JPH0836837A - Learning type signal recording and reproducing device - Google Patents

Abstract

PURPOSE:To obviate erroneous detection by noises and distortions and to stably reproduce signals by converting teacher signals recorded in the header part of a recording medium to reproduced signals by using a conversion rule formed by learning. CONSTITUTION:The reference teacher signals formed by a teacher signal generator 3 are recorded in the header part of the recording medium 6 at the time of recording the data. The relations between the results obtd. by detecting the teacher signals of the header part by a detector 7 and the waveforms of the reference teacher signals outputted from the teacher signal generator 3 have learned by a learning type waveform conversion section 2 at the time of reproducing. The conversion rule for exactly reproducing the analog signals detected from the recording medium 6 to the signals used as the basis for the recording is automatically formed and signal conversion is executed. As a result, the signals are stably reproduced without being affected by the noises and differences between the individuals of the recording media. Further, the formed conversion rule is preserved in a memory 4. The conversion rule is loaded from the memory 4 and the relations with the reference teacher signals are learned from the next time to early converge the learning.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a learning type signal recording / reproducing apparatus for recording / reproducing a signal on / from a recording medium such as an optical disc, a magneto-optical disc, a magnetic disc or a magnetic tape.

[0002]

2. Description of the Related Art As a conventional technique for reproducing a signal recorded on a recording medium such as an optical disk, a magnetic disk, or a magnetic tape, a detected signal is converted into 0 or 1 by using a predetermined threshold value. A device that outputs the result as a reproduction signal has been proposed.

The structure of such a conventional signal recording / reproducing apparatus is briefly shown in FIG. The signal recorded on the recording medium 81 is detected by the signal detector 82. The signal detector 82 is connected to the comparator 83 together with a threshold memory 84 that stores the threshold. The signal detected by the signal detector 82 is input to the comparator 83, where it is compared with the threshold value read from the threshold value memory 84. Comparator 83
Is 1 if the signal from the signal detector 82 is greater than or equal to the threshold value.
, Is output as a reproduction signal if it is smaller than the threshold value. By obtaining the reproduced signal in this way, the conventional signal recording / reproducing apparatus achieves stable signal reproduction that is not affected by signal distortion due to noise or the like.

[0004]

However, in such a conventional signal recording / reproducing apparatus, for example, as shown in FIG.
When a signal with a large ripple is input from the signal detector 82 to the comparator 83, an incorrect reproduction signal may be output. Such erroneous detection may occur due to various factors such as a case where the signal detected by the signal detector 82 is affected by noise, a difference between individual recording mediums, or a difference in conditions for signal detection. .

The present invention has been made in view of such a situation as described above, and an object of the present invention is to learn so that a signal can be stably reproduced without being affected by noise and a difference between individual recording media. A type signal recording / reproducing apparatus is provided.

[0006]

A learning type signal recording / reproducing apparatus of the present invention comprises a signal detecting means for detecting a teacher signal and an information signal recorded on a recording medium, and a signal detecting means for detecting the teacher signal and the information signal. Learning-type signal recording including conversion means for converting the teacher signal and the information signal into a reproduction teacher signal and an information reproduction signal, respectively, based on a predetermined conversion rule, and teacher signal generating means for generating a reference teacher signal. In the reproducing apparatus, the conversion means detects the teacher signal prior to detecting the information signal and generates the predetermined conversion rule based on the reproduction teacher signal and the reference teacher signal. And a waveform conversion unit, thereby achieving the above object.

The converting means may include a memory means for storing the predetermined conversion rule generated by the learning type waveform converting section.

The converting means extracts a change in the signal value at each point on the transition axis of the detected teacher signal, and outputs the change in the extracted signal value as a series signal. The learning type waveform converting section further has a memory control section for storing a predetermined conversion rule in the memory means, and the learning type waveform converting section learns a relationship between the extracted change in the signal value and the waveform of the reference teacher signal. By doing so, the predetermined conversion rule may be automatically generated.

The learning type waveform converting section converts the detected teacher signal into the reproduced teacher signal according to the other conversion rule when another conversion rule is already stored in the memory means. The predetermined conversion rule may be generated based on the reproduction teacher signal and the reference teacher signal.

The learning type waveform conversion section loads the other conversion rule by the memory control section when another conversion rule is already stored in the memory means,
The detected teacher signal may be converted into the reproduction teacher signal according to the other conversion rule, and the predetermined conversion rule may be generated based on the reproduction teacher signal and the reference teacher signal.

The converting means further comprises a comparing means for comparing the reproduction teacher signal with the reference teacher signal, and the memory means substantially matches the reproduction teacher signal with the reference teacher signal. At this time, the predetermined conversion rule may be stored.

The learning type waveform conversion section has a network having a plurality of layers, and at least one of the plurality of layers has a signal input section for inputting the sequence signal and the signal input section. A quantizer for receiving the sequence signal from the unit and quantizing the sequence signal; at least one first path input terminal; and at least one first path input terminal connected to the at least one first path input terminal. By changing the connection manner of one path output terminal, the at least one first path input terminal and the at least one first path output terminal according to the quantized sequence signal, A plurality of first unit recognition units, each of which has a path selection unit for selecting a path, are included, and a layer immediately before the output layer of the plurality of layers includes a teacher for inputting the reference teacher signal. Signal input terminal And at least one second path input terminal and at least one second path input terminal connected said at least one second path input terminal,
Each has a path learner that changes the strength of the connection with the at least one second path input terminal and one of the at least one second path output terminals pointed to by the reference teacher signal. A plurality of second unit recognition units are included, and the learning type waveform conversion unit is a unit having the largest output value among the plurality of third unit recognition units included in the output layer of the plurality of layers. An output result determination unit that detects the recognition unit may be further provided.

The sequence signal extraction unit extracts the signal value of a predetermined point on the transition axis of the detected teacher signal, and a current signal extraction unit that extracts the signal value of the predetermined point extracted by the current signal extraction unit. A change amount extraction unit that calculates a difference between the signal value and the signal value of the teacher signal before the predetermined point may be included.

The learning type waveform conversion section has a network of a plurality of layers each including a plurality of units,
The predetermined conversion rule may be generated by changing the strength of connection between adjacent layers of the plurality of layers by learning using the reproduction teacher signal and the reference teacher signal.

The learning type waveform conversion section has a network of a plurality of layers each including a plurality of units,
The predetermined conversion rule may be generated by changing a connection method between adjacent layers of the plurality of layers by learning using the reproduction teacher signal and the reference teacher signal.

The predetermined conversion rule may be stored in the memory means as a way of connecting between the adjacent layers.

The method of connection between the adjacent layers is
It may be changed by changing the strength of each connection between the unit contained in one of the adjacent layers and the unit contained in the other of the adjacent layers.

The multi-layer network has first, second, third and fourth layers arranged in this order, the first layer receiving the detected teacher signal. , The fourth layer outputs the reproduction teacher signal, and the reference teacher signal is given to a unit included in the third layer, whereby the unit included in the third layer and the fourth unit are included. The strength of the connection with the units included in the layer may be varied.

The strength of the connection between the unit included in the third layer and the unit included in the fourth layer is such that the reproduction teacher signal output from the fourth layer and the reference teacher signal are output from each other. It may be changed to substantially match.

When the data is recorded on the recording medium, the learning type signal recording / reproducing apparatus may record the reference teacher signal generated by the teacher signal generating means before recording the data.

The at least one second unit in the plurality of second unit recognition units of the learning type waveform conversion unit.
The connection method between the path input terminal of and the at least one second path output terminal may be stored in the memory means as the predetermined conversion rule.

The sequence signal extracting section generates another sequence signal by extracting a change in the signal value at each point on the transition axis of the detected information signal, and when the information signal is detected, The memory control section of the learning type waveform conversion section loads the predetermined conversion rule stored in the memory means, and in the plurality of first unit recognition units, the signal input section is different from the other one. A sequence signal is input, the quantizer quantizes the other sequence signal, and the route selection unit connects the at least one first route input terminal and the at least one first route output terminal. In accordance with the quantized other sequence signal, and in the plurality of second unit recognition units, the at least one second path input terminal and the at least one second
The path output terminals may be connected in the manner of connection indicated by the predetermined conversion rule, thereby converting the information signal into the reproduction information signal according to the predetermined conversion rule.

[0023]

In the learning type signal recording / reproducing apparatus of the present invention, when recording data on the recording medium, the reference teacher signal having a predetermined waveform is recorded in, for example, the header portion, and then the actual data is recorded. I'll do it. This reference teacher signal is generated by the teacher signal generator of the learning type recording / reproducing apparatus of the present invention. When the data is reproduced from the recording medium on which the reference teacher signal is recorded, the learning-type signal recording / reproducing apparatus first detects the teacher signal recorded on the header portion of the recording medium by the signal detector. Then, the learning type waveform conversion unit is made to learn the relationship between the detection result and the reference teacher signal output from the teacher signal generator. As a result, the learning-type waveform conversion unit automatically generates a conversion rule for accurately reproducing the analog signal detected from the recording medium into the signal that is the basis of the recording.

When the learning is completed, the signal recorded on the recording medium is converted according to the generated conversion rule. In this way, the learning-type signal recording / reproducing apparatus of the present invention can stably reproduce a signal without being affected by noise or a difference between individual recording media.

The conversion rule generated by learning is stored in the non-volatile memory unit. From the next playback, before detecting the teacher signal recorded on the recording medium, first,
The learning-type waveform conversion unit loads the conversion rule from the nonvolatile memory unit. Then, the learning-type waveform conversion unit calculates the relationship between the signal obtained by converting the teacher signal recorded on the recording medium by the signal detector according to the conversion rule and the reference teacher signal output by the teacher signal generator. Let them learn. This allows learning to converge faster,
As a result, the rise time of the device can be very fast. In addition, the conversion rule corrected by learning is
Each time it is modified, it is stored in the non-volatile memory unit.

[0026]

Embodiments of the present invention will be described below with reference to the drawings.

FIG. 1A is a block diagram showing the configuration of the learning type signal recording / reproducing apparatus 100 of the present invention. The learning-type signal recording / reproducing apparatus 100 does not obtain a reproduction signal by simply comparing the signal detected by the signal detection unit with a threshold value, but converts the detected signal according to a conversion rule generated by learning. Thus, a reproduction signal is obtained. This transformation rule can be modified by subsequent learning generated as detailed below.

The learning type signal recording / reproducing apparatus 100 comprises a sequence signal extracting section 1, a learning type waveform converting section 2, a teacher signal generator 3,
It has a non-volatile memory unit 4, a conversion result comparator 5, and a signal detector 7. The sequence signal extraction unit 1 extracts each point on the transition axis of the analog signal detected by the signal detector 7 that sequentially detects the signal recorded on the recording medium 6, that is, the change in the signal level at various times. , And inputs this to the learning type waveform conversion unit 2.

In addition to the output of the sequence signal extraction unit 1, the learning type waveform conversion unit 2 also receives a reference teacher signal having a predetermined waveform generated by the teacher signal generator 3. When the signal detected by the signal detector 7 is a signal based on the recording of the reference teacher signal, the learning type waveform converter 2
Is the change in the level of the detected signal and the teacher signal generator 3
Learn the relationship with the reference teacher signal from. This allows
The learning-type waveform conversion unit 2 converts a signal detected from the recording medium 6 into a signal that is the basis of recording on the recording medium 6, that is, converts the detected signal into 0 or 1 A rule to convert to is automatically generated. If the signal detected by the signal detector 7 corresponds to normal data, the learning-type waveform conversion unit 2 converts the signal detected according to the generated conversion rule into the signal that is the basis of recording. Convert. The generated conversion rule is stored in the non-volatile memory unit 4.

The operation of the learning type signal recording / reproducing apparatus 100 having such a configuration will be described in more detail below.

When data is recorded on the recording medium 6 using the apparatus 100 of FIG. 1A, the reference teacher signal generated by the teacher signal generator 3 is used prior to the data recording.
As shown in (b), it is recorded on the recording medium 6, for example, in the header portion.

When the apparatus 100 reproduces the recording medium on which the reference teacher signal is recorded in advance, the learning type waveform converting section 2 first accesses the non-volatile memory section 4. If the conversion rule is already stored in the non-volatile memory unit 4, the learning type waveform conversion unit 2 loads the conversion rule. Next, the signal detector 7 detects the teacher signal recorded in the header portion of the recording medium 6. The detected signal is sent to the sequence signal extraction 1, and the change in the level of the signal in the signal detected here is extracted.

The learning type waveform converter 2 converts the detected signal based on the change in the level of the detected signal and the conversion rule. Here, when the conversion rule is not stored in the non-volatile memory unit 4 and is not loaded from the memory unit 4, the conversion rule in the initial state before learning is used. The signal thus converted is output from the learning-type waveform conversion section 2 as a reproduction signal.

Then, the conversion result comparator 5 compares the reproduced teacher signal with the reference teacher signal. As a result of the comparison, if the reproduced teacher signal and the reference teacher signal match, the conversion rule used for the conversion is stored in the non-volatile memory unit 4 as it is. If the conversion rule in the initial state before learning was used for the conversion at this time, the conversion rule in the initial state is stored in the non-volatile memory unit 4 as a rule to be loaded at the next reproduction. However, if they do not match, learning of the relationship between the change in the level of the teacher signal reproduced from the header section and the reference teacher signal generated by the teacher signal generator 3 is performed in the learning-type waveform conversion section 2. That is, the rule obtained by learning is stored in the non-volatile memory unit 4 as a new conversion rule. That is, the correction by learning the conversion rule and the update of the conversion rule in the non-volatile memory unit 4 are performed. Upon completion of correction and update of the conversion rule, the learning type waveform conversion unit 2 thereafter converts the detected signal into 0 or 1 according to this conversion rule and outputs this as a reproduction signal. As a result, the learning-type signal recording / reproducing apparatus of the present invention can stably reproduce the signal on which the recording is based, even if the detected signal has distortion due to noise or individual difference of the recording medium. it can.

Next, the main parts of the learning type signal recording / reproducing apparatus 100 will be described.

FIG. 2 is a diagram simply showing the configuration of the sequence signal extraction unit 1. The sequence signal extraction unit 1 has a current signal extraction unit 1 and a change amount extraction unit 12. The current signal extraction unit 1 receives the analog signal detected by the signal detector 7, extracts the height of the waveform, that is, the signal level, and supplies it to the change amount extraction unit 12. The change amount extraction unit 12 has a signal storage unit 121 and a difference calculation unit 122. The signal storage unit 121 includes n + 1 signal storage units 1211, and each point in a predetermined range of the transition axis of the detected analog signal (waveform signal), that is, signals at various times within a predetermined section. The levels (signal values) of are sequentially stored. The difference calculation unit 122 calculates the difference between the signal value of the current analog signal extracted by the current signal extraction unit 11 and each of the signal values stored in the signal storage unit 121.

More specifically, the sequence signal extraction unit 1
When an analog signal is input to the current signal extraction unit 11, the current signal extraction unit 11 extracts the height (signal value) of the waveform of the analog signal, and sequentially extracts the height (signal value) of the change amount extraction unit 12 from the signal storage unit 121 and the difference calculation unit 122. Give to both. The signal storage unit 121 is
Similarly to the shift register, the signal values sequentially input are sequentially stored in the signal storage unit 1211 while shifting the storage position. Therefore, the signal value extracted at time t
When s (t) is input from the current signal extraction unit 11, FIG.
As shown in, the maximum n + 1 signal values s (t) to s (tn) are stored in the signal storage unit 121.

On the other hand, the difference calculation unit 122 receives the signal value s (t) at the time t input from the current signal extraction unit 11 and the signal value s (ti) (i ≠ t stored in the signal storage unit 121). )
The difference ds (ti) with each of the above is obtained as shown in the following equation (1), and a maximum of n-1 difference values ds (t1) to ds (tn) are output.

Ds (ti) = s (t) −s (ti) (1) In this way, the sequence signal extraction unit 1 is configured so that the current signal extraction unit 1
The current signal value s (t) extracted by 1 and the difference values ds (t1) to ds (tn) obtained by the change amount extraction unit 12 are treated as a set of series signals by the learning-type waveform conversion unit 2 Output to.

Next, the configuration and operation of the learning type waveform conversion section 2 will be described.

FIG. 3 briefly shows the structure of the learning type waveform conversion section 2. Here, it is assumed that the set of series signals output from the series signal extraction unit 1 includes five series signals of s (t) and ds (t1) to ds (t4). In the present invention, a network having a four-layer hierarchy in which a plurality of unit recognition units n are combined is used as the learning type waveform conversion section 2. Such networks are, for example, USP 5,2
65,224. The learning type waveform conversion unit 2
Input sequence signals one by one to the unit of the first layer, and
It is determined whether the waveform at is "0" or "1". The fourth layer of this network is the third
It includes two unit recognition units p1 and p2 interconnected with the output terminals of the units of the layers, each outputting a degree of match to "0" and a degree of match to "1". These outputs are given to the output result determination unit 20. The output result determination unit 20 compares these output values,
If the output value from p1> the output t from p2, then “0”
Otherwise, "1" is output as the final determination result of the network.

The first layer is an input layer having the same number of units as the number of sequence signals. Here, since there are five sequence signals s (t) and ds (t1) to ds (t4), the number of units included in the first layer is also five. The second layer and the third layer will be described in detail later.

The memory control unit 21 is connected to the third layer of the four-layer network. The memory controller 21 includes an output terminal of the unit recognition unit of the third layer and a fourth terminal.
The load of mutual coupling with the input terminals of the unit recognition units p1 and p2 of the layer is loaded from the nonvolatile memory unit 4 or saved in the nonvolatile memory unit 4.

Next, the structure of the unit recognition unit of the second layer will be described with reference to FIGS. 4 and 5.

FIG. 4 is a diagram simply showing the structure of the unit recognition unit included in the second layer. All the plurality of unit recognition units included in the second layer have the configuration shown in FIG. One unit recognition unit n included in the second layer has a signal input unit n1, a quantizer n2, and a path selector n3. The signal input unit n1 sends the sequence signal input via the signal input terminal n1a to the quantizer n2. The quantizer n2 quantizes the received sequence signal and supplies the quantized value to the route selection unit n3. The route selection unit n3 has an input terminal connected to the output of the quantizer n2, a route input terminal n3a, and route output terminals n3b1 to n3b3. The route input terminal and the route output terminal can be connected between units when a plurality of unit recognition units are combined to form a network. In addition, the route input terminal n3a and the route output terminals n3b1 to n
The method of connection with 3b3 can be changed according to the value input from the quantizer n2.

FIG. 5 shows an example of a more detailed configuration of such a route selection unit n3. In this example, the route selection unit n3 has one route input terminal n3a and three route output terminals n3b1, n3b2, n.
3b3, load of connection between path input terminal n3a and each path output terminal
It includes n3w1, n3w2, n3w3, and a load setter n3s for setting these loads. The loads n3w1, n3w2, and n3w3 are weights that are multiplied by the signal input from the route input signal n3a, and the multiplied result is the route output terminals n3b1, n3.
Output from b2 and n3b3. The load setter n3s, among the connection strengths of the path input terminal n3a and each path output terminal, the load n3w1 so that the strength of the connection with the path output terminal indicated by the value from the quantizer n2 is the largest. , N3w2 and n3w3 are set.

Next, the structure of the unit recognition unit of the third layer will be described.

FIG. 6 is a diagram showing the structure of the unit recognition unit included in the third layer. The unit recognition units included in the third layer all have the configuration shown in FIG. Figure 6
The unit recognition unit n shown in is a path input terminal n3a,
The route output terminals n3b1 to n3b16 and the loads n3w1 to n3w16 for changing the strength of these connections are provided, and the signals input from the route input terminal n3a are multiplied by the respective loads, and Is output.

The unit recognition unit included in the third layer further has a teacher signal input terminal n1t. From the teacher signal input terminal n1t, the number of the waveform signal (“0” or “1”) to be learned in the learning process, that is, the unit number of the unit recognition unit p1 or p2 included in the fourth layer is input as a teacher signal. To be done. Here, in the "learning process", for example, the teacher signal recorded in the header portion is detected, and the relationship between the change in the signal level of the detected teacher signal and the reference teacher signal from the teacher signal generator 3 is learned. Therefore, it means the process of generating the conversion rule. Therefore,
The reference teacher signal from the teacher signal generator 3 is input to the teacher signal input terminal n1t in the learning process. Terminal
The teacher signal input from n1t is given to the path learning device n3c. In the learning process, the route learner n3c changes the weights n3w1 to n3w16 based on the teacher signal, thereby maximizing the strength of the connection between the route output terminal and the route input terminal indicated by the teacher signal.

In the actual recognition process, that is, in the process of actually reproducing the detected signal using the conversion rule already generated, no signal is input to the teacher input terminal n1t. Therefore, the weights n3w1 to n3w16 are not changed by the route learning device n3c and retain the state after learning. Then, the route signal input from the route input terminal n3a is weighted and output to the route output terminals n3b1 to n3b16. The weights n3w1 to n3w16 after learning are stored in the non-volatile memory unit 4 by the memory control unit 21, and are held until the weights n3w1 to n3w16 are reset next. Further, the loads n3w1 to n3w16 stored in the nonvolatile memory unit 4 can be loaded by the memory control unit 21.

Here, the configuration of the route learning device n3c will be briefly described. As shown in FIG. 6, the route learning device n3c includes a maximum output terminal detector n3c1, a comparison / determination device n3c2, and a weight increase device n3c3.
It has. The maximum output terminal detector n3c1 detects the maximum output among the path output terminals n3b1 to n3b16 and gives the detection result to the comparison / determination unit n3c2. Comparison judge
n3c2 compares the detection result of the maximum output terminal detector n3c1 with the teacher signal input from the teacher signal input terminal nt1 in the learning process. More specifically, the comparison / determination device n3
c2 compares the number of the terminal with the maximum output, which is indicated by the detection result of the maximum output terminal detector n3c1, with the number of the output terminal pointed to by the teacher signal, and when both terminal numbers do not match. Outputs 0, and if they match, 1
Is output. The output 0 or 1 of the comparison / determination unit n3c2 is given to the load increase unit n3c3. The load increaser n3c3 is a comparison judge
Based on the output from n3c2, among the connections of the path input terminal n3a and the path output terminals n3b1 to n3b16, the strength of the connection with the path output terminal indicated by the teacher signal,
That is, the load is increased so that the load becomes maximum.
Therefore, if the output from the comparison / determination unit n3c2 is 1,
This means that the strength of the connection between the path output terminal and the path input terminal, which is indicated by the teacher signal, is in the maximum state, so that it is not necessary to change the load. However, when the output from the comparison / determination unit n3c2 is 0,
This means that there is a path output terminal connected to the path input terminal that is stronger than the path output terminal indicated by the teacher signal. Therefore, the load increaser n3c3 is
The load is changed so that the detection result of the maximum output terminal detector n3c1 matches the teacher signal.

FIG. 7 is a diagram showing the configuration of the unit recognition units p1 and p2 included in the fourth layer. Each of the unit recognition units p1 and p2 has a plurality of path input terminals.
n3a, one output terminal n3b, an adder n3as for adding the inputs from the input terminal n3a, and a threshold processor n3bs for thresholding the output from the adder n3as and outputting it to the output terminal n3b. A sigmoid function, a step function, or the like can be used as the function used by the threshold value processor n3bs for the threshold value processing.

Next, the learning operation of the learning type waveform conversion section 2 having the network composed of the layers described above will be described.

First, 1 is given as a route signal to the route input terminal n3a of the unit recognition unit n of the second layer.

The five sequence signals s (t) and ds (t are extracted by the sequence signal extraction unit 1 into the five units in the first layer.
When 1) to ds (t4) are input, these sequence signals are given to the quantizer n2 of the unit recognition unit n of the second layer. Each of the quantizers n2 quantizes the received sequence signal and, based on the quantized values, a load n3 that is a weight of a connection between the unit recognition unit of the second layer and the unit recognition unit of the third layer.
Set w1, n3b2 and n3w3. Thereby, the path output terminals n3b1, n3b2 and n3 of the unit recognition unit n of the second layer
From each of b3, a product obtained by multiplying the load corresponding to the route signal "1" given to the route input terminal n3a is output.

The route signals output from the route output terminals n3b1, n3b2 and n3b3 of each unit recognition unit of the second layer are sent to the route input terminals of the corresponding unit recognition unit of the third layer. Each unit recognition unit of the third layer weights the route signal input from the route input terminal with the loads n3w1 to n3w16 set at that time, and outputs the weighted signal to the fourth layer via the route output terminals n3b1 to n3b16. At this time, in each of the third unit recognition units, the maximum output terminal detector n3c1 detects the number of the route output terminal having the maximum output, and supplies it to the comparison / determination unit n3c2. At the same time, the number of the waveform to be learned, that is, the teacher signal indicating which of the unit recognition units p1 and p2 of the fourth layer should generate an output is input from each teacher recognition terminal nt1 to each unit recognition unit. This is also given to the comparison decision unit n3c2. The comparison / determination unit n3c2 determines which output of the path output terminals should be maximized from the given teacher signal, and determines the number of the output terminal indicated by the teacher signal as the maximum output terminal detector n3c1. Compare with the output terminal number detected by. Then, the comparing and determining unit n3c2 inputs 0 to the load increasing unit n3c3 when they do not match, and when they match.

When the input is 0, that is, when the route output terminal that outputs the maximum output is different from the route output terminal indicated by the teacher signal, the load increaser n3c3 and the teacher input signal n3a are different from each other. Increase the strength of the connection with the path output terminal, that is, the load. In this way, in each of the unit recognition units of the third layer, the route learning device n3c applies a load such that the route output terminal that outputs the maximum output matches the route output terminal indicated by the teacher signal. learn. When the learning is completed, the learned weights n3b1 to n3b16 are stored in the non-volatile memory unit 4 via the memory control unit 21 as a conversion rule for obtaining the signal which is the basis of the recording from the detected signal. .

As described above, the learning of the learning type waveform converter 2 of this embodiment is input to the signal input terminal n1a through the input terminal of the first layer in each unit recognition unit of the second layer. The sequence signal is quantized, and the method of coupling with the upper layer is set according to the quantized value. In each unit recognition unit of the third layer, the route input terminal and the route output are output based on the given teacher signal. It is only necessary to change the strength of the connection with the terminal by the route learning device. Therefore, very fast learning is possible.

Next, the actual waveform conversion operation of the learning type waveform conversion section 2 shown in FIG. 3 will be described.

Learning is performed until the information signal recorded after the header portion and corresponding to the actual data is detected by the signal detecting portion 7 and input to the learning type waveform converting portion 2 via the sequence signal extracting portion 1. Since it is the same as the operation at that time, its explanation is omitted.

In the learning type waveform conversion section 2, first, 1 is given as a route signal to the route input terminal n3a of each unit recognition unit of the second layer, as in the case of the learning operation. Also, the sequence signals s (t), ds (t1) to ds extracted by the sequence signal extraction unit 1
(t4) is input to the signal input terminal n1a of the unit recognition unit of the second layer via the five input terminals of the first layer. The quantizer n2 of each unit recognition unit in the second layer quantizes the input sequence signal. Based on this quantized value, in the route selection unit n3, the load setter n3s sets the loads n3w1 to n3w3 for coupling with the unit recognition unit in the upper layer, and the load is applied to the route signal input from the route input terminal n1a n3w
Multiplying 1 to n3w3 is the route output terminal n3b1 to n
It is output to the path input terminal of the unit recognition unit of the third layer via 3b3. In each unit recognition unit of the third layer, the teacher signal is not input during the waveform conversion operation. Therefore, the loads n3w1 to n3w16 of the unit recognition unit of the third layer hold the state at the time of learning, weight the route signal input from the route input terminal n3a, and output the weighted signal to the route output terminals n3b1 to n3b16. Output. Path output terminal n3b
1 to n3b16 send route signals to the route input terminals n3a of all unit recognition units in the fourth layer.

The adder n3as in the path input section of each unit recognition unit of the fourth layer adds all the input path signals,
Send to the threshold processor n3bs. The threshold processor n3bs performs threshold processing on the output from the adder n3as and outputs the output to the output terminal n3b. Therefore, if the value obtained by the addition is larger than the predetermined threshold value, the output from the unit recognition unit is generated. That is, each of the unit recognition units p1 and p2 outputs the degree of whether the input sequence signal is “0” or “1”.

Unit recognition units p1 and p2 of the fourth layer
The output from each is sent to the output determination unit 20. The output determination unit 20 compares the output value of the unit p1 with the output value of the unit p2, and outputs “0” as a final result if p1> p2, and otherwise outputs “1”.

As described above, the learning type waveform conversion section 2 of the present embodiment quantizes the sequence signal input to the signal input terminal n1a via the input terminal of the first layer in each unit recognition unit of the second layer. And the method of connection with the upper layer is set according to the quantized value, and in each unit recognition unit of the third layer, the route input terminal n3 is set by the weight set in the learning process.
The path signals input from a are weighted and the weighted path signals are output to the path input terminals of all the unit recognition units of the fourth layer via the path output terminals n3b1 to n3b16. Therefore, based on the learning result, it is possible to perform threshold processing on the input sequence signal very quickly and accurately.

[0065]

As described above, in the learning type signal recording / reproducing apparatus of the present invention, at the time of recording, the teacher signal is generated by the teacher signal generator, and the teacher signal is generated in the header portion before the data to be actually recorded. Record and then record the data. During reproduction, first, the teacher signal recorded in the header portion of the recording medium is detected, and the relationship between the detected teacher signal and the teacher signal generated from the teacher signal generator is learned. As a result, a conversion rule for converting the detected signal into a signal on which the signal is recorded is automatically generated. Here, the load of each unit recognition unit corresponds to the conversion rule. By using the conversion rule generated by learning in this way, the signal recorded after the header part is converted into a reproduction signal. Therefore, even if a signal having distortion due to noise or a difference between recording media is detected, a reproduced signal can be accurately obtained.

Further, in the learning type signal recording / reproducing apparatus of the present invention, the conversion rule generated by learning is stored in the non-volatile memory section, and the conversion rule is loaded from the non-volatile memory section at the next reproduction. That is, the learning of the load is started from the state set by the previous learning. For this reason, the learning converges quickly, and the rise time of the signal recording / reproducing device can be made very short.

[Brief description of drawings]

FIG. 1 (a) is a block diagram showing a configuration of an embodiment of a learning type signal recording / reproducing apparatus of the present invention, and FIG. 1 (b) is a case of recording a signal by the learning type signal recording / reproducing apparatus of the present invention. It is a figure which shows the waveform of a recording signal.

FIG. 2 is a diagram showing an embodiment of a sequence signal extracting section of the learning type signal recording / reproducing apparatus of the present invention.

FIG. 3 is a diagram showing an embodiment of a learning type waveform converting section of the learning type signal recording / reproducing apparatus of the present invention.

FIG. 4 is a diagram showing a first embodiment of a unit recognition unit used in the learning-type waveform conversion section of FIG.

5 is a diagram showing a second embodiment of the unit recognition unit used in the learning-type waveform conversion section of FIG.

FIG. 6 is a diagram showing a third embodiment of a unit recognition unit used in the learning type waveform conversion section of FIG.

FIG. 7 is a diagram showing a fourth embodiment of the unit recognition unit used in the learning-type waveform conversion section of FIG.

FIG. 8 is a block diagram showing a configuration of a conventional signal recording / reproducing apparatus.

FIG. 9 is a waveform diagram showing how an input waveform signal of a conventional signal recording / reproducing device is subjected to threshold processing.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 sequence signal extraction part 2 learning type waveform conversion part 3 teacher signal generator 4 non-volatile memory part 5 conversion result comparator 6 recording medium 7 signal detector 11 current signal extraction part 12 change amount extraction part 20 output result determination part 21 memory Control unit

Claims (17)

[Claims] 1. A signal detecting means for detecting a teacher signal and an information signal recorded on a recording medium, and the teacher signal and the information signal detected by the signal detecting means are respectively converted into predetermined conversion rules. A learning-type signal recording / reproducing apparatus comprising: a conversion means for converting a reproduction teacher signal and an information reproduction signal based on the learning signal; and a teacher signal generating means for generating a reference teacher signal, wherein the conversion means is the information signal. Prior to the detection of the learning signal recording unit, a learning type signal converting unit that detects the teacher signal and generates the predetermined conversion rule based on the reproduction teacher signal and the reference teacher signal. Playback device. 2. The learning-type signal recording / reproducing apparatus according to claim 1, wherein the converting means has a memory means for storing the predetermined conversion rule generated by the learning-type waveform converting section. 3. The sequence signal extraction unit, wherein the conversion means extracts a change in the signal value at each point on the transition axis of the detected teacher signal, and outputs the extracted change in the signal value as a sequence signal. And a memory control unit that stores the predetermined conversion rule in the memory unit, and the learning-type waveform conversion unit has a relationship between a change in the extracted signal value and a waveform of the reference teacher signal. The learning-type signal recording / reproducing apparatus according to claim 2, wherein the predetermined conversion rule is automatically generated by learning. 4. The learning type waveform conversion section, when another conversion rule is already stored in the memory means,
The method according to claim 2, wherein the detected teacher signal is converted into the reproduced teacher signal according to the other conversion rule, and the predetermined conversion rule is generated based on the reproduced teacher signal and the reference teacher signal. Signal reproduction device. 5. The learning type waveform conversion section, when another conversion rule is already stored in the memory means,
The memory control unit loads the other conversion rule, converts the detected teacher signal into the reproduction teacher signal according to the other conversion rule, and based on the reproduction teacher signal and the reference teacher signal, The signal reproduction device according to claim 3, which generates a predetermined conversion rule. 6. The converting means further comprises a comparing means for comparing the reproduction teacher signal and the reference teacher signal, and the memory means substantially stores the reproduction teacher signal and the reference teacher signal. 5. The signal reproducing apparatus according to claim 2, wherein the predetermined conversion rule is stored when the above-mentioned predetermined conversion rule is satisfied. 7. The learning type waveform conversion section has a network having a plurality of layers, and at least one of the plurality of layers has a signal input section for inputting the sequence signal, A quantizer for receiving the sequence signal from a signal input unit and quantizing the sequence signal; at least one first path input terminal; and at least one connected to the at least one first path input terminal Changing the connection between one first path output terminal, the at least one first path input terminal and the at least one first path output terminal according to the quantized sequence signal. A plurality of first route selection units for selecting a route according to
A unit recognition unit of the plurality of layers, the layer immediately before the output layer of the plurality of layers, a teacher signal input terminal for inputting the reference teacher signal, at least one second path input terminal, Of at least one second path input terminal connected to the at least one second path input terminal, at least one second path input terminal, and at least one second path output terminal A plurality of second unit recognition units each having a path learner that changes the strength of the connection with the one pointed by the reference teacher signal of, and the learning-type waveform conversion unit includes: The learning type according to claim 3, further comprising an output result determination unit that detects a unit recognition unit having a largest output value among the plurality of third unit recognition units included in the output layer of the plurality of layers. No. recording and reproducing apparatus. 8. The sequence signal extracting section extracts a signal value of a detected point of the teacher signal at a predetermined point on the transition axis, and the predetermined signal extracted by the current signal extracting section. The learning-type signal recording according to claim 3 or 7, further comprising: a change amount extraction unit that calculates a difference between the signal value of the point and the signal value of the teacher signal before the predetermined point. Playback device. 9. The learning type waveform conversion unit has a network of a plurality of layers each including a plurality of units, and the strength of the connection between adjacent layers of the plurality of layers is determined by the reproduction teacher. The learning-type signal recording / reproducing apparatus according to claim 1, wherein the predetermined conversion rule is generated by changing the learning rule using a signal and the reference teacher signal. 10. The learning-type waveform conversion unit has a network of a plurality of layers each including a plurality of units, and the reproduction teacher determines a connection method between adjacent layers of the plurality of layers. The learning-type signal recording / reproducing device according to claim 2, wherein the predetermined conversion rule is generated by changing the learning rule using a signal and the reference teacher signal. 11. The learning-type signal recording / reproducing apparatus according to claim 10, wherein the predetermined conversion rule is stored in the memory means as a method of connecting the adjacent layers. 12. The method of connecting between the adjacent layers is performed by changing the strength of each connection between a unit included in one of the adjacent layers and a unit included in the other of the adjacent layers. Modified, claim 10 or 1
1. The learning type signal recording / reproducing apparatus described in 1. 13. The multi-layer network has first, second, third and fourth layers arranged in this order, the first layer providing the detected teacher signals. Input, the fourth layer outputs the reproduced teacher signal, and the reference teacher signal is provided to a unit included in the third layer,
As a result, the unit and the fourth unit included in the third layer are
The strength of the connection with the units contained in the layer of is changed,
The learning-type signal recording / reproducing apparatus according to claim 12. 14. The strength of the connection between the unit included in the third layer and the unit included in the fourth layer is
14. The learning type signal recording / reproducing apparatus according to claim 13, wherein the reproduction teacher signal output from the fourth layer and the reference teacher signal are changed so as to substantially coincide with each other. 15. The learning type signal recording / reproducing apparatus according to claim 1, wherein when recording data on a recording medium, the reference teacher signal generated by the teacher signal generating means is recorded prior to the recording of the data. . 16. A method of connecting the at least one second path input terminal and the at least one second path output terminal in the plurality of second unit recognition units of the learning type waveform conversion section, The learning type signal recording / reproducing apparatus according to claim 7, wherein the predetermined conversion rule is stored in the memory means. 17. The sequence signal extraction unit generates another sequence signal by extracting a change in the signal value at each point on the transition axis of the detected information signal, and the information signal is detected. Sometimes, the memory control unit of the learning-type waveform conversion unit loads the predetermined conversion rule stored in the memory unit, and the plurality of first conversion rules are loaded.
Unit recognition unit, the signal input unit inputs the other sequence signal, the quantizer quantizes the other sequence signal, and the route selection unit inputs the at least one first route input. The connection between the terminal and the at least one first path output terminal is changed according to the quantized other series signal, and in the plurality of second unit recognition units, the at least one second unit recognition unit. The path input terminal and the at least one second path output terminal are connected in a connection manner indicated by the predetermined conversion rule, thereby converting the information signal into the reproduction information signal according to the predetermined conversion rule. The learning type signal recording / reproducing apparatus according to claim 16.