JP4885799B2 - Signal evaluation apparatus, signal evaluation method, signal evaluation program, and computer-readable recording medium - Google Patents

Description

  The present invention relates to a signal evaluation apparatus, a signal evaluation method, a signal evaluation program, and a computer-readable recording medium for evaluating the signal quality of a two-dimensional signal in an apparatus for Viterbi decoding a two-dimensional signal such as a hologram memory reproduction signal.

  In recent years, a hologram memory that records and reproduces information two-dimensionally, that is, as page data using holography, is attracting attention as a next-generation high-density recording / reproducing system. As shown in FIG. 17, in this system, when recording information data, a recording signal is recorded based on the information data to be recorded using a spatial light modulator 60 (for example, a liquid crystal panel or the like) composed of a plurality of pixels. The light is spatially modulated to generate a two-dimensional interference fringe by causing the recording signal light passing through the lens 61 and the coherent recording reference light to interfere with each other, and an image formed by the interference fringe is used as two-dimensional information. Recording on the hologram medium 62. The hologram medium 62 includes a rewritable medium using an inorganic photorefractive crystal typified by lithium niobate and a write-once medium using a photopolymer that is an organic polymer material.

  On the other hand, when reproducing the information data by reading the recorded interference fringes from the hologram medium 62, the reflection generated by irradiating the hologram medium 62 with the reproduction reference light at the same incident angle as the recording reference light. Light or transmitted light is received through a lens 63 by a light receiving element 64 (for example, a CCD) having a plurality of pixels to generate a reproduction signal, and the original information data is reproduced using the generated reproduction signal. Since reproduction is performed in units of two-dimensional page data in this way, it is possible to significantly improve the reproduction speed over conventional optical discs that perform reproduction in one dimension.

  By the way, in the hologram memory system, various noises may be mixed in the recording process and the reproducing process (in particular, noises are often generated due to inhomogeneity of the recording medium). Signal quality deteriorates. In addition, there is an influence of a reproduction signal from adjacent pixels, that is, inter-pixel interference, which also deteriorates the signal quality.

  Therefore, conventionally, methods for accurately reproducing information by reducing the influence of noise and inter-pixel interference have been studied. In particular, a technique for Viterbi decoding of a reproduction signal output from a light receiving element has been studied.

  Patent Document 1 discloses a decision feedback Viterbi decoding method. This method is a method of Viterbi decoding and is called a Decision Feedback Viterbi Algorithm.

On the other hand, as a system for Viterbi decoding a one-dimensional reproduction signal, there are a hard disk and an optical disk. In such a system, a value called SAM (Sequenced Amplitude Margin) has been proposed as an evaluation value for evaluating the quality of a reproduction signal. Although the SAM value itself can be used to determine the local quality of the reproduced signal, it is necessary to make a determination based on the frequency distribution of the SAM value in order to evaluate the quality of the entire reproduced signal. In the case of a one-dimensional reproduction signal, the frequency distribution of SAM values has a shape in which normal distributions having a plurality of discrete ideal SAM values as an average value are combined. The ideal SAM value is a SAM value obtained for an ideal waveform signal having an assumed partial response (PR) characteristic. Among these normal distributions, only those that are likely to cause an error, that is, those having a small ideal SAM value are classified and extracted, and then the average value and standard deviation are determined based on the characteristics of the probability density distribution function of the normal distribution. A method has been proposed in which a quality evaluation value that is theoretically equivalent to an error rate such as the ratio of the above (Patent Documents 2 and 3) or an estimated error rate (Patent Document 4) is calculated.
US Pat. No. 5,740,184 (issued April 14, 1998) JP 10-21651 A (publication date January 23, 1998) JP 2003-141823 A (publication date May 16, 2003) JP 2003-272304 A (publication date September 26, 2003)

  As described above, in a system that performs Viterbi decoding of a one-dimensional reproduction signal, a method for calculating a signal quality evaluation value equivalent to the error rate of the reproduction signal from a frequency distribution of SAM values has been proposed. However, as a method for calculating a signal quality evaluation value equivalent to the error rate based on the SAM value of the two-dimensional signal, it is impossible to apply the conventional methods such as Patent Documents 2 to 4 as they are because of the following problems. Met.

  Each of Patent Documents 2 to 4 is a system that performs Viterbi decoding of a one-dimensional reproduction signal. In particular, a specific pattern that easily causes an error (that is, a pattern corresponding to a specific ideal SAM value having a small value) is examined in advance. In addition, a configuration for detecting this specific pattern from the recording bits is essential. The number of these specific patterns is from several to several tens at most, and the number of corresponding ideal SAM values is from one to at most about three.

  On the other hand, in the case of the SAM value in the two-dimensional Viterbi decoding in which the trellis diagram is more complex than the one-dimensional, the number of types of specific patterns that are particularly prone to errors increases. If it is to be realized, the circuit scale becomes extremely large, which is practically impossible.

  The present invention has been made in view of the above problems, and its purpose is to provide an evaluation value equivalent to an error rate, which can evaluate the signal quality with high reliability when Viterbi decoding a two-dimensional reproduction signal. An object of the present invention is to provide a signal evaluation apparatus, a signal evaluation method, a signal evaluation program, and a computer-readable recording medium.

  In order to solve the above problems, the signal evaluation apparatus according to the present invention performs Viterbi decoding assuming a predetermined partial response characteristic on the input signal in the signal evaluation apparatus that outputs the quality evaluation value of the input signal. Viterbi decoding means for outputting a path metric difference between a correct answer path and an error path, and an ideal path metric difference corresponding to the input signal obtained from the plurality of path metric differences output from the Viterbi decoding means. Extraction means for extracting a path metric difference belonging to a predetermined value range, and quality evaluation means for calculating a quality evaluation value of the input signal based on a group of path metric differences extracted by the extraction means. It is characterized by that.

  In addition, in order to solve the above-described problem, the signal evaluation method according to the present invention performs Viterbi decoding assuming a predetermined partial response characteristic in the input signal in the signal evaluation method for outputting the quality evaluation value of the input signal. The Viterbi decoding step for outputting the path metric difference between the correct path and the error path, and the ideal path metric corresponding to the input signal obtained from the path metric difference output in the Viterbi decoding step. An extraction step for extracting a path metric difference in which the difference belongs to a predetermined value range; and a quality evaluation step for calculating a quality evaluation value of the input signal based on a group of path metric differences extracted in the extraction step. It is characterized by including.

  Here, the predetermined partial response characteristic is used to calculate the ideal path metric difference, and is expressed as, for example, a two-dimensional impulse response represented by a 3 × 3 matrix.

  The predetermined value range is a predetermined value determined for an assumed partial response characteristic, and defines a range for classification into a group to which the ideal path metric difference belongs.

  As a result, the extraction unit can extract only a group of path metric differences corresponding to the extracted ideal path metric difference from among the path metric differences that are input signals Viterbi decoded by the Viterbi decoding unit. Therefore, the quality evaluation unit calculates a quality evaluation value of a group of path metric differences corresponding to the ideal path metric difference extracted by the extraction unit among the path metric differences decoded by the Viterbi decoding unit. be able to.

  That is, the quality evaluation value is calculated only from the path metric difference belonging to the predetermined value range among the path metric differences as the input signal.

  For this reason, in order to calculate the quality evaluation value, it is not necessary to calculate all the path metric differences that are the input signals.

  Therefore, it is possible to efficiently evaluate the performance of recording media, recording devices, and playback devices used for recording and playback of two-dimensional playback signals while maintaining reliability. You can do it properly.

  Further, in the signal evaluation apparatus according to the present invention, the quality evaluation unit calculates a statistical value representing a degree of spread of the frequency distribution of the extracted path metric difference group, and the quality of the input signal is calculated from the statistical value. It is preferable to calculate an evaluation value.

  The statistical value indicating the extent of the frequency distribution has a strong correlation with the quality evaluation value of the input signal. Therefore, the quality evaluation means can efficiently calculate the quality evaluation value of the input signal while maintaining reliability.

  Further, in the signal evaluation apparatus according to the present invention, the quality evaluation unit calculates a standard deviation of the group of the extracted path metric differences as the statistical value, and the standard deviation as the quality evaluation value, It is preferable to calculate a ratio with the average value of the group of extracted path metric differences.

  As a result, a quality evaluation value that theoretically corresponds to the error rate can be calculated based on the average value and standard deviation of the frequency distribution of the path metric difference of the input signal.

  Further, in the signal evaluation apparatus according to the present invention, the quality evaluation unit calculates a standard deviation of the group of the extracted path metric differences as the statistical value, and the standard deviation and a predetermined value as the quality evaluation value It is preferable to calculate a ratio with a constant value.

  The predetermined constant value can be arbitrarily set. For example, the average value of the ideal path metric difference of the corresponding group can be obtained in advance and used. Thereby, the function which calculates | requires an average value can be abbreviate | omitted.

  In addition, in order to solve the above problems, the signal evaluation apparatus according to the present invention performs Viterbi decoding assuming a predetermined partial response characteristic in the input signal in the signal evaluation apparatus that outputs the quality evaluation value of the input signal. Viterbi decoding means for outputting a path metric difference between a correct path and an error path, and an ideal path corresponding to the input signal obtained from the plurality of path metric differences output from the Viterbi decoding means. Corresponding to the input signal obtained by obtaining a path metric difference from a plurality of path metric differences output from the Viterbi decoding means and a first extraction means for extracting a path metric difference whose metric difference belongs to a predetermined first value range A second extraction for extracting a path metric difference belonging to a predetermined second value range in which an ideal path metric difference to be different from the first value range. And means, based on the group of path metric differences extracted respectively by said first and second extracting means is characterized by comprising: a quality evaluation unit for calculating the quality evaluation value of the input signal.

  Thereby, the quality evaluation means calculates the quality evaluation value from the path metric difference belonging to the first value range and the path metric difference belonging to the second value range. For this reason, compared with the case where the said quality evaluation value is calculated only from the path metric difference which belongs to the 1st value range, a quality evaluation value can be calculated more correctly.

  Further, the signal evaluation apparatus according to the present invention uses a partial response characteristic ideal waveform signal in accordance with a partial response characteristic assumed in the Viterbi decoding means by using a test bit corresponding to the input signal for the ideal signal associated with the input signal. By performing Viterbi decoding according to the partial response characteristics on the ideal response signal generated by the ideal signal generating means for converting to the partial response characteristics ideal waveform signal converted by the ideal signal generating means, a path metric difference between the correct path and the error path is obtained. It is preferable that an ideal path metric difference generating unit that generates the ideal path metric difference is provided.

  Accordingly, it is possible to reliably identify the ideal SAM value corresponding to the two-dimensional signal and accurately classify the frequency distribution of the path metric difference of the two-dimensional signal for each ideal SAM value. Thereby, the possibility that an accurate evaluation value can be calculated can be further increased.

  In the signal evaluation apparatus according to the present invention, the Viterbi decoding means outputs a decoded bit corresponding to the path metric difference together with the path metric difference, and the Viterbi decoding is performed on the decoded bit corresponding to the input signal. An ideal signal generating means for converting into a partial response characteristic ideal waveform signal in accordance with a partial response characteristic assumed in the means, and Viterbi decoding is performed on the partial response characteristic ideal waveform signal converted by the ideal signal generating means in accordance with the partial response characteristic Accordingly, it is preferable to include ideal path metric difference generation means for generating the ideal path metric difference that is a path metric difference between the correct path and the error path.

  Here, the decoded bit is used to select the path metric of the correct path when Viterbi decoding means performs Viterbi decoding on the input signal.

  Thereby, in order to specify the ideal SAM value corresponding to the two-dimensional signal, a memory or the like for storing bits corresponding to the two-dimensional signal is prepared separately. Since it is not necessary to prepare a mechanism for reading bits, the quality evaluation value of a two-dimensional signal can be calculated with a simple configuration.

In this case, there is a possibility that the correspondence between the two-dimensional signal and the ideal SAM value may be made erroneously, but the ratio is almost the same as the error rate and is very small, usually 10 ⁻² or less. The influence on the calculation results such as the average and standard deviation is extremely small. Therefore, errors included in quality evaluation values such as finally calculated error rates are small enough to be ignored.

  In the signal evaluation apparatus according to the present invention, an ideal path metric difference corresponding to the input signal obtained by the extracting unit in association with the path metric difference output from the Viterbi decoding unit is obtained. A storage unit preliminarily storing a flag map indicating whether or not it belongs to a predetermined value range, and determining whether or not to extract a path metric difference output from the Viterbi decoding unit based on the flag map; Is preferred.

  Accordingly, it is possible to reliably identify the ideal SAM value corresponding to the two-dimensional signal and accurately classify the frequency distribution of the path metric difference of the two-dimensional signal for each ideal SAM value. Thereby, there exists an effect which can improve possibility that an exact evaluation value can be calculated more. Furthermore, since it is not necessary to separately provide a Viterbi decoding circuit for generating an ideal SAM value, there is an effect that can be realized with a very simple configuration.

  The above may be realized by a computer. In this case, a Viterbi decoding program for realizing the above in the computer by operating the computer as each of the above means and a computer-readable recording medium recording the Viterbi decoding program are also included in the scope of the present invention.

  As described above, the signal evaluation apparatus according to the present invention performs Viterbi decoding on the input signal assuming a predetermined partial response characteristic, thereby outputting Viterbi decoding means for outputting a path metric difference between a correct path and an error path. Extraction means for extracting a path metric difference in which an ideal path metric difference corresponding to the input signal obtained from the plurality of path metric differences output from the Viterbi decoding means belongs to a predetermined value range; and Quality evaluation means for calculating a quality evaluation value of the input signal based on a group of path metric differences extracted by the extraction means.

  The signal evaluation method according to the present invention also includes a Viterbi decoding step of outputting a path metric difference between a correct path and an error path by performing Viterbi decoding assuming a predetermined partial response characteristic on an input signal, and the Viterbi decoding step. An extraction step for extracting a path metric difference in which an ideal path metric difference corresponding to the input signal from which the path metric difference is obtained belongs to a predetermined value range is extracted from the path metric difference output from the decoding step, and extracted by the extraction step. A quality evaluation step of calculating a quality evaluation value of the input signal based on the group of path metric differences.

  Therefore, since the quality evaluation value of the input signal can be calculated only from the path metric difference belonging to the extracted predetermined value range, when the two-dimensional reproduction signal is Viterbi-decoded, the signal quality is evaluated with high reliability. There is an effect that can be done.

Embodiment 1
An embodiment of the present invention will be described with reference to FIGS. FIG. 1 is a block diagram illustrating a configuration of a hologram memory reproduction signal evaluation system 20 according to the first embodiment. The hologram memory reproduction signal evaluation system 20 includes a hologram memory reproduction device 1, a two-dimensional equalizer 2, and a signal evaluation device 3.

  The signal evaluation device 3 shown in FIG. 1 corresponds to the signal evaluation device described in the claims.

  The signal evaluation device 3 includes a two-dimensional SAM detection circuit (Viterbi decoding unit) 4, an ideal signal generation unit (ideal signal generation unit) 40, a two-dimensional SAM detection circuit (ideal path metric difference generation unit) 7, and an extraction unit (extraction unit). , First extraction means) 45 and a quality evaluation circuit (quality evaluation means) 50.

  The ideal signal generation unit 40 includes a test bit memory 5 and a PR filter 6. The extraction unit 45 includes a value range identification circuit 8 and an average value / standard deviation calculation circuit 9. The quality evaluation circuit 50 includes an average value / standard deviation calculation circuit 9 and a divider 10.

  Now, operations of reproduction and signal evaluation in the hologram memory reproduction signal evaluation system 20 shown in FIG. 1 will be described with reference to FIG. FIG. 2 is a flowchart showing a process flow of the hologram memory reproduction signal evaluation system 20 according to the first embodiment.

  In step S1, the hologram memory reproducing apparatus 1 reproduces a hologram medium on which a known test bit is recorded. That is, a reproduction signal (input signal) is output based on a known test bit. The specific reproduction operation in the hologram memory reproduction apparatus 1, that is, the operation in which the reproduction reference light is irradiated onto the hologram medium and the two-dimensional reproduction signal is output from the light receiving element is as described in the background art. Further, since the operation is not directly related to the present invention, the detailed description is omitted.

  As the test bit, a pattern (usually a random pattern) having a sufficient number of samples for obtaining an evaluation value of signal quality and having no bias so that the evaluation can be appropriately performed is selected. The test bit may be recorded in advance on the hologram medium, or may be recorded on the hologram medium before the signal quality is evaluated. Note that if the test bit is defined in the hologram memory standard, it is possible to use a common test bit between devices of different manufacturers, so that there is an effect that compatibility is improved.

  The test bit memory 5 stores the same bits as the test bits reproduced by the hologram memory reproducing apparatus 1. A test bit memory 5 is prepared with a memory for storing bits, and these bits are read out in accordance with the reproduction timing of the two-dimensional reproduction signal.

  Further, instead of the test bit memory 5, a pseudo random pattern generation circuit using a linear shift register may be used. As a result, the amount of hardware can be reduced.

  In step S 1, the two-dimensional reproduction signal output from the hologram memory reproduction apparatus 1 is subjected to waveform equalization by the two-dimensional equalizer 2. The two-dimensional equalizer 2 includes a two-dimensional FIR (Finite Impulse Response) filter and removes inter-waveform interference to bring the two-dimensional reproduction signal closer to the assumed PR characteristic.

  The above is the reproduction processing of the hologram memory reproduction signal evaluation system 20 in step S1.

  Next, in step S2, the two-dimensional SAM detection circuit 4 performs Viterbi decoding on the two-dimensional reproduction signal (input signal) and outputs a SAM value for each bit. A specific description of the operation of the two-dimensional SAM detection circuit 4 in step S2 will be described later.

  Next, in step S3, the two-dimensional SAM detection circuit 7 outputs an ideal SAM value corresponding to the SAM value output in step S2. A specific process will be described. A test bit corresponding to the two-dimensional reproduction signal is read from the test bit memory 5 and passes through the PR filter 6 to output a PR ideal waveform (partial response characteristic ideal waveform signal). The PR filter 6 is an FIR filter having PR characteristics assumed in the two-dimensional SAM detection circuits 4 and 7. Viterbi decoding is performed on the PR ideal waveform by the two-dimensional SAM detection circuit 7, and an ideal SAM value is output. Since the operation of the two-dimensional SAM detection circuit 7 is the same as that of the two-dimensional SAM detection circuit 4, it will be described later.

  Next, in step S4, the value range identification circuit 8 determines whether or not the ideal SAM value is within a predetermined range, that is, whether it is V1 or more and less than V2 (predetermined value range). V1 and V2 are predetermined values determined for each assumed PR characteristic, and define ranges for classifying ideal SAM values into a plurality of groups. FIG. 3 shows an example of the frequency distribution of ideal SAM values. It can be seen that the ideal SAM values have many different types, but can be classified into several groups. Since the reproduction signal corresponding to the ideal SAM value having a small value has a high probability of causing an error because of the property that the SAM value <0 corresponds to the generation of the Viterbi decoding error bit, the group 1 (predetermined in the case of FIG. If the SAM value for the reproduction signal corresponding to the ideal SAM value in the first value range is taken out and a frequency distribution is created, it is possible to evaluate with a high correspondence with the error rate. In the case of FIG. 3, if V1 = 3000 and V2 = 5000 are set, only 88 types of ideal SAM values of group 1 can be extracted.

  When it is determined in step S4 that the ideal SAM value is within the predetermined range, the process proceeds to step S5, and the corresponding SAM value is set as a calculation target in the average value / standard deviation calculation circuit 9. Specifically, the SAM value and the square error of the SAM value with respect to the temporary average value μ ′ are cumulatively added to calculate the average value μ and the temporary standard deviation σ ′, respectively. On the other hand, if it is determined in step S4 that the ideal SAM value is outside the predetermined range, the process in step S5 is skipped. In this way, SAM values belonging to a specific group can be determined and extracted based on the corresponding range of ideal SAM values, instead of being determined based on the pattern.

  Next, in step S6, it is determined whether or not the output of the SAM value for all test bits has been completed. If not completed, the process returns to step S2, so that the processes from step S2 to S6 are repeated until all the test bits are completed.

  When the end determination is made in step S6, the process proceeds to step S7. In step S 7, the average value / standard deviation calculation circuit 9 calculates μ and σ, and the divider 10 calculates the ratio σ / μ. The average value μ is obtained by dividing the sum of the SAM values cumulatively added in step S5 by the number. The standard deviation σ is obtained by dividing a temporary standard deviation σ ′ obtained by dividing the sum of the square errors of the SAM values with respect to the temporary average value cumulatively added in step S5 by the number by an accurately calculated average value μ ( It is obtained by correcting in 1). If the temporary average value μ ′ can be estimated to some extent accurately, the error between the true average values μ and μ ′ becomes smaller and σ≈σ ′, so the temporary standard deviation σ ′ is regarded as the standard deviation σ. Thus, the circuit may be simplified by omitting the correction of equation (1).

  By the above processing, finally, σ / μ can be obtained as the signal quality evaluation value of the reproduction signal of the hologram memory.

  In this embodiment, the ideal SAM value is obtained in real time using the test bit memory 5, the PR filter 6, the two-dimensional SAM detection circuit 7, and the value range identification circuit 8, and the value is within a predetermined range. This is a configuration in which SAM values belonging to a specific group are extracted. However, when a predetermined test bit is used, an ideal SAM value is obtained in advance, and a flag map indicating whether or not each reproduction signal corresponds to a SAM value belonging to a specific group is prepared in advance. A configuration may be adopted in which SAM values belonging to a specific group are extracted based on the flag map.

  FIG. 4C shows an example of the flag map. Assume that the ideal SAM value of the test bit in FIG. 4A is obtained as shown in FIG. When V1 = 3000 and V2 = 5000 are set as shown in FIG. 3, the flag map is determined as shown in FIG. The meaning of the flag indicates that “1” belongs to group 1 and “0” does not belong to group 1. In the actual signal evaluation apparatus, if there is only the flag map of FIG. 4C, a determination equivalent to the determination performed in step S4 is possible. Therefore, the test bit memory 5, the PR filter 6, the two-dimensional SAM detection circuit 7, and the value range identification circuit 8 are unnecessary, and a flag map memory 47 (storage unit) that stores a flag map is required instead. Then, the flag map memory 47 reads the flag map in accordance with the reproduction timing of the two-dimensional reproduction signal.

  FIG. 5 shows an example of the signal evaluation apparatus 30 provided with the flag map memory 47. The signal evaluation apparatus 30 includes a two-dimensional SAM detection circuit 4, an extraction unit 46, and a quality evaluation circuit 50.

  The extraction unit 46 includes an average value / standard deviation calculation circuit 9 and a flag map memory 47. The quality evaluation circuit 50 includes an average value / standard deviation calculation circuit 9 and a divider 10.

  Since the data amount of the flag map is the same as the data amount of the test bit, the circuit scale can be reduced by the PR filter 6, the two-dimensional SAM detection circuit 7, and the value range identification circuit 8. In particular, since the two-dimensional SAM detection circuit 7 has a large circuit scale, the reduction effect is large. Further, instead of storing the flag map in the memory, the flag map itself may be recorded as data on the hologram medium, and the flag map may be obtained by first reproducing the data.

  Next, the reason why the signal quality evaluation value corresponding to the error rate in Viterbi decoding can be obtained by the signal evaluation device 3 according to the present embodiment will be described in detail. FIG. 6A is a diagram illustrating an example of a frequency distribution of SAM values corresponding to the ideal SAM values of group 1 extracted in step S4. The three frequency distributions show cases where the quality of the reproduced signal is different and the bit error rate (BER) is 1E-5 or less, 1E-4, 5E-3, respectively. As can be seen from FIG. 6A, each frequency distribution has a distribution shape close to a normal distribution.

  According to statistics, if the probability (cumulative relative frequency) of x or less in the normal distribution of the mean value μ and the standard deviation σ is F (x), the cumulative density function Φ (z) for the standard normal distribution (Expression (2) )) And F (x), the relationship of Expression (3) is established.

  Now, because of the property that the SAM value <0 and Viterbi decoding error bit generation correspond, the cumulative relative frequency of the portion smaller than 0 in the frequency distribution of the SAM value has a very strong correlation with the error rate. As can be seen from FIG. 6 (a), when considering that the SAM value has a distribution shape substantially similar to a normal distribution, the error rate is calculated using the equation (3) as F (0) = Φ (−μ / It can be estimated fairly accurately as σ).

  From the above, it is theoretically explained that σ / μ (ratio of standard deviation and average value) calculated by the hologram memory reproduction signal evaluation system 20 according to the present embodiment theoretically corresponds to the error rate in Viterbi decoding. It was.

  Hereinafter, the operation of the two-dimensional SAM detection circuit 4 will be described in detail (the operation of the two-dimensional SAM detection circuit 7 is exactly the same).

(Two-dimensional SAM detection circuit 4)
The two-dimensional SAM detection circuit 4 applies decision feedback Viterbi decoding to the input two-dimensional reproduction signal. Accordingly, the decision feedback Viterbi decoding executed by the two-dimensional SAM detection circuit 4 will be described below.

  In this method, the reproduction signal is Viterbi-decoded along the horizontal direction of the page data, that is, the row direction. Specifically, the process of determining the decoded bit for each row is performed while shifting the row downward, that is, in the column direction. At that time, judgment feedback is performed. That is, the previous decoding result (feedback row) is used for Viterbi decoding of the next row.

  An outline of the decision feedback Viterbi decoding method will be described below with reference to FIGS. FIG. 7 is a diagram showing an impulse response of an assumed PR characteristic in the decision feedback Viterbi decoding method. FIG. 8 is a trellis diagram expressing the decision feedback Viterbi decoding method.

  The assumed PR characteristic is expressed as a two-dimensional impulse response h represented by a 3 × 3 matrix shown in FIG. The trellis state is defined as a 2 × 2 matrix. The trellis diagram is expressed as shown in FIG.

  The “two-dimensional impulse response” means an output signal that is output from the linear system when a unit impulse signal is input to the linear system. The unit impulse signal means a two-dimensional bit matrix in which (a) the signal level of only 1 bit is 1, and (b) the signal levels of other bits are all 0.

  In the two-dimensional impulse response h in FIG. 7, the signal level is generated at the center pixel. Not only that, it also occurs in neighboring neighboring pixels. That is, the occurrence of inter-pixel interference is assumed.

(Trellis state)
When the trellis state is represented by [k, l; m, n], each of k, l, m, and n takes a 1-bit value. That is, both are 0 or 1. From this, 2 4 = 16 kinds of trellis states exist. As shown in FIG. 8, there are four branches from the trellis state to the next trellis state for each trellis state. Similarly, as shown in FIG. 8, there are four branches for each trellis state.

  The assumed waveform level assumed by each branch is determined by a 3 × 3 matrix in which a feedback row is combined with a 2 × 3 matrix in which two trellis states connected to the branch are combined. Details of the feedback line will be described later.

(Example of assumed waveform level)
Hereinafter, a branch connecting [0, 0; 0, 1] to [0, 1; 1, 0] will be described as an example. The level assumed by this branch is determined by the matrix [1, 1, 0; 0, 0, 1; 0, 1, 0], assuming that the feedback row above it is [1, 1, 0]. Is done. At this time, when the impulse response h and the matrix [1, 1, 0; 0, 0, 1; 0, 1, 0] are calculated by two-dimensional convolution, the result is 0.31. This value is the assumed waveform level of this branch.

  The decision feedback Viterbi decoding process is performed in the row direction from left to right in the page data. When the waveform obtained by scanning the reproduction signal in the row direction is a reproduction signal waveform, the square error between the level of the reproduction signal waveform and the assumed waveform level of each branch is called a branch metric. That is, the branch metric is a value for each branch. In addition, the cumulative branch metric of the path leading to each trellis state is called a path metric.

  In the decision feedback Viterbi decoding method, the path having the smallest path metric among the four paths input to the trellis state is left as a surviving path. This process is performed for each trellis state and repeated in the row direction. As a result, the surviving paths converge to one when going back (leftward) by a predetermined time. Therefore, this survival path is determined as the correct answer path. The predetermined time required for the above processing is called a path memory length.

  The above process is in principle the same as the conventional one-dimensional Viterbi decoding method. Therefore, further detailed description is omitted.

(Handling of correct path)
However, the decision feedback Viterbi decoding method is different from the one-dimensional Viterbi decoding in handling the correct path. That is, in the case of one-dimensional Viterbi decoding, the correct answer path corresponds to the bit sequence itself. On the other hand, in the two-dimensional Viterbi decoding, the correct answer path is a matrix. Therefore, the correct answer path corresponds to a plurality of bit rows. In the above example, there are two rows.

Of these two bit rows, the second row does not consider the influence from the third and subsequent bit rows. That is, the influence of inter-pixel interference is only partially considered. This is not reliable. On the other hand, for the first row, all the waveform interference from the top, bottom, left and right is taken into consideration and Viterbi decoded. Therefore, only the first line is output as a decoded bit.

  One bit row is obtained by Viterbi decoding in the row direction. This process is repeated while shifting one column at a time in the column direction (downward). As a result, the entire two-dimensional reproduction signal is decoded.

(Introduction of judgment feedback)
Here, the concept of decision feedback is introduced. Judgment feedback is a method of reflecting the decoding result of the next row as a feedback row in Viterbi decoding of the next row. Specifically, it is as described above. That is, the decoding result of the next row is used for the process of determining the assumed waveform level of the branch of the next row. Thereby, the decoding result is reflected.

  The procedure for decision feedback will be described in more detail below. The uppermost row and the upper row can be regarded as “0”. Therefore, when information is reproduced based on the reproduction signal from the top row, all feedback rows are set to zero.

  Next, information is reproduced based on the reproduction signal from the second row from the top. At this time, it is assumed that the bit row is correctly decoded in the first row (that is, the uppermost row). Therefore, this is used as a feedback row for decoding the second row.

  Furthermore, reproduction is performed based on the reproduction signal from the third row from the top. At this time, it is assumed that the bit row is correctly decoded in the second row. Therefore, this is used as a feedback row for decoding the third row.

  As described above, in the decision feedback Viterbi decoding method, it is assumed that the decoding process is accurately performed in the row immediately above the current decoding target row. Therefore, determination feedback in the column direction is performed while taking into consideration the effect of the upper row. At the same time, the playback signal from the current row is Viterbi decoded. That is, the Viterbi decoding process in which the influence of inter-pixel interference in the column direction is more reflected is executed.

  By the above processing, Viterbi decoding can be executed more accurately than the conventional one-dimensional Viterbi decoding processing, that is, the processing using the change of the reproduction waveform only in the row direction as it is.

(Configuration of two-dimensional SAM detection circuit 4)
The configuration of the two-dimensional SAM detection circuit 4 will be described below with reference to FIG. As shown in this figure, the two-dimensional SAM detection circuit 4 includes a two-dimensional Viterbi decoding circuit 11, a path selection circuit 12, and a subtractor 13.

(Two-dimensional Viterbi decoding circuit 11)
The two-dimensional Viterbi decoding circuit 11 performs two-dimensional Viterbi decoding on the two-dimensional reproduction signal. Inter-pixel interference in the two-dimensional reproduction signal is assumed as a two-dimensional impulse response h represented by a matrix of 3 rows and 3 columns shown in FIG. The trellis state is defined as a 2 × 2 matrix.

  The two-dimensional Viterbi decoding circuit 11 determines the correct trellis state based on the self-decoding bit. Also, the path metrics of the four paths that are input to the determined correct trellis state are calculated. In this way, the calculated four path metrics are output to the path selection circuit 12.

(Path selection circuit 12)
The path selection circuit 12 selects the path metric of the correct path from the four input path metrics. At this time, a path metric to be selected is determined based on the decoded bit input from the two-dimensional Viterbi decoding circuit 11. Further, the path metric of the selected correct answer path is output to the subtractor 13.

  The path selection circuit 12 further selects, from among the four input path metrics, an error path having the smallest path metric among error paths having decoding target rows different from the correct path. Also at this time, the error path to be selected is determined based on the decoded bits input from the two-dimensional Viterbi decoding circuit 11. The path selection circuit 12 outputs the path metric selected at this time to the subtractor 13 as the path metric of the error path, similarly to the path metric of the correct path.

(Subtractor 13)
The subtracter 13 subtracts the path metric of the input correct path from the path metric of the input error path. Thereby, the SAM value is calculated.

(Details of trellis diagram)
In the present embodiment, the trellis diagram represents the operation of the two-dimensional Viterbi decoding circuit 11. This trellis diagram is similar to the trellis diagram shown in FIG. For example, the bit width of inter-pixel interference in the assumed response matrix h is 3. In addition, a decision feedback Viterbi decoding method is applied to this trellis diagram. Thus, the trellis state is defined as a 2 × 2 matrix.

  The two-dimensional Viterbi decoding circuit 11 can use a feedback row as a row immediately above the decoding target row. As a result, the width of the row becomes 3-1 = 2. 2 rows and 2 columns correspond to 4 bits in total. Therefore, the number of trellis states is 2 4 = 16. In addition, the number of branches input to the trellis state is four for each trellis state. Further, the number of branches output from the trellis state is four for each trellis state.

(Operation of the two-dimensional SAM detection circuit 4)
The operation of the two-dimensional SAM detection circuit 4 will be described.

The two-dimensional Viterbi decoding circuit 11 performs Viterbi decoding of the two-dimensional reproduction signal in the row direction based on the trellis diagram of FIG. Here, it is assumed that the two-dimensional reproduction signal of the input recording bit is x. An assumed waveform level corresponding to the two-dimensional reproduction signal is assumed to be w. At this time, the two-dimensional Viterbi decoding circuit 11 calculates a square error (x−w) ² between x and w as a branch metric.

  Further, the branch metrics calculated for each trellis state are accumulated for a period from when the Viterbi decoding is started to when a path reaches a certain trellis state. Thereby, a path metric which is a value obtained by accumulating branch metrics for each trellis state is calculated. The two-dimensional Viterbi decoding circuit 11 determines a surviving path based on the calculated path metric. Simultaneously with this decoding operation, the path metrics of the four paths input to the correct trellis state are calculated. Also, the calculated four path metrics are output to the path selection circuit 12.

(Specific example of path metric calculation)
The above specific example will be described using an example of transition of the trellis diagram shown in FIG. FIG. 10 is a diagram illustrating the transition of the trellis state in which the decoded bits of the m-th row and the n-th column are determined.

  In FIG. 10, it is assumed that the decoded bits output from the two-dimensional Viterbi decoding circuit 11 are in the following state.

b (m, n-2) = 0
b (m, n-1) = 0
b (m, n) = 1
b (m + 1, n-2) = 0
b (m + 1, n-1) = 0
b (m + 1, n) = 0
At this time, the correct trellis state in the m-th row and the n-th column is [0, 1; 0, 0]. Therefore, the two-dimensional Viterbi decoding circuit 11 outputs the path metrics of the four paths input to the trellis state [0, 1; 0, 0]. Specifically, path metrics P 0 (n), P 1 (n), P 2 (n), and P 3 (n) are output to the path selection circuit 12.

  The path metrics P0 (n) to P3 (n) are input to the path selection circuit 12. At the same time, from the two-dimensional Viterbi decoding circuit 11, the decoded bits b (m, n-2), b (m, n-1), b (m, n), b (m + 1, n-2), b (m + 1, n-1) and b (m + 1, n) are input. The path selection circuit 12 selects the path metric of the correct path based on each input decoded bit. Further, an error path having a minimum path metric is selected from error paths having decoding target rows different from the correct answer path.

In the example of FIG. 10, the correct path is a path corresponding to a transition from the trellis state [0, 0; 0, 0] to [0, 1; 0, 0]. At this time, the path metric is P0 (n). As an error path,
Path from [0, 0; 1, 0] to [0, 1; 0, 0] Path from [1, 0; 0, 0] to [0, 1; 0, 0] [1, 0; 1 , 0] to [0, 1; 0, 0].

  Among these, the second and third error paths have a decoding target row [1, 0, 1] different from the decoding target row [0, 0, 1] of the correct path. Therefore, the path selection circuit 12 selects the smaller one of P2 (n) and P3 (n) as the path metric of the error path. For example, assume that P2 (n) <P3 (n). In this case, the path selection circuit 12 selects P0 (n) as the path metric of the correct answer path. Further, P2 (n) is selected as the path metric for the error path. The path selection circuit 12 outputs both the selected P0 (n) and P2 (n) to the subtractor 13.

(Calculation of SAM value)
The subtracter 13 subtracts the path metric of the input correct path from the path metric of the input error path and outputs the result. For example, assume that P0 (n) and P2 (n) are input to the subtractor 13. At this time, the subtractor 13 calculates and outputs P2 (n) -P0 (n) as the SAM value.

(Two-dimensional SAM detection circuit 7)
The operation of the two-dimensional SAM detection circuit 7 is exactly the same as that of the two-dimensional SAM detection circuit 4. However, the input signals are different, and the reproduction signal of the hologram medium is input to the two-dimensional SAM detection circuit 4, but the PR ideal waveform is input to the two-dimensional SAM detection circuit 7. Accordingly, the SAM value obtained for the reproduction signal of the hologram medium is output from the two-dimensional SAM detection circuit 4, and simultaneously, the ideal SAM value corresponding to the SAM value is output from the two-dimensional SAM detection circuit 7.

  The signal evaluation apparatus 3 of the present embodiment was implemented by simulation. A two-dimensional reproduction signal of 256 × 256 test bits with a changed amount of noise is input to the mounted signal evaluation device 3, and the output signal quality evaluation value σ / μ and the bit error rate of the Viterbi decoding result are The relationship is shown in FIG. From FIG. 11, it can be confirmed that σ / μ is a signal quality evaluation value having a strong correlation with the bit error rate of the Viterbi decoding result. From this, it is possible to evaluate σ / μ of the recording medium before shipment and determine that it is acceptable when it is below a predetermined reference value, or make various adjustments so that σ / μ becomes as small as possible in the recording device or reproducing device. Optimal adjustment can be realized.

[Embodiment 2]
In the first embodiment, the average value μ and the standard deviation σ are measured as statistical values for specifying the frequency distribution of the two-dimensional SAM value, and the ratio σ / μ is used as the signal quality evaluation value. However, the correlation with the error rate is particularly strong is the standard deviation σ indicating the spread of the frequency distribution, and the average value μ is appropriately adjusted so that the amplitude of the reproduction signal matches the assumed PR waveform level. The value is almost close to a predetermined value (average value of ideal SAM values of group 1). Therefore, the function of calculating the average value μ from the average value / standard deviation calculation circuit 9 of the first embodiment is omitted, and instead of μ, a predetermined value M (predetermined constant value previously determined as the average value of the ideal SAM values of the group 1) is omitted. ) May be used as the signal quality evaluation value. This is close to the definition of jitter generally used in optical disc systems. Jitter is an evaluation value obtained by standardizing an edge fluctuation of a reproduced waveform as a standard deviation and dividing by a clock time T (predetermined value), and is theoretically strongly correlated with an error rate in the threshold detection method. Since the configuration of the second embodiment has an advantage that the circuit can be simplified, an appropriate configuration may be selected according to the system design.

[Embodiment 3]
In the first embodiment, among the plurality of groups of ideal SAM values, only the group closest to 0 (group 1 in the example of FIG. 3) is classified, and the frequency distribution of the corresponding two-dimensional SAM value is specified. The average value μ and the standard deviation σ were measured as values, and the ratio σ / μ was used as the signal quality evaluation value. In the third embodiment, a configuration will be described in which μ and σ are respectively measured for frequency distributions corresponding to a plurality of groups as well as the group having the ideal SAM value closest to 0, and the estimated bit error rate is calculated therefrom.

  FIG. 12 is a block diagram illustrating a configuration of the signal evaluation apparatus 31 according to the third embodiment. Constituent elements having functions equivalent to those of the constituent elements in the signal evaluation apparatus 3 according to the first embodiment are given the same reference numerals and description thereof is omitted.

  The signal evaluation device 31 includes an extraction unit (second extraction means) 48 in addition to the configuration of the signal evaluation device 3. A quality evaluation circuit 51 is also provided.

  The extraction unit 48 includes a value range identification circuit 14 and an average value / standard deviation calculation circuit 15. The quality evaluation circuit 51 includes an average value / standard deviation calculation circuit 9, an average value / standard deviation calculation circuit 15, and an estimated BER calculation circuit 16.

  The value range identification circuit 14 has a function similar to that of the value range identification circuit 8, but the range of the value to be identified is different and determines whether or not it is greater than or equal to V2 and less than V3. V2 and V3 are predetermined values determined for each assumed PR characteristic. In the example of FIG. 3, in order to classify the ideal SAM values of group 2 (predetermined second value range), V2 = 5000, V3 = Set to about 7500.

  The average value / standard deviation calculation circuit 15 is similar in function to the average value / standard deviation calculation circuit 9, and the frequency distribution of the SAM values corresponding to the ideal SAM values determined to belong to the group 2 by the value range identification circuit 14. An average value μ2 and a standard deviation σ2 are calculated.

  The estimated BER calculation circuit 16 calculates an estimated bit error rate (input signal quality evaluation value) from μ1, σ1, μ2, and σ2. The calculation method will be described later.

  The extraction unit 48 is a second extraction unit, the value range identification circuit 14 is a path metric difference group determination unit, the average value / standard deviation calculation circuit 15 is a statistical value calculation unit, and the estimated BER calculation circuit 16 is a quality evaluation value calculation unit. Each corresponds.

  Now, the operation of the signal evaluation apparatus 31 configured as shown in FIG. 12 will be described with reference to FIGS. 12 and 13 as follows.

  FIG. 13 is a flowchart showing the flow of processing in the signal evaluation apparatus 31 according to the third embodiment. Detailed description of the same processing as that of the first embodiment shown in the flowchart of FIG. 2 is omitted.

  In step S21, a hologram medium on which a known test bit is recorded is reproduced and a two-dimensional reproduction signal is output. The processing here is the same as step S1 of FIG. 2 described in the first embodiment.

  Next, in step S22, the two-dimensional SAM detection circuit 4 performs Viterbi decoding on the two-dimensional reproduction signal, and outputs a SAM value for each bit. The operation of the two-dimensional SAM detection circuit 4 in step S22 is the same as that in step S2 in FIG.

  Next, in step S23, an ideal SAM value corresponding to the SAM value output in step S2 is output. This operation is the same as step S3 in FIG.

  Next, in step S24, the value range identification circuit 8 determines whether or not the ideal SAM value is not less than V1 and less than V2. Thereby, only 88 types of ideal SAM values of group 1 in FIG. 3 can be extracted.

  If it is determined in step S24 that the ideal SAM value is greater than or equal to V1 and less than V2, the process proceeds to step S25, and the corresponding SAM value is set as a calculation target in the average value / standard deviation calculation circuit 9. When the process of step S25 is completed, the process proceeds to step S28. On the other hand, if it is determined in step S24 that the ideal SAM value is not in the range of V1 or more and less than V2, the process proceeds to step S26.

  In step S26, the value range identification circuit 14 determines whether or not the ideal SAM value is not less than V2 and less than V3. Thereby, only 351 types of ideal SAM values of group 2 in FIG. 3 can be extracted.

  If it is determined in step S26 that the ideal SAM value is greater than or equal to V2 and less than V3, the process proceeds to step S27, and the corresponding SAM value is set as a calculation target in the average value / standard deviation calculation circuit 15. On the other hand, if it is determined in step S26 that the ideal SAM value is not in the range of V2 or more and less than V3, the process proceeds to step S28.

  Next, in step S28, it is determined whether or not the output of the SAM value for all test bits has been completed. If not completed, the process returns to step S22. Therefore, the processes from step S22 to S28 are repeated until the process is completed for all the test bits.

  When the end determination is made in step S28, the process proceeds to step S29. In step S29, the average value / standard deviation calculation circuit 9 calculates μ1 and σ1, the average value / standard deviation calculation circuit 15 calculates μ2 and σ2, and the estimated BER calculation circuit 16 determines the estimated bit error rate. .

  Through the above processing, the estimated bit error rate can be finally obtained as the signal quality evaluation value of the reproduction signal of the hologram memory.

  Here, the reason why the estimated bit error rate in Viterbi decoding can be obtained by the signal evaluation apparatus 31 according to the present embodiment will be described in detail.

  FIG. 6A corresponds to the frequency distribution of the SAM values corresponding to the ideal SAM values of the group 1 extracted in step S24, and FIG. 6B corresponds to the ideal SAM values of the group 2 extracted in step S26. It is a figure which shows an example of the frequency distribution of a SAM value. The three frequency distributions show cases where the quality of the reproduction signal is different and the BER is 1E-5 or less, 1E-4, 5E-3, respectively. As can be seen from FIGS. 6A and 6B, the frequency distributions of both groups 1 and 2 have a distribution shape close to a normal distribution.

  Based on the same theory as in the first embodiment, the cumulative relative frequency where the SAM value <0 in the frequency distribution of group 1 can be estimated by F1 (0) = Φ (−μ1 / σ1). Further, in the frequency distribution of group 2, the cumulative relative frequency where the SAM value <0 can be estimated by F2 (0) = Φ (−μ2 / σ2).

  Because of the property that SAM value <0 and Viterbi decoding error bit generation correspond, the error rate in the pattern corresponding to the ideal SAM value belonging to group 1 is F1 (0), and the pattern corresponding to the ideal SAM value belonging to group 2 The error rate at 1 is almost equal to F2 (0). Therefore, the error rate for all patterns, in other words, for all bits can be estimated by adding F1 (0) and F2 (0).

  Strictly speaking, it is necessary to add up the error rates in the patterns corresponding to the ideal SAM values belonging to all groups such as group 3 and group 4 in FIG. Since the thing is dominant, it can be considered that it can be estimated sufficiently accurately by adding only the group 1 and the group 2.

  Strictly speaking, when one SAM value is negative, a Viterbi decoding error does not always occur by 1 bit, and depending on the pattern, a plurality of decoding error bits may occur. If the expected value of the number of decoding error bits corresponding to the ideal SAM value belonging to is separately measured and reflected by multiplying it as a coefficient when F1 (0) and F2 (0) are added, the estimation can be made more accurately. It becomes possible. Furthermore, at the time of addition, it is necessary to multiply the appearance probability of the patterns belonging to group 1 and group 2 as a coefficient. The appearance probability may be implemented as a constant by checking the test bits to be used in advance, or the appearance frequency may be counted by the value range identification circuits 8 and 14 during reproduction. For simplicity, assuming that the expected value of the number of decoding error bits in group 1 and group 2 is 1, the appearance probability of the pattern of group 1 is P1, and the appearance probability of the pattern of group 2 is P2, the estimated bit error rate Is calculated by equation (4).

Estimated bit error rate = P1 × F1 (0) + P2 × F2 (0) (4)
The signal evaluation apparatus 31 of the third embodiment is implemented by simulation. A 256 × 256 test bit two-dimensional reproduction signal with a varying amount of noise is input to the implemented signal evaluation device 31, and the relationship between the output estimated bit error rate and the measured bit error rate of the Viterbi decoding result is expressed as follows: As shown in FIG. From FIG. 14, it can be confirmed that the estimated bit error rate and the actually measured bit error rate agree very well.

  When the amount of noise is small, the difference between the estimated bit error rate and the actually measured bit error rate is large. This is because the probability that the estimated bit error rate is 0 or less due to the natural spread of the normal distribution is obtained, that is, only errors due to natural noise are estimated, whereas the measured bit error rate is This is because errors due to unnatural noise such as defects are also included.

  Usually, errors due to defects are at most a few bits, so in the situation where the error rate is bad, the number of errors due to natural noise is much larger, so the influence can be ignored, but in situations where the error rate is good, it is caused by natural noise. Since there are almost no errors, the effect of errors due to defects is greatly increased.

  However, it is originally desirable to evaluate the reproduction signal quality without the influence of defects. Also, in this simulation example, the total number of test bits is only 65536 bits, and an error rate of 1E-5 or less cannot be measured (the number of errors is 0). Although it cannot be performed, a very low error rate can be calculated with the estimated bit error rate. Therefore, if various adjustments of the recording device and the reproduction device are performed based on the estimated bit error rate, it is possible to realize a truly optimum adjustment.

[Embodiment 4]
In the first embodiment, predetermined test bits are recorded in advance on the hologram medium.

  The hologram memory reproducing device 1 reproduces this test bit. That is, it is necessary to record test bits on the hologram medium in advance. Therefore, the following two problems arise. First, the use efficiency of the hologram medium is lowered. Second, the quality of the signal itself generated by recording or reproducing actual information cannot be directly evaluated.

  The signal evaluation apparatus 32 of the present embodiment avoids the above two problems by feeding back and using its own Viterbi decoding result.

  FIG. 15 is a block diagram illustrating a configuration of the signal evaluation device 32. Differences between the signal evaluation device 32 and the signal evaluation device 3 will be described below. That is, the signal evaluation device 32 does not have the test bit memory 5, and instead, the two-dimensional SAM detection circuit 4 outputs the decoded bit. The decoded bits are input to an ideal signal generation unit (ideal signal generation means) 41. The ideal signal generation unit 41 includes a PR filter 6, and the decoded bit passes through the PR filter 6. This differs from the signal evaluation device 3 in that the generated PR ideal waveform is input to the two-dimensional SAM detection circuit 7. Since the operation other than the PR ideal waveform generation operation is exactly the same as the operation of the signal evaluation apparatus 3 described in the first embodiment, detailed description thereof will be omitted.

In this case, if the two-dimensional SAM detection circuit 4 causes a Viterbi decoding error, an incorrect PR ideal waveform is generated, and as a result, the group of ideal SAM values is misclassified. However, in general, the signal quality to be evaluated is at most an error rate of 10 ^-2 units at best, and even if the group classification of ideal SAM values is incorrect corresponding to the error rate, the average value or standard Since the influence on the calculation result of the deviation is very small, it can be said that the error given to the calculation result of the final signal quality evaluation value σ / μ is small enough to be ignored.

On the other hand, with such a configuration, the test bit memory 5 can be omitted, and the test bits can be recorded on the hologram medium in advance or recorded on the hologram medium before signal quality evaluation. The effect of eliminating the troublesome work is obtained. Furthermore, signal quality evaluation can be performed not only on test bits but also on arbitrary bits.

  In the first to fourth embodiments, the example in which the average value and the standard deviation are used as the statistical values for specifying the frequency distribution has been described. However, the present invention is not limited to this. For example, the mode value may be used instead of the average value. Any statistical value that can identify the frequency distribution, such as using variance instead of standard deviation, may be used. The gist of the present invention is to classify the ideal SAM values for each group, classify the frequency distribution of the corresponding SAM values, and evaluate the signal quality based on the statistical values that identify the respective frequency distributions.

  Each block of the signal evaluation apparatus described in the first to fourth embodiments may be configured by hardware logic, or may be realized by software using a computer as follows.

  That is, the signal evaluation device 3, the signal evaluation device 31, and the signal evaluation device 32 are a CPU (central processing unit) that executes instructions of a control program that realizes each function, a ROM (read only memory) that stores the program, A RAM (random access memory) for expanding the program, a storage device (recording medium) such as a memory for storing the program and various data, and the like are provided. The object of the present invention is the program code (execution format program, intermediate code program, source program) of the control program of the signal evaluation device 3, the signal evaluation device 31, and the signal evaluation device 32, which is software that implements the functions described above. Is supplied to the signal evaluation device 3, the signal evaluation device 31, and the signal evaluation device 32, and the computer (or CPU or MPU) is recorded on the recording medium. This can also be achieved by reading out and executing.

  Examples of the recording medium include a tape system such as a magnetic tape and a cassette tape, a magnetic disk such as a floppy (registered trademark) disk / hard disk, and an optical disk such as a CD-ROM / MO / MD / DVD / CD-R. Card system such as IC card, IC card (including memory card) / optical card, or semiconductor memory system such as mask ROM / EPROM / EEPROM / flash ROM.

  Further, the signal evaluation device 3, the signal evaluation device 31, and the signal evaluation device 32 may be configured to be connectable to a communication network, and the program code may be supplied via the communication network. The communication network is not particularly limited. For example, the Internet, intranet, extranet, LAN, ISDN, VAN, CATV communication network, virtual private network, telephone line network, mobile communication network, satellite communication. A net or the like is available. Further, the transmission medium constituting the communication network is not particularly limited. For example, even in the case of wired such as IEEE 1394, USB, power line carrier, cable TV line, telephone line, ADSL line, etc., infrared rays such as IrDA and remote control, Bluetooth ( (Registered trademark), 802.11 wireless, HDR, mobile phone network, satellite line, terrestrial digital network, and the like can also be used. The present invention can also be realized in the form of a computer data signal embedded in a carrier wave in which the program code is embodied by electronic transmission.

  Thus, in this specification, the means does not necessarily mean physical means, but includes cases where the functions of the means are realized by software. Further, the function of one means may be realized by two or more physical means, or the functions of two or more means may be realized by one physical means.

  In any of the above embodiments, the hologram memory system has been described as an example of the reproduction system. However, the present invention is not limited to this, and the same effect can be achieved in a system that reproduces a two-dimensional signal. That is, the present invention can be applied to a two-dimensional barcode reproduction system represented by a QR code as another reproduction system.

  Note that the present applicant has previously devised a new SAM value suitable for a system that performs Viterbi decoding of a two-dimensional reproduction signal (Japanese Patent Application No. 2006-75680). In the process of Viterbi decoding a two-dimensional signal, as shown in FIG. 16, one correct path and a plurality of error paths are input to the correct state of the trellis diagram. The trellis diagram is a state transition diagram corresponding to the Viterbi decoding process. In the case of a two-dimensional signal, there are a plurality of error paths corresponding to one correct path, but it has the smallest path metric among error paths having a decoding target line different from the decoding target line of the correct path. The major feature is that an error path is selected and the path metric difference between this and the correct path is obtained as a SAM value. (In the case of one-dimensional, one error path corresponds to one correct path, so it is simple. The difference between the two path metrics is defined as the SAM value). The path metric is an error of the input signal with respect to the ideal signal of the path.

In the case of the example in FIG. 17, the correct path input to the correct state [0, 1; 0, 0] is a path corresponding to the transition of [0, 0; 0, 0] → [0, 1; 0, 0]. And its path metric is P0 (n) (the trellis state is represented as [k, l; m, n]. [K, l; m, n] is the row vector [k, l] and [m , N] are arranged in the column direction (up and down). The corresponding error path is as follows:
[0, 0; 1, 0] → [0, 1; 0, 0] path [1, 0; 0, 0] → [0, 1; 0, 0] path [1, 0; 1, 0 ] → [0, 1; 0, 0] There are three paths.

  Among these, the second and third error paths have decoding target rows [1, 0, 1] different from the decoding target rows [0, 0, 1] of the correct answer path, so P2 (n) and The smaller one of P3 (n), that is, P2 (n) is selected as the path metric of the error path. Eventually, 0.34−0.27 = 0.07, which is the difference between P2 (n) and P0 (n), is calculated as the SAM value.

  In order not to generate an error in the Viterbi decoding process, the path metric of the correct path needs to be smaller than the path metric of the wrong path. That is, it is necessary that the SAM value> 0. Also, the larger the SAM value, the less likely an error will occur. In this way, in a system that performs Viterbi decoding of a two-dimensional signal, the quality of a reproduction signal can be evaluated with high reliability based on the SAM value.

  However, in the case of SAM values in two-dimensional Viterbi decoding, where the trellis diagram is more complex than one dimension, there are more than a thousand specific patterns that are particularly prone to errors, and the corresponding ideal SAM values are many from tens of There are hundreds of types. Therefore, if it is attempted to realize a configuration for detecting these 1,000 or more types of specific patterns with the apparatus, the circuit scale becomes extremely large, which is practically impossible.

  As described above, with the configuration of the present embodiment, in the system that performs Viterbi decoding of a two-dimensional signal, the quality of the reproduced signal is evaluated with high reliability based on the SAM value without increasing the circuit scale. be able to.

  The present invention is not limited to the above-described embodiments, and various modifications are possible within the scope shown in the claims, and embodiments obtained by appropriately combining technical means disclosed in different embodiments. Of course, it is included in the technical scope of the present invention.

  According to the signal evaluation apparatus and the signal evaluation method of the present invention, a recording medium, a recording apparatus, and a reproduction apparatus are evaluated using an index value having a very high correlation with an error rate in a Viterbi decoding apparatus for a two-dimensional signal. Or a highly reliable pass / fail judgment and various adjustments of a pre-shipment product based on the evaluation result, a hologram memory reproducing device for reproducing a two-dimensional signal, and a two-dimensional bar represented by a QR code. It can be applied to a code playback device.

It is a block diagram which shows the structure of the hologram memory reproduction signal evaluation system of Embodiment 1. 2 is a flowchart showing a flow of processing for calculating σ / μ as a quality evaluation value of a two-dimensional reproduction signal in the hologram memory reproduction signal evaluation system of FIG. 1. It is a figure which shows an example of the frequency distribution of an ideal SAM value. It is a figure which shows an example of a flag map. It is a block diagram which shows the other structure of the hologram memory reproduction signal evaluation system of Embodiment 1. It is a figure which shows an example of the frequency distribution of the SAM value corresponding to the ideal SAM value of each group. It is a figure which shows the assumed impulse response in the decision feedback Viterbi decoding method. It is a trellis diagram expressing the decision feedback Viterbi decoding method. It is a block diagram which shows the structure of a two-dimensional SAM detection circuit. It is a figure which shows the transition of the trellis state about four branches input into the trellis state [0, 1; 0, 0]. It is a figure which shows the simulation result of the relationship between (sigma) / micro and a bit error rate. It is a block diagram which shows the structure of the signal evaluation apparatus which concerns on Embodiment 3. 13 is a flowchart showing a flow of processing for calculating an estimated bit error rate as a quality evaluation value of a two-dimensional reproduction signal in the signal evaluation apparatus of FIG. It is a figure which shows the simulation result of the relationship between an estimated bit error rate and a measurement bit error rate. It is a block diagram which shows the structure of the signal evaluation apparatus which concerns on Embodiment 4. It is a figure which shows the trellis diagram of two-dimensional Viterbi decoding. It is a block diagram which shows the structure of the conventional hologram memory recording / reproducing apparatus.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Hologram memory reproducing apparatus 2 Two-dimensional equalizer 3, 30, 31, 32 Signal evaluation apparatus 4 Two-dimensional SAM detection circuit (Viterbi decoding means)
5 Test bit memory 6 PR filter 7 Two-dimensional SAM detection circuit (ideal path metric difference generation means)
8, 14 Value range identification circuit (path metric difference group determination means)
9, 15 Average value / standard deviation calculation circuit (quality evaluation means, statistical value calculation means)
10 Divider (quality evaluation means, quality evaluation value calculation means)
11 Two-dimensional Viterbi decoding circuit 12 Path selection circuit 13 Subtractor 16 Estimated BER calculation circuit (quality evaluation means, quality evaluation value calculation means)
40, 41 Ideal signal generator (ideal signal generator)
45 Extraction unit (extraction means, first extraction means)
46 Extraction unit (extraction means)
47 Flag map memory (storage unit)
48 Extraction unit (second extraction means)
50, 51 Quality evaluation circuit (quality evaluation means)

Claims (11)

In a signal evaluation apparatus that outputs a quality evaluation value of an input signal that is a two-dimensional reproduction signal ,
Assuming a predetermined partial response characteristic for the input signal, the first bit of the correct path consisting of a matrix obtained by performing two-dimensional Viterbi decoding for performing Viterbi decoding in the row direction on the input signal is decoded bit. Viterbi decoding means for outputting the path metric difference between the correct path and the error path obtained as
Extraction means for extracting a path metric difference in which an ideal path metric difference corresponding to the input signal from which a path metric difference is obtained belongs to a predetermined value range from a plurality of path metric differences output from the Viterbi decoding means;
A signal evaluation apparatus comprising: quality evaluation means for calculating a quality evaluation value of the input signal based on a group of path metric differences extracted by the extraction means.   The quality evaluation means calculates a statistical value representing a degree of spread of a frequency distribution of the group of the extracted path metric differences, and calculates a quality evaluation value of the input signal from the statistical value. Item 2. The signal evaluation device according to Item 1. The quality evaluation means calculates a standard deviation of the extracted path metric difference group as the statistical value, and calculates the standard deviation and an average value of the extracted path metric difference group as the quality evaluation value. The signal evaluation apparatus according to claim 2 , wherein a ratio of the two is calculated. The quality evaluation means calculates a standard deviation of the group of extracted path metric differences as the statistical value, and calculates a ratio between the standard deviation and a predetermined constant value as the quality evaluation value. The signal evaluation apparatus according to claim 2 . In a signal evaluation apparatus that outputs a quality evaluation value of an input signal that is a two-dimensional reproduction signal ,
Assuming a predetermined partial response characteristic for the input signal, the first bit of the correct path consisting of a matrix obtained by performing two-dimensional Viterbi decoding for performing Viterbi decoding in the row direction on the input signal is decoded bit. Viterbi decoding means for outputting the path metric difference between the exact path and the error path obtained as
A first extraction for extracting a path metric difference in which an ideal path metric difference corresponding to the input signal from which a path metric difference is obtained belongs to a predetermined first value range from a plurality of path metric differences output from the Viterbi decoding means. Means,
An ideal path metric difference corresponding to the input signal obtained from the plurality of path metric differences output from the Viterbi decoding means belongs to a predetermined second value range different from the first value range. A second extraction means for extracting a path metric difference;
A signal evaluation apparatus comprising: quality evaluation means for calculating a quality evaluation value of the input signal based on a group of path metric differences respectively extracted by the first and second extraction means. Ideal signal generating means for converting the test bit corresponding to the input signal into a partial response characteristic ideal waveform signal according to the partial response characteristic assumed in the Viterbi decoding means;
An ideal for generating the ideal path metric difference, which is a path metric difference between a correct path and an error path, by performing Viterbi decoding on the partial response characteristic ideal waveform signal converted by the ideal signal generation means according to the partial response characteristic. signal evaluation apparatus according to any one of claims 1 to 5, characterized in that it comprises a path metric difference generating means. The Viterbi decoding means outputs a decoded bit corresponding to the path metric difference together with the path metric difference; and
Ideal signal generating means for converting the decoded bit corresponding to the input signal into a partial response characteristic ideal waveform signal according to the partial response characteristic assumed in the Viterbi decoding means;
An ideal for generating the ideal path metric difference, which is a path metric difference between a correct path and an error path, by performing Viterbi decoding on the partial response characteristic ideal waveform signal converted by the ideal signal generation means according to the partial response characteristic. signal evaluation apparatus according to any one of claims 1 to 5, characterized in that it comprises a path metric difference generating means. The extraction means includes
In association with the path metric difference output from the Viterbi decoding means, a flag map indicating whether or not the ideal path metric difference corresponding to the input signal from which the path metric difference is obtained belongs to a predetermined value range is stored in advance. Storage section
2. The signal evaluation apparatus according to claim 1, wherein whether to extract a path metric difference output from the Viterbi decoding unit is determined based on the flag map. In a signal evaluation method for outputting a quality evaluation value of an input signal which is a two-dimensional reproduction signal ,
Assuming that the input signal has a predetermined partial response characteristic, the first row in the correct path consisting of a matrix obtained by performing two-dimensional Viterbi decoding for performing Viterbi decoding in the row direction on the input signal is decoded bit. Viterbi decoding step for outputting the path metric difference between the correct path and the error path obtained as
An extraction step for extracting a path metric difference in which an ideal path metric difference corresponding to the input signal that has obtained the path metric difference belongs to a predetermined value range from the path metric difference output in the Viterbi decoding step;
And a quality evaluation step of calculating a quality evaluation value of the input signal based on the group of path metric differences extracted in the extraction step.   A signal evaluation program for operating the signal evaluation apparatus according to claim 1, wherein the computer functions as each of the above means.   The computer-readable recording medium which has recorded the signal evaluation program of Claim 10.

Images