, such as learning and updating the waveform equalization coefficient. Can be accurately reproduced.

  Further, a super-resolution optical disc is provided with a compensation value training area composed of a known bit pattern for each TOC (Table Of Contents) information and a predetermined section of the disk medium, and when the compensation value training area is reproduced. Instead of using the output from the decision feedback Viterbi decoder 66, a compensation value may be obtained using a known bit pattern. The compensation value training area is, for example, as shown in FIG. 23A, the TOC area 1 in the inner periphery of the recording area 2 of the optical disk 15, or the optical disk 15 as shown in FIG. The recording area 2 can be periodically arranged as indicated by black dots 3. As a result, the compensation value can be stably learned without being affected by the error bit output from the decision feedback Viterbi decoder 66.

  Of course, in the training area, a known bit pattern may be recorded in advance for the ROM at the time of manufacture, and a pre-recorded bit pattern may be used for the rewritable / write-once optical disc. Alternatively, an unrecorded portion for training for recording a known bit pattern may be prepared, and the prepared known bit pattern may be written. At that time, as described above, in order to learn the compensation value at a plurality of recording densities for the purpose of improving the stability of the PLL at a high density, a training area at a plurality of recording densities is prepared. High density reproduction is possible. For example, the compensation value training area recorded at a low density in the TOC area 1 of the optical disk 15 shown in FIG. 28 is used for the compensation value training recorded at a high density in the TOC area 2 at another position of the optical disk. An area can be provided.

  It is also possible to record a compensation value suitable for the medium on the super-resolution optical disc 15 and read it into the optical disc apparatus 20 during reproduction. For example, if the wobble information or TOC information of the optical disk 15 includes information about the compensation value (preliminary compensation value information), the CPU 40 also reads the information about the preliminary compensation value from the optical disk 15 and mounts it in the RAM 41 when the disk is mounted. If there is a notification that the reproduction request command has been received from the host device 90, the value of the pattern compensation memory is set based on the information on the preliminary compensation value stored in the RAM 41 and the reproduction condition. Also good.

  The temperature distribution of the super-resolution layer changes and the shape of the aperture or mask changes depending on the playback conditions such as the super-resolution layer configuration, recording density, playback power, playback linear velocity, and light source drive waveform. Therefore, the manner of nonlinearity expression is different. Therefore, it is preferable that one or more parameters are defined and have an optimum preliminary compensation value at that time. In this case, for example, the preliminary compensation value information corresponds to reproduction at 4 × speed, and when reproduction by the user at 16 × speed is requested, the preliminary compensation value information corresponds to reproduction at 16 × speed. A compensation value may be predicted.

Further, the linear intersymbol interference prediction value calculated from the determination result from the past Viterbi decoder and the actual signal error value corresponding to the time have a probability distribution for each bit pattern. Therefore, the compensation value at the same time, the dispersion value VV (n) also is stored in the B (n) the compensation value set storage memory _{51 1-51 16,} the following equation to calculate the branch metric BM (n) to (3) By using this, maximum likelihood estimation can be performed in consideration of the influence of the variance value, and the effect of suppressing bit errors can be improved. Note that k is a constant of 1 to 4.

BM _t (n) = (PP (n) × PR + CV (n) −RF) ² / (VV (n)) ^k (3)

As the structure, B (n) the compensation value set storage memory _{51 1-51 16} of each branch B (n), stores the variance VV (n) is added to the compensation value CV (n), the compensation value CV Similarly to (n), the variance value VV (n) is transmitted to the branch metric calculator 77 with a compensation function based on the tentative determination result, that is, the bit string B (0000) to B (1111). The branch metric calculator 77 with a compensation function calculates the branch metric BM according to the above formula. Incidentally, it is shown in FIG. 24 the variance value with a B (6) of the compensation value set storage memory 51 ₆ table as an example. Of course, a deviation value obtained by taking the root of the dispersion value instead of the dispersion value may be stored. A similar method is disclosed in Japanese Patent Application Laid-Open No. 2005-223854, but the effect is greatly enhanced in the decision feedback Viterbi decoder 66 in the present embodiment for the following reason.

When the nonlinear compensation bit length is short, the probability distribution of the equalization error value of each bit pattern is added by the central limit theorem and becomes close to a normal distribution. However, the decision feedback nonlinear Viterbi decoder of the present invention is highly effective because the probability distribution peculiar to each bit pattern appears by increasing the compensation bit length and the variance value is different. Of course, this dispersion value consideration is not necessary in the case of the decision feedback type non-linear Viterbi decoder alone shown in Patent Document 6 because of the necessity of non-linear compensation with a long bit length, so that the effect is thin. The effect is enhanced. The following table 2 is a table showing the effect when the dispersion value is taken into consideration. As shown in Table 2, when the nonlinear correction bit length is short, the effect of reducing the bit error rate is as thin as about 10%. However, when the nonlinear correction bit length is increased, the bit error rate can be reduced to half or less. .

  Also, in FIG. 25, as an example, a block diagram of the nonlinear component correction signal output mechanism 31 is shown together with a block diagram of the decision feedback Viterbi decoder 66. The signal input to the decision feedback Viterbi decoder 66 includes a nonlinear component, but the nonlinear component can be removed from the input signal by compensating the nonlinear component with the value of the pattern compensation memory 75. The nonlinear component compensator 13 reads a compensation value of a pattern corresponding to the binarized data string from the pattern compensation memory 75, and compensates from the signal delayed by the delay unit 12 by the decoding time in the decision feedback Viterbi decoder 66. Can be subtracted to obtain a reproduction signal in which the nonlinear component is compensated.

FIG. 35 is a block diagram showing a modification of the nonlinear component correction signal output mechanism 31. Here, the output selector 13b has the same function as the output selector 73, and selects from the path metric memory 74 one surviving path corresponding to the minimum path metric value, that is, most likely. However, the output selector 73 is different in that the judgment value is output from the final stage of the path memory 76 (for example, the path memory cell 90 _{16 in} FIG. 14), and the provisional judgment output of the path memory 76 as can be seen from the connection. Among the outputs from the path memory cell 90 _{4 in} FIG. 14, for example, a value corresponding to the selected path is output as binarized data (temporary determination value). As a result, although the error rate is inferior to the final stage of the path memory, the tentative determination result is inferior in latency to the determination value output in the decoder, so the delay time in the PLL loop is shortened and the PLL is stabilized. To do.

  By outputting the reproduction signal thus obtained to the PLL 67 for clock extraction as shown in FIG. 26, the PLL 67 can stably extract the reproduction clock and reduce bit errors. it can. Further, by monitoring the eye pattern / jitter value of the reproduced signal, it can be used as an index of the reproduced signal quality of the system.

  Here, the PLL 67 will be further explained. FIG. 29 shows a detailed block diagram of the PLL 67. The input RF signal is binarized at an appropriate signal level by the slicer 113 and enters the edge detection phase comparator 112. Here, with respect to a reference frequency signal from a VCO (Voltage Controlled Oscillator Voltage Controlled Oscillator) 110 and a signal input from the slicer 113, the phase difference between the rising and falling edges of both is obtained. The loop filter 111 determines a loop time constant of the PLL 67, and is configured by an LPF (low-pass filter), a lead lag filter, or the like. With the above configuration, the PLL 67 extracts the recovered clock.

  Here, because the PLL 67 uses a slicer and an edge detection phase comparator, the CN ratio of the shortest mark is small, and the operation of the PLL becomes unstable during reproduction of high-density recording or reproduction using super-resolution. This causes an increase in bit errors. Therefore, the above problem can be solved by using the cross-correlation PLL 67 b shown in FIG. 31 instead of the PLL 67. In addition, the configuration of the decoder 28e is also changed from the configuration shown in FIG. 26 to the configuration shown in FIG.

Here, since the PR intersymbol interference generator 67a has the same function as the PR intersymbol interference generator 10 of FIG. 22, instead of adding the PR intersymbol interference generator 67b, the PR intersymbol interference of FIG. The output of the creator 10 may be used. Further, the PR class used for the PR intersymbol interference generator 67b is preferably as close as possible to the PR class used in the equalization circuit 65. However, for simplification of the circuit, for example, PR (1), that is, the binarized data is passed through. Even with such a configuration, it can operate sufficiently. However, in order to use the cross-correlation operation later, in any PR class, the DC (direct current) component of the signal is cut, that is, for example, a binary signal of 1 and 0 is set to 1 and −1 in PR (1). It is necessary to perform processing.

A description of the cross-correlation PLL 67b will be given below. The cross-correlation phase comparator 114 performs phase comparison between the outputs of the PR intersymbol interference generator 67a, the interpolator 64, the equalization circuit 65, or the decision feedback nonlinear Viterbi decoder 69 with a compensation signal output. FIG. 32 shows the configuration of the cross correlation phase comparator 114. A signal (input A) from the PR intersymbol interference generator 67a is delayed by 2T (two clocks) by the multiplier 101a and the delay device 100b and input to the multiplier 101b. Further, a signal (input B) from the interpolator 64, the equalization circuit 65, or the decision feedback nonlinear Viterbi decoder 69 with a compensation signal output is delayed by 1T (one clock) by the delay device 100a, and the multiplier 101a , And input to the multiplier 101b. That is, the product of the input A and the input B delayed by 1T, the product of the input A delayed by 1T and the input B delayed by 2T are calculated, and the difference is obtained by the subtractor 102. In consideration of the fact that there is a loop filter 111 that is an LPF (low-pass filter) thereafter, this operation is performed by cross-correlation between input A and input B delayed by 1T, input A delayed by 1T, and input delayed by 2T. The difference of the cross correlation of B is taken. That is, the cross-correlation phase comparator 114 and the loop filter 111 collectively function as two cross-correlation calculation circuits.

  Here, the above equation (4) is an arithmetic expression for cross-correlation. x is input A, y is input B, t is time, and R is a cross-correlation value. K is an integer. Here, Σ (sigma) in the equation is played by the loop filter 111 in the configuration of FIG. If there is no phase difference between the input A and the input B, the cross-correlation value is maximized. Therefore, it is sufficient to lock the PLL so that the cross-correlation value is maximized. However, in order to construct a feedback loop, it is necessary to obtain a phase error signal in which the error signal becomes 0 when the phase difference is 0 and the error amount sign can be recognized. is there. Therefore, in this configuration, utilizing the fact that the cross-correlation value is symmetric about the zero phase error amount, the cross-correlation between the input A and the input B shifted by 1T, and the input shifted from the input A by 1T. A phase comparison is performed by obtaining a difference from the cross-correlation of B to obtain the necessary phase error signal. In other words, the phase error signal = R (1) −R (−1) when expressed using the equation (4). FIG. 33 shows an operation flowchart of the cross-correlation phase comparator 114 that summarizes the above operations. After passing through the loop filter 111, a signal having a frequency corresponding to the phase error amount is reproduced by the VCO 110, and a reproduction clock of the RF signal can be extracted.

Here, the output of the interpolator 64, the equalization circuit 65, or the decision feedback nonlinear Viterbi decoder 69 with a compensation signal output can be used as the input of the cross-correlation PLL 67b. In particular, the decision with the compensation signal output is obtained by removing the nonlinear component. By using the output of the feedback type non-linear Viterbi decoder 69, it becomes possible to perform phase comparison with a signal from which non-linear components have been removed, so that more stable operation of the PLL becomes possible.

Although not shown in the figure, the output of the interpolator 64, the equalization circuit 65, or the decision feedback nonlinear Viterbi decoder 69 with a compensation signal output, and the output of the corresponding PR intersymbol interference generator 67a, respectively. Is appropriately inserted in front of the cross-correlation PLL 67b.

In the upper input in FIG. 31 to the cross-correlation PLL 67b, in the case of the configuration of the decoder 28e (FIG. 30) using the interpolator 64 for adjusting the reproduction clock timing, the interpolator as described above. 64 or is the output of the equalizer circuit 65 or compensation signal output with the decision feedback nonlinear Viterbi decoder 69 can be used, the upper input in Fig. 31 to the cross-correlation PLL67b, subsequent reproduction clock timing adjustment, the decoder For example, in the case of a decoder configuration that adjusts the reproduction clock timing in the ADC where the reproduction clock generated by the PLL is the sampling clock of the ADC, the cross-correlation PLL 67b is used. The upper input in FIG. 31 is any stage after the ADC and before the decoder. Of course, good.

Further, in this configuration, because of the use of the binarization determination result of the decoder 69, the PLL does not operate correctly when the determination result is incorrect. Therefore, the compensation value for nonlinear correction for reducing the error rate of the decoder 69 and the waveform equalization coefficient in the equalization circuit 65 are important. Therefore, the compensation value and waveform equalization coefficient are first learned for the low density recording density described above, and then reproduction is performed at a slightly higher recording density. Then, the compensation value is learned again, and the waveform equalization coefficient is calculated. A method of obtaining a compensation value / waveform equalization coefficient for reproduction with a high density step by step, such as updating, and a method of setting the rewrite frequency of the compensation value to a value equal to or lower than the open loop crossing frequency of the feedback loop in the PLL 67 Using PLL increases the stability of the PLL.

If the optical disc has the training area in which a known bit pattern is recorded, the known bit pattern is used as an input to the cross-correlation PLL 67b instead of using the binarization determination result of the decoder 69. PLL stability increases. A configuration example of the decoder 28e in that case is shown in FIG. The switch 67d can switch the output of the decoder 69 and the output of the known bit pattern output device recorded in the training area. Thus, in order to learn the compensation value for nonlinear correction and the waveform equalization coefficient in the equalization circuit 65, when reproducing the training area, the PLL is stabilized using a known bit pattern, and an appropriate compensation value or waveform is obtained. An equalization coefficient can be obtained.

  Further, by setting the PR class in the PR intersymbol interference generator 10 to a class different from the PR class in the FIR filter 63, for example, the PR class in the FIR filter 63 is changed to PR (1, 2, 2, 2, 1), When PR (1, 2, 2, 2, 1) is also used as the PR class in the PR intersymbol interference generator 10, there is no 2T mark carrier, and it is difficult to extract a recovered clock by a PLL using a slicer. However, the PR class in the FIR filter 63 is PR (1, 2, 2, 2, 1), and the PR class in the PR intersymbol interference generator 10 is changed to PR (0, 1, 2, 1, 2). 0) and PR (1, 1, 2, 1, 1), and so on, by setting the two PR classes to different classes, the signal supplied to the PLL is PR (0, 1, 2, 1, 0) and PR (1, 1, 2, 1, 1), and the PLL operates normally.

The PR bit length is not related to the ML bit length. For example, the ML is 5 bits as shown below, but PR (1, 2, 3, 4, 4, 3, 2, 1) The PR bit length may be 8 bits longer than the ML bit length. In that case, it is possible to use the branch metric calculation formula shown in the following formula (5). However, n is an integer of 1 to 16, PR is a matrix [12344] ^T , and RF is a value of an RF signal waveform-equalized to a PR characteristic as an input signal. PP (n) is a 5-bit bit string corresponding to each branch B (n). CV (n) is a compensation value corresponding to each branch B (n). In this way, the upper 5 bits (the same number of bits as the ML length) are used as the PR characteristics when calculating branch metrics. In this state, the pattern compensation value is learned. By doing so, the intersymbol interference is reproduced by the pattern compensation value for the lower 3 bits.

BM _t (n) = (PP (n) × PR + CV (n) −RF) ² (5)

Further, it is possible to use a provisional determination result for the lower 3 bits of PR characteristics using the branch metric calculation formula shown in the following formula (6). However, n is an integer of 1 to 16, PR is a matrix [12334421] ^T , and RF is a value of an RF signal waveform-equalized to a PR characteristic as an input signal. Further, the upper 5 bits (the same number of bits as the ML length) of PP (n) are the bit string corresponding to each branch B (n), and the lower 3 bits (the remaining number of bits) are 3 bits corresponding to each branch (n). Is a temporary determination bit string. CV (n) is a compensation value corresponding to each branch B (n).

BM _t (n) = (PP (n) × PR + CV (n) −RF) ² (6)

  As described above, it is possible to use a long PR class that is more suitable for a super-resolution reproduction signal, and further reduce the bit error rate.

  In the above embodiment, the optical disk apparatus capable of only reproducing information has been described. However, the present invention is not limited to this, and any optical disk apparatus capable of reproducing at least information among recording, reproducing, and erasing of information may be used.

  Moreover, although the case where the optical pickup device includes one light source has been described in the above embodiment, the present invention is not limited thereto, and for example, a plurality of light sources that emit light having different wavelengths may be included.

  As described above, the information reproducing apparatus, information recording / reproducing apparatus, information reproducing method, and optical disc of the present invention are suitable for reproducing information with a recording mark pitch smaller than the diffraction limit with high accuracy.

1 is a block diagram showing a configuration of an optical disc device according to an embodiment of the present invention. It is a figure for demonstrating a super-resolution optical disk. It is a figure for demonstrating the relationship between the reproduction power and CNR in a super-resolution optical disk. FIG. 4A is a diagram for explaining a minute opening region, and FIG. 4B is a diagram for explaining a minute mask region. It is a figure for demonstrating the effect | action of a micro mask area | region. It is a figure for demonstrating the relationship between the light intensity distribution of the reflected light in a super-resolution optical disk, and reproduction power. FIG. 7A to FIG. 7B are conceptual diagrams showing clock timing and light intensity of reflected light from the optical disc. It is a block diagram of the decoder 28e. It is a block diagram of Viterbi decoder 66 '. 6 is a state transition diagram corresponding to a PR class PR (1, 2, 2, 2, 1) used in the optical disc apparatus 20. FIG. FIG. 11 is a trellis diagram corresponding to the state transition diagram of FIG. 10. 3 is a block diagram of a path memory 72. FIG. 6 is a block diagram of a decision feedback Viterbi decoder 66. FIG. It is a block diagram of the path memory with provisional determination function. It is a figure for demonstrating acquisition of a temporary determination result. 3 is a block diagram of a pattern compensation memory 75. FIG. FIGS. 17A to 17C are diagrams for explaining a method of determining the compensation value CV (n). 18A is a diagram showing an eye pattern of an output signal of the ADC 62, and FIG. 18B is a diagram showing an eye pattern of an output signal of the asymmetric FIR filter 63. As shown in FIG. FIG. 19A is a diagram showing an eye pattern of an output signal from the decision feedback Viterbi decoder 66 when there is no decision feedback bit, and FIG. 19B is a decision feedback when the decision feedback bit is 3 bits. It is a figure which shows the eye pattern of the output signal of the type | mold Viterbi decoder 66. FIG. It is a figure which shows the comparison of the effect of a prior art and this invention. It is a figure which shows the graph of the relationship between the nonlinear correction bit number and error rate in the shortest recording mark length of about 162 nm. 4 is a diagram for explaining a compensation value learning mechanism 30. FIG. FIG. 23A and FIG. 23B are diagrams for explaining each area of the optical disc 15. Variance value with a B (6) is a diagram for a table explaining the compensation value set storage memory 51 _6. 3 is a block diagram of a nonlinear component correction signal output mechanism 31. FIG. It is a block diagram which shows the modification of the decoder 28e. It is a flowchart for demonstrating the process performed when performing high-density recording. 3 is a diagram showing a TOC area of the optical disc 15. FIG. It is a block diagram of PLL67. It is a block diagram of the decoder 28e using the cross correlation PLL67b. It is a block diagram of cross correlation PLL67b. 4 is a block diagram of a cross correlation phase comparator 114. FIG. 4 is an operation flowchart of a cross-correlation phase comparator 114. It is a block diagram which shows the modification of the decoder 28e using the cross correlation PLL67b. FIG. 11 is a block diagram showing a modification of the nonlinear component correction signal output mechanism 31.

Explanation of symbols

DESCRIPTION OF SYMBOLS 15 ... Optical disk, 20 ... Optical disk apparatus, 23 ... Optical pick-up apparatus, 28d ... RF signal generation circuit, 28e ... Decoder, 65 ... Equalization circuit , 66 ... Decision feedback type Viterbi decoder, 75 ... Pattern compensation memory, 76 ... Temporary Path memory with determination function, 77... Branch metric calculator with compensation function, 12... Delay unit, 13... Nonlinear component compensator, 67.

Claims (30)

An information reproducing apparatus for reproducing information using a PRML method,
An optical head including a light source, an optical system including an objective lens that collects light emitted from the light source on an optical disc, and a photodetector that receives reflected light from the optical disc;
Signal generating means for generating an RF signal from the output signal of the photodetector;
A waveform equalizer that equalizes the RF signal to a preselected partial response characteristic of the PRML method for the optical disc in which information is recorded with a recording mark having a pitch smaller than the diffraction limit;
A pattern compensation memory in which a plurality of compensation values for each bit pattern of the recording mark are recorded;
A path memory having a function of outputting past determination results for a plurality of states according to the PRML method;
A branch metric calculator that calculates a likelihood of each received signal of the waveform equalized RF signal using a compensation value according to the past determination result among the plurality of compensation values. apparatus.   2. The information reproducing apparatus according to claim 1, wherein the pattern compensation memory stores an average value of waveform equalization errors for each bit pattern of the waveform equalized RF signal as a compensation value. The information reproducing apparatus according to claim 2, wherein information relating to an average value of the waveform equalization error amount is recorded on the optical disc . The pattern compensation memory is added to the average value of the waveform equalization error amount, the information reproducing apparatus according to claim 2 or 3, characterized in that storing a variance value of the waveform equalization error amount as the compensation value. 5. The information reproducing apparatus according to claim 4, wherein information relating to a variance value of the waveform equalization error amount is recorded on the optical disc .   The pattern compensation memory includes a second partial response characteristic different from the partial response characteristic, an ideal reproduction signal calculated by a convolution operation with recorded bit data, and the waveform equalized RF signal. The information reproducing apparatus according to claim 1, wherein an average value of the error amounts is stored as a compensation value for each bit pattern. The pattern compensation memory is information reproducing apparatus according to any one of claims 1 to 6, characterized in that storing a compensation value corresponding to a bit pattern of 8 bits or more. An RF signal read from the optical disc, according to claim 1 to 7, further comprising a clock extraction circuit for extracting a reproduction clock signal having a cross-correlation calculation circuit for performing phase comparison signal and decoded by the PRML system the RF signal The information reproducing apparatus according to any one of the above. In the clock extraction circuit, a first cross-correlation operation circuit that calculates a cross-correlation for an RF signal read from the optical disc and a signal obtained by delaying the RF signal by the PRML method for a first time; An RF signal read from the optical disc; and a second cross-correlation operation circuit that calculates a cross-correlation for a signal obtained by delaying the RF signal decoded by the PRML method for a second time, and a first time delay And a subtractor for setting the amount larger than the second time delay amount and taking a difference between the output of the first cross-correlation operation circuit and the output of the second cross-correlation operation circuit. 8. The information reproducing apparatus according to 8 . And known bit pattern recorded on the optical disc, the information reproducing apparatus according to claim 1 to 9 to perform clock extraction step of extracting a reproduction clock signal with the RF signal read from the optical disc. 10. The information according to claim 8 , wherein the clock extraction circuit extracts a reproduction clock signal using a known bit pattern recorded on the optical disc instead of the signal decoded by the PRML method. Playback device. A delayer for delaying the waveform-equalized RF signal by a time required to decode the RF signal by the PRML method;
And a nonlinear component compensator that compensates for a nonlinear component of the RF signal based on a difference between a value of the RF signal delayed by the delay unit and a compensation value stored in the pattern compensation memory. the information reproducing apparatus according to any one of 1-11. 13. The information reproducing apparatus according to claim 12 , wherein the nonlinear component compensator includes an output selector that performs provisional determination using past determination results for the plurality of states in the path memory. 14. The information reproducing apparatus according to claim 12 , further comprising a clock extraction circuit that extracts a reproduction clock signal from an output signal of the nonlinear component compensator. Compensation value information related to the compensation value is recorded on the optical disc, and the pattern compensation memory includes at least one of recording density, reproduction power, linear velocity during reproduction, and drive waveform of the light source, and the optical disc. the information reproducing device according to the compensation value information recorded in any one of claims 1 to 14, characterized in that to set the compensation value to. The bit length of the partial response characteristic, the information reproducing apparatus according to any one of claims 1 to 15, characterized in that longer than the bit length of the state to perform maximum likelihood estimation. The information reproducing apparatus according to claim 1, wherein the optical disc has a training area in which a known bit pattern is recorded. An information reproduction method for reproducing information using a PRML method,
Reading an RF signal from an optical disc in which the information is recorded with recording marks having a pitch smaller than the diffraction limit;
A waveform equalization step for equalizing the RF signal to a preselected partial response characteristic of the PRML method;
A compensation value calculating step of calculating a compensation value for each bit pattern of the waveform equalized RF signal;
A branch metric value calculating step of calculating a likelihood of each received signal of the waveform equalized RF signal using the compensation value according to a past determination result for a plurality of states according to the PRML method; Information reproducing method including: 19. The information reproducing method according to claim 18 , wherein, in the compensation value calculating step, an average value of waveform equalization error amounts for each bit pattern corresponding to the partial response characteristics is calculated as a compensation value. In the compensation value calculation step, an ideal reproduction signal calculated by convolution with the second partial response characteristic, which is different from the partial response characteristic, and the recording bit data, and the waveform equalized RF signal 19. The information reproducing method according to claim 18 , wherein an average value of error amounts is calculated as a compensation value. 19. The information reproducing method according to claim 18 , wherein, in the compensation value calculating step, a variance value of the waveform equalization error amount is calculated as a compensation value in addition to the average value of the waveform equalization error amount. For phase comparison, and an RF signal read from the optical disc, the clock extraction process and for extracting a reproduction clock signal having a cross-correlation calculation step of a signal obtained by decoding the RF signal with the PRML method; further comprising a claim 18 21. The information reproducing method according to any one of 21 . A first cross-correlation calculating step of calculating a cross-correlation for an RF signal read from the optical disc and a signal obtained by delaying the RF signal decoded by the PRML method for a first time in the clock extracting step; A second cross-correlation calculating step for calculating a cross-correlation between an RF signal read from the optical disc and a signal obtained by delaying the RF signal decoded by the PRML method for a second time, and a first time delay amount And a subtraction step for taking a difference between an output of the first cross-correlation calculation step and an output of the second cross-correlation calculation step. 22. An information reproducing method according to 22 . The clock extracting step of extracting a reproduction clock signal using a known bit pattern recorded on the optical disc and an RF signal read from the optical disc, according to any one of claims 18 to 23 . Information reproduction method. In the clock extraction step, any one of claims 22 to 24, characterized in that extracting a reproduction clock signal with a known bit pattern recorded on the optical disk instead of the decoded signal by the PRML method The information reproduction method according to item. A delaying step of delaying an RF signal read from the optical disc by a time required to decode the RF signal by the PRML method;
A nonlinear component compensation step of compensating a nonlinear component of the RF signal based on a difference between the value of the RF signal delayed in the delay step and the compensation value;
The information reproduction method according to any one of claims 18 to 25 , further comprising a step of extracting a reproduction clock from the RF signal in which the nonlinear component is compensated in the nonlinear component compensation step. 27. The information reproducing method according to claim 26 , further comprising: an output selection step of performing a provisional determination of an RF signal using the past determination result in the nonlinear component compensation step. The information reproduction method according to any one of claims 18 to 27 , wherein a bit length of the partial response characteristic is longer than a bit length of the state in which maximum likelihood estimation is performed. In the compensation value calculation step, information is reproduced using a first compensation value calculation step for obtaining a compensation value at a low density recording density, and the compensation value obtained in the first compensation value calculation step. The information reproduction method according to any one of claims 18 to 27 , further comprising a second compensation value calculation step for obtaining a compensation value at a high recording density. An information recording / reproducing apparatus for recording and reproducing information with respect to an optical disc,
An information reproducing apparatus according to any one of claims 1 to 17 ;
An information recording / reproducing apparatus comprising: an information recording apparatus for recording information on the optical disc with recording marks having a pitch smaller than a diffraction limit.