JP4519776B2 - Viterbi decoding apparatus, playback system, Viterbi decoding method, Viterbi decoding program, and recording medium - Google Patents

Description

  The present invention relates to Viterbi decoding that can be suitably applied to an optical memory system that two-dimensionally records and reproduces digital information such as a hologram memory.

  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. 9, in this system, when information data is recorded, a recording signal is recorded based on information data to be recorded using a spatial light modulator 20 (for example, a liquid crystal panel) composed of a plurality of pixels. The light is spatially modulated and interference fringes are generated by causing the recording signal light passing through the lens 21 and the coherent recording reference light to interfere with each other, and an image formed by the interference fringes is generated on the hologram medium 22 as two-dimensional information. Record.

  The hologram medium 22 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 information data is reproduced by reading the recorded interference fringes from the hologram medium 22, it is generated by irradiating the hologram medium 22 with the reproduction reference light at the same incident angle as the recording reference light. The reflected light or transmitted light is received by the light receiving element 24 (for example, CCD) having a plurality of pixels through the lens 23 to generate a reproduction signal, and the original information data is reproduced using the generated reproduction signal.

  Note that the two-dimensional reproduction signal obtained as described above can be regarded as indicating a two-dimensional waveform of the signal, and is arranged along a certain direction in the two-dimensional reproduction signal. It can also be understood that the reproduced signal indicates a one-dimensional waveform of the signal. As described above, the reproduction signal is regarded as a waveform and may be referred to as a “reproduction waveform signal”.

  As described above, since reproduction is performed in units of two-dimensional page data in the above system, it is possible to significantly improve the reproduction speed as compared with a conventional optical disk that performs reproduction in one dimension.

  Here, 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 the inhomogeneity of the recording medium). The S / N ratio decreases due to the influence. In addition, there is an influence of a reproduction signal from an adjacent pixel, that is, an inter-pixel interference, and the original signal may not be reproduced accurately.

  Therefore, as a technique for accurately reproducing information data while reducing the influence of noise and inter-pixel interference, a technique for performing Viterbi decoding processing on a reproduction signal output from a light receiving element has been studied. As a result, a Decision Feedback Viterbi Algorithm has been developed (see Patent Document 1).

  The determination feedback Viterbi decoding method performs the process of determining the decoded bits row by row by performing Viterbi decoding along the horizontal direction (row direction) of the page data while shifting the row downward (column direction). At this time, it is characterized in that decision feedback is performed using the previous decoding result (feedback row) for Viterbi decoding of the next row.

  The outline of the conventional decision feedback Viterbi decoding method will be described with reference to FIGS. Assuming that inter-pixel interference is a two-dimensional impulse response h ′ represented by a 3 × 3 matrix as shown in FIG. 10, the trellis state is defined as a 2 × 2 matrix, and the trellis diagram is as shown in FIG. It is expressed in Note that “two-dimensional impulse response” means inputting a unit impulse signal (a two-dimensional bit matrix in which the signal level of only one bit is 1 and the signal levels of all other bits are all 0) to a linear system. Means the output signal. In the two-dimensional impulse response h ′ in FIG. 10, it can be seen that the signal level is generated not only in the central pixel but also in the neighboring adjacent pixels, and the occurrence of inter-pixel interference is assumed.

  The trellis state [k, l; m, n] ([k, l; m, n]) means a matrix in which row vectors [k, l] and [m, n] are arranged in the column direction. In this case, k, l, m, and n are all 0 or 1, and there are 2 4 = 16 types of trellis states.

  There are four branches from each trellis state to the next trellis state, and there are also four branches input to each trellis state. An assumed value (also referred to as an “assumed waveform level”) assumed by each branch is a 2-by-3 matrix obtained by combining two trellis states connected to the branch, and a feedback row above that (the details of the feedback row are as follows). Determined by a matrix of 3 rows and 3 columns, which will be described later.

  For example, the level assumed by the branch of [0, 0; 0, 1] → [0, 1; 1, 0] is assumed to be a matrix [ 1, 1, 0; 0, 0, 1; 0, 1, 0]. Specifically, 0.31 obtained by a two-dimensional convolution operation between the impulse response h ′ and the matrix [1, 1, 0; 0, 0, 1; 0, 1, 0] is the assumed value of this branch.

  The Viterbi decoding process is performed in the row direction from left to right. The square error between the reproduced waveform signal obtained by scanning the reproduced signal in the row direction and the assumed waveform level of each branch is defined as the branch metric, and the cumulative branch metric of the path leading to each trellis state is defined as the path metric. Of the four paths to be input, the path with the smallest path metric is left as a surviving path.

  If this is repeated in the row direction, the surviving path converges to one if it is traced back by a predetermined time (leftward), so this is determined as the correct path (this “predetermined time” is the path memory length) Called). Since this process is in principle the same as the conventional one-dimensional Viterbi decoding, detailed description is omitted. However, the difference from the one-dimensional Viterbi decoding is that the one-dimensional correct path corresponds to the bit sequence itself. In the dimension, since the correct path is a matrix, it corresponds to a plurality of bit rows (2 rows in the above example).

  Of these two bit lines, the second line does not consider the influence from the third and subsequent bit lines, that is, only partially considers the influence of inter-pixel interference, and is not reliable. Only the first line that is Viterbi-decoded with all the waveform interference from the top, bottom, left and right taken into account is output as decoded bits.

  Since one bit line is obtained by Viterbi decoding in the row direction as described above, this process is repeated while shifting in the column direction (downward) one row at a time in order to decode the entire page data.

  Here, the concept of decision feedback is introduced. Judgment feedback is a technique of reflecting the decoding result of the next row as a feedback row in the Viterbi decoding of the next row. Specifically, as described above, the expected waveform level of the branch of the next row is set as follows. It is reflected by using it for the processing to decide.

  The procedure for decision feedback will be explained in a little more detail. When the reproduction is performed based on the reproduction waveform signal received in the uppermost row of the light receiving element 24, the amount of light received in the upper row of the uppermost row can be regarded as 0, so that the reproduction waveform signal from the uppermost row is obtained. When reproducing based on the above, all feedback lines are set to zero.

  Next, when reproducing based on the reproduction waveform signal from the second line from the top, it is assumed that the bit line is correctly decoded in the first line (the uppermost line), and this is used as the feedback line. Used for decoding.

  Furthermore, when reproducing based on the reproduction waveform signal from the third row from the top, it is assumed that the bit row is correctly decoded in the second row, and this is used as a feedback row to decode the third row. Use.

As described above, it is assumed that the decoding process has been accurately performed on the line immediately above the current decoding target line, taking into consideration the influence of the upper line (that is, performing decision feedback in the column direction) Since Viterbi decoding is performed on the playback waveform signal from the row, Viterbi decoding processing that takes into account the influence of inter-pixel interference in the column direction has been performed, and the change in the playback waveform only in the row direction was used. Compared with conventional one-dimensional Viterbi decoding processing, more accurate decoding processing can be performed.
US Pat. No. 5,740,184 (issued in 1998) Toyohiko Yadagai, “Physics of Modern People 5: Light and Fourier Transform”, Asakura Shoten, May 15, 1992, p.12-15

  In the above background art, inter-pixel interference is expressed as a two-dimensional impulse response h ′, and an assumed waveform level assumed by each branch is a bit matrix determined by two trellis states and a feedback row connected to the branch, h ′ Is determined by the two-dimensional convolution operation.

  Here, the signal received and output by each pixel of the CCD corresponds to the intensity (power) of light. On the other hand, the optical meaning of the two-dimensional convolution operation is to superimpose light waves from a plurality of light sources, but in the hologram memory system, the light waves from each pixel of the spatial light modulator 20 correspond to a plurality of light sources. Therefore, these light sources are coherent. Therefore, from the viewpoint of optical theory, it is appropriate to perform the superposition of light waves, that is, the two-dimensional convolution operation, based on the amplitude of light (see Non-Patent Document 1).

  However, since the above background art has not taken this into consideration, there is a problem in that the accuracy of setting the assumed waveform level of each branch is not sufficient and high-performance Viterbi decoding cannot be realized.

  The present invention has been made in view of the above problems, and an object thereof is to reduce an error rate of a decoding result in Viterbi decoding of a two-dimensional reproduction signal of page data.

  A Viterbi decoding device according to the present invention is a Viterbi decoding device that Viterbi-decodes a two-dimensional reproduction signal obtained by optically reproducing page data and indicating a signal value corresponding to the intensity of light. In order to solve the problem, a transition of a trellis state corresponding to the two-dimensional reproduction signal and a branch metric calculation means for obtaining a branch metric of the transition based on the two-dimensional reproduction signal, and Viterbi decoding based on the branch metric And a Viterbi decoding means for performing the above operation, and assuming that the assumed value of the transition of the trellis state assumed based on the amplitude of light is an assumed amplitude value, the branch metric calculating means responds to the square value of the assumed amplitude value. The branch metric is obtained using the assumed signal value.

  A Viterbi decoding method according to the present invention is a Viterbi decoding method for Viterbi decoding a two-dimensional reproduction signal obtained by optically reproducing page data and indicating a signal value corresponding to the intensity of light. In order to solve the problem, a branch metric calculation process for obtaining a branch metric of the transition based on the transition of the trellis state corresponding to the two-dimensional reproduction signal and the two-dimensional reproduction signal, and on the basis of the branch metric A Viterbi decoding process for performing Viterbi decoding, and assuming that the assumed value of the transition of the trellis state assumed based on the amplitude of light is an assumed amplitude value, the branch metric calculation process uses the square value of the assumed amplitude value. It is characterized in that a branch metric is obtained by using an assumed signal value corresponding to.

  Here, “optical reproduction” typically means that reflected light or transmitted light generated by irradiating light onto a medium on which page data is recorded is received by a CCD or CMOS sensor or the like. This corresponds to reproduction by receiving light by the element, but is not limited thereto, and any reproduction is possible as long as a two-dimensional reproduction signal indicating a value corresponding to the intensity of light can be obtained.

  In addition, “value corresponding to the light intensity” means a value that has a linear relationship with the light intensity (however, if the linearity is lost due to non-essential factors such as noise) Is not excluded). Therefore, the “value according to the light intensity” includes not only the light intensity value itself but also a value obtained by amplifying or shifting the light intensity value, but it is nonlinear such as the square root of the light intensity value. The value converted to is not included.

  In addition, the “assumed signal value according to the square value” means a value that is linearly related to the square value (however, the linearity is lost due to non-essential factors such as noise. Is not excluded). Therefore, “the assumed signal value corresponding to the square value” includes not only the square value itself but also a value obtained by amplifying or shifting the square value, but nonlinearly such as the square root of the square value. Does not include converted values.

  In the above configuration or method, when Viterbi decoding a two-dimensional reproduction signal obtained by optically reproducing page data and indicating a signal value corresponding to the intensity of light, an assumed amplitude value (that is, based on the amplitude of light). The branch metric is calculated using the assumed signal value corresponding to the square value of the assumed amplitude value instead of using the assumed trellis state transition value) as it is. This makes it possible to set an assumed signal value that is more suitable for Viterbi decoding than in the case where the assumed amplitude value is directly used as the assumed signal value. This is because the two-dimensional reproduction signal is a signal indicating a signal value corresponding to the intensity of light, and there is a relationship between the amplitude of the light and the intensity that the square of the absolute value of the amplitude is the intensity. .

  As described above, the assumed signal value more suitable for Viterbi decoding can be set. As a result, high-performance Viterbi decoding can be realized, and reproduction with a low error rate is possible.

  In the Viterbi decoding device according to the present invention, in the Viterbi decoding device, the Viterbi decoding unit performs Viterbi decoding on a plurality of rows including a decoding target row in the page data, and the Viterbi decoding unit performs the Viterbi decoding unit. Of the paths surviving as a result of decoding, a decoding bit output means for outputting a decoding bit line corresponding to the decoding target line as a decoding result and a decoding bit line output by the decoding bit output means for the immediately preceding decoding target line are stored. Two-dimensional convolution of a decoded bit storage means, a bit matrix determined from a trellis state transition and a decoded bit row output from the decoded bit storage means, and a two-dimensional impulse response matrix assumed as light amplitude in the reproduction An assumed signal value calculating means for calculating the assumed amplitude value by calculation to obtain the assumed signal value; It can be configured to include a.

  Here, “two-dimensional impulse response” means a unit impulse signal (a two-dimensional bit matrix in which the signal level of only one bit is 1 and the signal levels of other bits are all 0) in a linear system. This means an output signal when input, and “two-dimensional impulse response matrix” means a matrix composed of the output signals.

  In the above configuration, the branch metric of the transition is obtained based on the trellis state transition assumed for the two-dimensional reproduction signal and the two-dimensional reproduction signal, and the two-dimensional impulse response assumed as the amplitude of light in the reproduction. Two-dimensional convolution of the matrix and the bit matrix corresponding to the transition is performed to obtain an assumed amplitude value of the transition.

  As a result, the calculation based on the viewpoint of the optical theory described above, that is, the two-dimensional convolution calculation corresponding to the amplitude of the light is performed, so that an appropriate calculation in optical theory can be performed.

  In the Viterbi decoding device according to the present invention, in the Viterbi decoding device, a response for generating the two-dimensional impulse response matrix by obtaining a square root of a two-dimensional reproduction signal obtained by reproducing a bit matrix corresponding to a unit impulse signal Matrix operation means may be further provided.

  Here, “unit impulse signal” means a two-dimensional bit matrix in which the signal level of only 1 bit is 1 and the signal levels of other bits are all 0 as described above.

  With the above configuration, a two-dimensional impulse response matrix suitable for each recording medium and reproduction environment for recording page data can be generated, and a branch metric can be obtained using this. As a result, higher-performance Viterbi decoding can be realized, and reproduction with a lower error rate is possible.

  The reproduction system according to the present invention can be configured by including any one of the above Viterbi decoding apparatuses and a reproduction apparatus that obtains the two-dimensional reproduction signal by optically reproducing page data.

  The Viterbi decoding program according to the present invention is a program for causing a computer to function as each means in order to operate the computer as the Viterbi decoding device. A recording medium according to the present invention is a computer-readable recording medium in which the Viterbi decoding program is described.

  By using the program or the recording medium, the Viterbi decoding apparatus can be realized by causing a computer to function as each unit of the Viterbi decoding apparatus.

  As described above, the Viterbi decoding apparatus according to the present invention, assuming that the assumed value of the trellis state transition assumed based on the amplitude of the light is the assumed amplitude value, the branch metric calculation means calculates the square value of the assumed amplitude value. The branch metric is obtained using the assumed signal value according to the above.

  Further, as described above, the Viterbi decoding method according to the present invention is a method for obtaining a branch metric using an assumed signal value corresponding to a square value of an assumed amplitude value in the branch metric calculation process.

  This makes it possible to set an assumed signal value that is more suitable for Viterbi decoding than in the case where the assumed amplitude value is directly used as the assumed signal value. This is because the two-dimensional reproduction signal is a signal indicating a signal value corresponding to the intensity of light, and there is a relationship between the amplitude of the light and the intensity that the square of the absolute value of the amplitude is the intensity. .

  As described above, it is possible to set an assumed signal value more suitable for Viterbi decoding. As a result, high-performance Viterbi decoding can be realized, and reproduction with a low error rate can be achieved.

[Embodiment 1]
A first embodiment of the present invention will be described with reference to FIGS.

  FIG. 1 is a block diagram showing a configuration of a hologram memory reproduction system (reproduction system) 10 according to this embodiment to which the Viterbi decoding apparatus and the Viterbi decoding method of the present invention are applied. A hologram memory reproduction system 10 includes a hologram memory reproduction device (reproduction device) 1, a buffer memory 2, a two-dimensional waveform equalization circuit 3, a two-dimensional Viterbi decoding circuit 4, a decoded bit row output circuit 5, a decoded bit register 6, and an assumed waveform. A level setting circuit 7 and a branch metric calculation circuit 8 are provided.

  The hologram memory reproducing apparatus 1 corresponds to the structure on the reproducing side in the system described with reference to FIG. 9 in the background art column, and irradiates reproduction reference light to a hologram medium on which information is recorded as interference fringes. In this device, reflected light or transmitted light (coherent light) generated by the light is received by a light receiving element (for example, a CCD or CMOS sensor) having a plurality of pixels through a lens to generate a reproduction signal.

  Note that, as described in the Background Art section, a reproduction signal may be regarded as a waveform and referred to as a “reproduction waveform signal”.

  The buffer memory 2 is a storage device that stores the reproduction signal generated by the hologram memory reproduction device 1 in units of page data. Note that the reproduction signal in units of page data stored in the buffer memory 2 is a two-dimensional reproduction signal affected by the inter-pixel interference as described in the background art section.

  The two-dimensional waveform equalization circuit 3 performs two-dimensional waveform equalization on the reproduced waveform signal generated by the hologram memory reproducing device 1 (more precisely, the reproduced waveform signal stored in the buffer memory 2). Thus, the reproduced waveform signal is converted into a signal more suitable for the signal assumed in the assumed waveform level setting circuit 7. Since the configuration of the two-dimensional waveform equalization circuit 3 and the processing content equivalent to the waveform are well known, a detailed description thereof will be omitted.

  The two-dimensional Viterbi decoding circuit 4 is a circuit that performs two-dimensional Viterbi decoding by repeating the process of determining the surviving path for each trellis state based on the path metric obtained by accumulating the branch metrics obtained by the branch metric calculation circuit 8. It is.

  The decoded bit row output circuit 5 extracts and outputs an appropriate decoding result (decoded bit row corresponding to the decoding target row) from among the decoded bit sequences included in the path that survived the decoding in the two-dimensional Viterbi decoding circuit 4. It is a circuit to do.

  The decoded bit register 6 is a register that holds the decoded bit string output from the decoded bit row output circuit 5.

  The assumed waveform level setting circuit 7 uses the decoded bit string held in the decoded bit register 6 to calculate an assumed waveform level (assumed signal value) w used in the branch metric calculation circuit 8 and sends it to the branch metric calculation circuit 8. Circuit.

The branch metric calculation circuit 8 calculates a branch metric for each branch of the signal x waveform-equalized by the two-dimensional waveform equalization circuit 3 with a square error from the assumed waveform level w obtained by the assumed waveform level setting circuit 7. (X−w) ² is a circuit to be obtained.

  In the hologram memory reproduction system 10, the two-dimensional waveform equalization circuit 3, the two-dimensional Viterbi decoding circuit 4, the decoded bit row output circuit 5, the decoded bit register 6, the assumed waveform level setting circuit 7, and the branch metric calculation circuit 8. Constitutes the Viterbi decoding device 12.

  The two-dimensional Viterbi decoding circuit 4 is Viterbi decoding means, the decoded bit row output circuit 5 is decoded bit output means, the decoded bit register 6 is decoded bit storage means, the assumed waveform level setting circuit 7 is assumed signal value calculation means, and branch metrics. The arithmetic circuits 8 correspond to branch metric arithmetic means, respectively.

  FIG. 2 shows a predetermined response matrix h in the hologram memory reproducing apparatus 1. The operation of the two-dimensional Viterbi decoding circuit 4 assumes this response matrix h. Therefore, this response matrix h is referred to as an assumed response matrix h.

  The assumed response matrix h is a two-dimensional impulse response matrix assumed based on the amplitude of the light received by the light receiving element, not the intensity of the light received by the light receiving element during reproduction in the hologram memory reproducing apparatus 1. Note that “two-dimensional impulse response” means inputting a unit impulse signal (a two-dimensional bit matrix in which the signal level of only one bit is 1 and the signal levels of all other bits are all 0) to a linear system. The “two-dimensional impulse response matrix” means a matrix composed of the output signals.

  For simplification of explanation, the assumed response matrix h will be described in the case where the bit width of inter-pixel interference is 3. However, the assumed response matrix is not limited to this, and the assumed response matrix is close to the response characteristics of the hologram memory reproducing device 1. Should be assumed.

  The trellis diagram expressing the operation of the two-dimensional Viterbi decoding circuit 4 is the same as that of the background art shown in FIG. Since the bit width of inter-pixel interference of the assumed response matrix h is 3 and the decision feedback Viterbi decoding method is applied, the trellis state is defined as a matrix of 2 rows and 2 columns (one higher than the decoding target row). Since a feedback row can be used for the row of, the row width is 3-1 = 2). Since there are 2 rows and 2 columns (4 bits in total), the number of trellis states is 2 to the fourth power = 16, and there are four branches to be input and outputs to each trellis state.

  FIG. 3 shows an assumed waveform level of four branches input to the trellis state [1, 0; 0, 1] as an example of the assumed waveform level of the branch (assumed signal value assumed for the transition of the trellis state). Indicates the level. The assumed waveform level in FIG. 3 is obtained when the feedback row corresponding to the branch is [1, 0, 1]. The assumed waveform level is obtained as follows.

  The bit matrix defined by the branch of the trellis state [0, 1; 0, 0] → [1, 0; 0, 1], when combined with the feedback row [1, 0, 1], [1, 0, 1; 1, 0; 0, 0, 1].

  As shown in FIG. 4, when a two-dimensional convolution operation is first performed between this bit matrix and the assumed response matrix h, a value 0.5 at the center of the resulting matrix is assumed to be the light assumed in the transition of the trellis state. It is obtained as an assumed value corresponding to the amplitude. This is because the assumed response matrix h used in the two-dimensional convolution calculation is a two-dimensional impulse response matrix assumed based on the amplitude of the light, not the intensity of the light, as described above.

  On the other hand, since the signal detected by the CCD is a signal corresponding to the intensity of light, considering that there is a relationship between the amplitude of the light and the intensity, the square of the absolute value of the amplitude is the intensity. A value 0.25 obtained by squaring 0.5 obtained as the amplitude corresponds to the light intensity. Therefore, if this value 0.25 is used as the assumed waveform level, a more accurate assumed waveform level can be set.

  The assumed waveform levels of all branches can be obtained in the same manner. When the feedback row is [f1, f2, f3], the trellis state transition [p, q; s, t] → [q, r; t, The assumed waveform level w of the branch corresponding to u] is calculated as the square value of the center value of the two-dimensional convolution matrix of [f1, f2, f3; p, q, r; s, t, u] and h. The Such a method of determining the assumed waveform level w is a feature of the present invention.

  As in the prior art, if the two-dimensional convolution operation result of the bit matrix corresponding to the trellis state transition and the assumed response matrix h is used as it is as the assumed waveform level without being squared, a bit matrix ( For the assumed waveform level (assuming that the outside of the matrix is all 0), the result of the two-dimensional convolution operation of FIG. 4 itself is used.

  On the other hand, the reproduction signal obtained by detecting this bit matrix with the CCD is a signal close to the square of all the elements of the result of the two-dimensional convolution operation. For example, the level of the reproduction waveform signal in the third row is , 0.0009, 0.0576, 0.25, 0.1296, and 0.0036 (the actual reproduction waveform signal includes noise).

  Such a reproduced waveform signal does not correspond to the assumed waveform signal of 0.03, 0.24, 0.5, 0.36, 0.06 in level ratio. Even if the reproduced waveform signal is proportionally multiplied using an amplifier or the like, it cannot be matched with the squared value (for example, in order to match 0.25 in the middle of the reproduced waveform signal for 5 bits to the assumed waveform level). Even if the signal is amplified twice, the reproduced waveform signals 0.0018, 0.1152, 0.5, 0.2592, and 0.0072 after amplification only have the level of the middle bit equal to the assumed waveform level. Does not match the other bits).

  Therefore, when the assumed waveform level as in the prior art is used, the Viterbi decoding performance cannot be sufficiently obtained.

  As can be seen from the above description, the assumed waveform level of the branch varies depending on the feedback line. Therefore, the assumed waveform level setting circuit 7 calculates the assumed waveform level w of each branch based on the feedback row [f1, f2, f3], and the branch metric calculation circuit 8 varies the assumed waveform level w used for branch metric calculation. Set.

  The assumed waveform level w may be set by performing the calculation as in the above calculation example every time according to the feedback row [f1, f2, f3] input from the decoding bit register 6, or may be calculated in advance. The assumed waveform level may be stored in a memory such as a lookup table, and the corresponding value may be read from the memory and set in accordance with the feedback row.

  If the branch corresponding to the trellis state transition [p, q; s, t] → [q, r; t, u] survives, two bits [q; t] are determined as decoded bits. Only the bit q in the first row that is Viterbi-decoded with all waveform interference from the top, bottom, left and right taken into account is output as a decoded bit.

  Next, the reproducing operation by the hologram memory reproducing system 10 having the above-described configuration shown in FIG. 1 will be described with reference to the flowchart of FIG.

  In the hologram memory reproducing apparatus 1, reproduction reference light is irradiated onto a hologram medium (recording information of page information of 100 rows and 100 columns) and two-dimensional from a light receiving element (a CCD array of 100 rows and 100 columns). Since the operation of outputting the reproduction waveform signal is as described in the background art, details are omitted. The output two-dimensional reproduced waveform signal is stored in the buffer memory 2 (step S1).

  First, the decoding target row i is set to the first row (step S2), and Viterbi decoding in the row direction is started from the j = 1st column (step S3).

  Next, the reproduction waveform signal corresponding to the first row and the first column is read from the buffer memory 2 and input to the branch metric calculation circuit 8 as x through the two-dimensional waveform equalization circuit 3 (step S4).

  Here, the two-dimensional waveform equalization circuit 3 converts the reproduction waveform signal into a signal more suitable for the signal assumed in the branch metric calculation circuit 8, and this process is compared with a configuration in which the reproduction waveform signal is directly subjected to Viterbi decoding. Thus, decoding with a lower error rate can be realized. Therefore, although it is desirable to provide the two-dimensional waveform equalization circuit 3 from the viewpoint of error rate, the hologram memory reproduction system 10 can be configured by omitting the two-dimensional waveform equalization circuit 3 from the viewpoint of circuit scale. .

Next, the two-dimensional Viterbi decoding circuit 4 and the branch metric calculation circuit 8 execute Viterbi decoding in the row direction based on the trellis diagram of FIG. That is, the square error (x−w) ² between the input reproduction waveform signal x of the recording bit in the first row and the first column and the corresponding assumed waveform level w is obtained as a branch metric, and further to each trellis state. The surviving path is determined based on the branch metric accumulated in step 1, i.e., the path metric (step S5).

  Here, the assumed waveform level w used in the Viterbi decoding is obtained by the assumed waveform level setting circuit 7 from the bit matrix determined for each branch and the feedback rows [f1, f2, f3] stored in the decoded bit register 6. As described above, it is a numerical value.

  Next, if the path memory length of the two-dimensional Viterbi decoding circuit 4 is L, it is determined whether the current j-th column is before or after the L-th column (step S6), and the j-th column is after the L-th column. Then, the decoded bit row output circuit 5 extracts only the bit q of the first row from the bit 2 rows [q; t] in the trellis state of the (j−L) th column of the surviving path and outputs it as a decoded bit row. (Step S7).

  This decoded bit row is the decoding result itself, but is also stored in the decoded bit register 6 at the same time (step S8).

  The above processing from S4 to S8 is advanced by shifting one column at a time in the row direction (step S9), and when decoding for an extra portion of the path memory length L, that is, (100 + L) columns is completed (step S10). ) Shifting one row at a time in the column direction (step S11), and repeating the processing from S3 to S11 until all 100 decoded bit rows are obtained (step S12).

  At this time, since the decoded bit row on the first row is stored in the decoded bit register 6, this is used as the feedback row for Viterbi decoding of the next row.

  With the above operation, a decoding result bit map is finally obtained as page data.

  As described above, in the background art, the assumed waveform level of each branch cannot be determined correctly and high-performance Viterbi decoding cannot be realized. In the present embodiment, two trellis states connected to the branch and feedback Since the assumed waveform level can be obtained more accurately based on the square value of the two-dimensional convolution operation between the bit matrix determined from the row and the response matrix h, the effect of further reducing the error rate of the Viterbi decoding result can be obtained. Is possible.

[Embodiment 2]
A second embodiment of the present invention will be described with reference to FIGS.

  In the first embodiment, it is assumed that a predetermined response matrix h as shown in FIG. 2 is preliminarily assumed to be close to the response characteristics of the hologram memory reproducing device 1, and the assumed waveform level setting circuit 7 obtains the assumed waveform level. . In the present embodiment, a configuration for obtaining an assumed waveform level after obtaining a response matrix h by first performing test reproduction will be described.

  FIG. 6 is a block diagram showing a configuration of the hologram memory reproducing system 11 of this embodiment to which the Viterbi decoding device and the Viterbi decoding method of the present invention are applied. Note that, in the constituent elements of the hologram memory reproducing system 11, constituent elements having functions equivalent to those of the constituent elements of the hologram memory reproducing system 10 in Embodiment 1 are denoted by the same reference numerals and description thereof is omitted.

  The hologram memory reproduction system 11 is different from the hologram memory reproduction system 10 in that the hologram memory reproduction system 11 includes a response matrix calculation circuit 9 and a response matrix obtained by the response matrix calculation circuit 9. The assumed waveform level setting circuit 7 determines the assumed waveform level w using h.

  In the hologram memory reproduction system 11, the two-dimensional waveform equalization circuit 3, the two-dimensional Viterbi decoding circuit 4, the decoded bit row output circuit 5, the decoded bit register 6, the assumed waveform level setting circuit 7, the branch metric calculation circuit 8, The response matrix calculation circuit 9 constitutes a Viterbi decoding device 13, and the response matrix calculation circuit 9 corresponds to response matrix calculation means.

  The test reproduction operation by the hologram memory reproduction system 11 having the above-described configuration shown in FIG. 6 will be described with reference to the flowchart of FIG.

  When a hologram medium (not shown) is mounted on the hologram memory reproducing apparatus 1 (step S20), the hologram memory reproducing apparatus 1 first accesses a position where a specific pattern as shown in FIG. To reproduce a two-dimensional reproduction waveform signal (step S21).

  Since the specific pattern in FIG. 8 is a unit impulse signal having a central bit of 1 and all other values of 0, the reproduced waveform signal reproduced corresponds to a two-dimensional impulse response. When this two-dimensional impulse response is input to the response matrix calculation circuit 9, the square root of each matrix element is obtained (step S22) and is output to the assumed waveform level setting circuit 7 as a response matrix h (step S23).

  The subsequent normal reproduction operation is almost the same as the reproduction operation shown in the flowchart of FIG. 5 in the first embodiment, and thus detailed description thereof will be omitted. However, in step S5 of FIG. The difference is that an expected waveform level is set using the response matrix h obtained by the matrix calculation circuit 9 and a branch metric is calculated.

  In this way, every time a hologram medium is newly installed, a reproduction waveform signal corresponding to a two-dimensional impulse response is test-reproduced, and a response matrix h is obtained by calculating the square root, thereby obtaining a predetermined matrix. Compared to a configuration using a response matrix (pre-fixed), Viterbi decoding can be performed based on a more appropriate assumed waveform level in accordance with a difference in the characteristics of the hologram medium and a change in the reproduction environment. Low reproduction can be realized.

  However, the timing for performing the test reproduction is not limited to when the hologram medium is newly mounted, and may be periodically executed at regular time intervals. As a result, it is possible to follow fluctuations in the reproduction environment for a short time, and the reliability is further improved.

  In the hologram memory reproducing systems 10 and 11 described in the first and second embodiments, a Viterbi decoding method that performs two-dimensional Viterbi decoding in the row direction and repeats this in the column direction to decode the entire two-dimensional reproduction signal. However, since the definition of the row and column itself is meaningless, it is a matter of course that the configuration may be such that the row and the column are interchanged.

  Further, the relationship between the upper and lower sides and the left and right used in the description has no meaning, and a configuration in which the upper and lower sides and the left and right sides are interchanged is naturally acceptable. Further, since the present invention relates to a method for determining an assumed waveform level used for metric calculation, as long as metric calculation is performed, the effect can be obtained regardless of the Viterbi decoding method.

  In the first and second embodiments, the hologram memory system has been described as an example of the reproduction system. However, the reproduction system 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.

  Each block of the hologram memory reproducing systems 10 and 11 described in the first and second embodiments may be configured by hardware logic, or may be realized by software using a computer as follows.

  That is, the decoding apparatus (the apparatus excluding the hologram memory reproducing apparatus 1 and the buffer memory 2 in the hologram memory reproducing system 10 in FIG. 1 or the hologram memory reproducing system 11 in FIG. 6) realizes each function of this apparatus. CPU (central processing unit) that executes Viterbi decoding program instructions, ROM (read only memory) that stores the program, RAM (random access memory) that expands the program, memory that stores the program and various data, etc. It can also be realized by a computer having a storage device or the like.

  That is, an object of the present invention is to supply a computer with a recording medium in which a program code (execution format program, intermediate code program, source program) of a decoding program, which is software that realizes the above-described functions, is readable by a computer. This can also be achieved by the computer reading and executing the program code recorded on the recording medium.

  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.

  The characteristics of the present invention can also be expressed as follows. That is, the Viterbi decoding apparatus according to the present invention is an apparatus that reproduces a two-dimensional reproduction signal of page data using Viterbi decoding, and is a metric for obtaining an error of the reproduction signal with respect to an assumed signal value assumed for a trellis state transition. And a Viterbi decoding unit that performs Viterbi decoding based on the error, wherein the metric calculation unit is based on a square value of a two-dimensional convolution of a bit matrix determined by trellis state transition and a predetermined response matrix. The assumed signal value obtained in this way is used.

  Further, the Viterbi decoding device according to the present invention is the Viterbi decoding device, wherein the Viterbi decoding device outputs a decoded bit row corresponding to a decoding target row, as a decoding result, among the paths surviving as a result of the Viterbi decoding; A decoded bit storage means for storing the decoded bit row output by the decoded bit output means with respect to a decoding target row, a bit matrix determined from a trellis state transition and a decoded bit row stored by the decoded bit storage means, and a predetermined response An assumed signal value calculation means for obtaining the assumed signal value based on the square value of the two-dimensional convolution with the matrix may be provided.

  Also, the Viterbi decoding apparatus according to the present invention is a bit matrix test corresponding to a unit impulse signal in which only one predetermined bit has a value of 1 and all other bits have a value of 0 in the Viterbi decoding apparatus. Test reproduction means for generating a reproduction signal, and response matrix calculation means for obtaining the predetermined response matrix by obtaining the square root of the test reproduction signal generated by the test reproduction means may be further provided.

  Further, the Viterbi decoding method according to the present invention is a method of Viterbi decoding a two-dimensional reproduction signal of page data, a metric calculation process for obtaining an error of a reproduction signal with respect to an assumed signal value assumed for a trellis state transition, and the error Viterbi decoding processing for performing Viterbi decoding based on the metric calculation processing, the metric calculation processing is based on a square value of a two-dimensional convolution of a bit matrix determined by trellis state transition and a predetermined response matrix This is a method using values.

  In addition, the Viterbi decoding method according to the present invention is a recording medium on which a bit matrix for testing in which only one predetermined bit has a value of 1 and all other bits have a value of 0 in the above Viterbi decoding method is recorded. A test reproduction process for generating a test reproduction signal corresponding to a unit impulse signal, and a response matrix calculation process for obtaining a square root of the test reproduction signal generated in the test reproduction process to obtain the predetermined response matrix. May be included.

  According to the Viterbi decoding apparatus or Viterbi decoding method of the present invention, the assumed waveform level of each branch can be set correctly, so that high-performance Viterbi decoding can be realized and reproduction with a low error rate is possible. Become.

  In addition, by determining the response matrix corresponding to the impulse response using the test reproduction signal and obtaining the assumed waveform level, it is possible to use the assumed waveform level that is more suitable for each hologram memory medium and the reproduction environment. The error rate can be further reduced.

  The present invention is not limited to the above-described embodiments, and various modifications can be made within the scope of the claims, and embodiments obtained by appropriately combining technical means disclosed in different embodiments. Is also included in the technical scope of the present invention.

  The present invention that enables decoding with a low error rate for a two-dimensional signal can be applied to a hologram memory reproducing device for reproducing a two-dimensional signal, a two-dimensional barcode reproducing device represented by a QR code, and the like.

It is a block diagram which shows the structure of the hologram memory reproduction system in the 1st Embodiment of this invention. It is a figure which shows one example of the response matrix in the hologram memory reproduction | regeneration system of FIG. It is a figure which shows one example of the assumed waveform level corresponding to the branch in the hologram memory reproduction system of FIG. It is a figure explaining an example of the two-dimensional convolution calculation for calculating | requiring the assumed waveform level of FIG. 3 is a flowchart showing a reproduction operation of the hologram memory reproduction system of FIG. 1. It is a block diagram which shows the structure of the hologram memory reproduction system in the 2nd Embodiment of this invention. It is a flowchart which shows the reproduction | regeneration operation | movement of the hologram memory reproduction system of FIG. It is a figure which shows the unit impulse signal used for test reproduction | regeneration of the hologram memory reproduction | regeneration system of FIG. It is a block diagram which shows the structure of the conventional hologram memory recording / reproducing apparatus. It is a figure which shows the assumed impulse response of the conventional hologram memory reproduction | regeneration system. It is a trellis diagram expressing the conventional decision feedback Viterbi decoding method.

Explanation of symbols

1 Hologram memory playback device (playback device)
2 buffer memory 3 2D waveform equalization circuit 4 2D Viterbi decoding circuit (Viterbi decoding means)
5 Decoded bit row output circuit (decoded bit output means)
6 Decoding bit register (decoding bit storage means)
7 Assumed waveform level setting circuit (assumed signal value setting means)
8 Branch metric calculation circuit (branch metric calculation means)
9 Response matrix calculation circuit (response matrix calculation means)
DESCRIPTION OF SYMBOLS 10 Hologram memory reproduction system 11 Hologram memory reproduction system 12 Viterbi decoding apparatus 13 Viterbi decoding apparatus

Claims (7)

In a Viterbi decoding device that Viterbi-decodes a two-dimensional reproduction signal obtained by optically reproducing page data and indicating a signal value corresponding to the intensity of light,
An assumed amplitude value, which is an assumed value of trellis state transition corresponding to the two-dimensional reproduction signal, assumed based on the amplitude of light is calculated, and an assumed signal value corresponding to the square value of the assumed amplitude value is obtained. Assumed signal value calculation means;
Branch metric calculation means for obtaining a branch metric of the transition based on the assumed signal value calculated by the assumed signal value calculation means and the two-dimensional reproduction signal;
Viterbi decoding apparatus according to claim Rukoto a Viterbi decoding means for Viterbi decoding is performed on the basis of the branch metric. In a Viterbi decoding device that Viterbi-decodes a two-dimensional reproduction signal obtained by optically reproducing page data and indicating a signal value corresponding to the intensity of light,
Branch metric calculation means for obtaining a branch metric of the transition based on the transition of the trellis state corresponding to the two-dimensional reproduction signal and the two-dimensional reproduction signal;
Viterbi decoding means for performing Viterbi decoding based on the branch metric,
When the assumed value of the transition of the trellis state assumed based on the amplitude of light is an assumed amplitude value,
The branch metric calculation means obtains a branch metric using an assumed signal value corresponding to a square value of the assumed amplitude value,
The Viterbi decoding means performs Viterbi decoding for a plurality of rows including a decoding target row in the page data,
Of the paths surviving as a result of Viterbi decoding performed by the Viterbi decoding means, decoded bit output means for outputting a decoded bit row corresponding to the decoding target row as a decoding result;
Decoded bit storage means for storing the decoded bit row output by the decoded bit output means with respect to the immediately preceding decoding target row;
The assumed amplitude value by a two-dimensional convolution operation of a bit matrix determined from a trellis state transition and a decoded bit row output from the decoded bit storage means and a two-dimensional impulse response matrix assumed as the amplitude of light in the reproduction further comprising an assumed signal value calculating means for obtaining said assumed signal value to calculate the characteristics and to the ruby Tabi decoding apparatus.   The response matrix calculation means for generating the two-dimensional impulse response matrix by obtaining a square root of the two-dimensional reproduction signal obtained by reproducing a bit matrix corresponding to a unit impulse signal is provided. The Viterbi decoding device described. A Viterbi decoding device according to any one of claims 1 to 3,
A reproduction system comprising: a reproduction device that obtains the two-dimensional reproduction signal by optically reproducing page data. In a Viterbi decoding method for Viterbi decoding a two-dimensional reproduction signal obtained by optically reproducing page data and indicating a signal value corresponding to light intensity,
An assumed amplitude value, which is an assumed value of trellis state transition corresponding to the two-dimensional reproduction signal, assumed based on the amplitude of light is calculated, and an assumed signal value corresponding to the square value of the assumed amplitude value is obtained. Assumed signal value calculation processing,
A branch metric calculation process for obtaining a branch metric of the transition based on the assumed signal value calculated in the assumed signal value calculation process and the two-dimensional reproduction signal;
And a Viterbi decoding process for performing Viterbi decoding based on the branch metric .   A Viterbi decoding program for causing a computer to function as each of the means in order to cause the computer to operate as the Viterbi decoding device according to any one of claims 1 to 3.   A computer-readable recording medium in which the Viterbi decoding program according to claim 6 is described.

Images