JP2861318B2 - Data detection method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention will be described in the following order.

A. Industrial application fields B. Summary of the invention C. Conventional technology D. Problems to be solved by the invention E. Means to solve the problems F. Function G. Embodiment G1 Configuration of one embodiment (No. (FIGS. 1 and 2) G2 Operation of one embodiment (FIGS. 1 to 8) H. Effects of the Invention A. Industrial Application Field The present invention relates to a data detection method suitable for high-density recording. .

B. Summary of the Invention The present invention relates to a data detection method for detecting conversion data from a reproduction signal of a medium on which conversion data in units of N bits is recorded, and a medium reproduction system for allowing n intersymbol interferences. , And the data to be detected is set in N-bit units. A pattern of a reproduction signal corresponding to the data to be detected is compared with a code pattern group corresponding to each converted data to determine an approximate pattern. In addition to the selection, the error rate at the time of detection is reduced by referring to the selected approximate pattern at the time of comparing the signal pattern of the subsequent reproduction signal with a plurality of predetermined patterns. Thus, it is possible to reliably detect converted data recorded at a remarkably high density.

C. Prior Art Conventionally, in digital magnetic recording, basically, binary values "1" and "0" of a digital signal correspond to the polarity of magnetization or the presence or absence of magnetization reversal. Then, according to the physical characteristics of the recording medium, the transmission band of the system, etc., binary coding is performed according to various modulation (writing) methods such as NRZ, FM, MFM, and GCR (8-10 conversion). Value and magnetization are 1
Correspondence is made appropriately in bit units.

In general, for these modulation methods, the minimum magnetization reversal interval (Tmin) is large and the maximum magnetization reversal interval (Tmax) is large.
It is required that the recording density is small, the change in the local recording density is small, and the DC component is small.

At the time of demodulation (reading), differential detection or integration detection is used in accordance with the DC component of the modulation code, and data is demodulated by detection in units of 1 bit.

In the conventional digital magnetic recording as described above, it is assumed that there is no intersymbol interference, and it is necessary that the level of a reproduced signal in a high frequency band is sufficiently high. In other words, the maximum repetition frequency of the recording signal and the density of the recording information are determined by the S / N of the high-frequency reproduction signal corresponding to Tmin of various modulation codes.

For this reason, as shown in FIG. 9, the current maximum repetition frequency f max is in the falling region of the reproduction level-frequency characteristic,
It is set in the area where high S / N can be obtained. In this descending region, the level of the reproduced signal falls at a gradient of, for example, 12 dB / oct due to various losses during recording and reproduction.

Further, an ideal transmission characteristic (frequency spectrum) as shown in FIG. 10A is required, and a sine-sloping equalization characteristic satisfying the first Nyquist criterion as shown in FIG. 10B is used.

In the figure, the frequency fo corresponds to the modulation code Tmin.

Also, the Nyquist's first criterion is that on the receiving side,
This is a condition that, when the waveform is sampled at regular intervals, sample values other than the center become 0.

For example, in the case of 8-10 conversion employed in current digital audio tape recorders (DATs) and the like, the relative speed between a coating type metal (MP) tape and a magnetic head is slightly more than 3 m / s, and the maximum repetition frequency f max Is set to 4.7 MHz, the gap length is 0.25 μm, and the recording wavelength is 0.67 μm. In this case, the track pitch is 10 to 15 μm and the linear recording density is 60
The limit is about 80kbpi.

D. Problems to be solved by the invention By the way, when recording at high density, the maximum repetition frequency
It is conceivable to increase f max.

However, if the recording density is to be increased, for example, by a factor of two, at a repetition frequency of 2f max, the level of the reproduced signal decreases as shown in FIG. There is a problem that it becomes impossible to detect an image.

As described above, current magnetic systems generally use recording media and conversion devices at their limits, reducing various losses during recording and playback, and significantly improving the level of playback signals in high frequencies. It is extremely difficult to do so.

On the other hand, when the recording density is increased, the following problem of intersymbol interference occurs in the reproduced waveform.

When one magnetization reversal exists in isolation on the magnetic medium, a pulse-like voltage waveform (isolated pulse) as shown in FIG. 11 is obtained as a reproduction signal. This isolated pulse, that is, the waveform of the impulse response is approximated to, for example, a lawrence type waveform as shown in the following equation (1), and the spread (pulse width) on the time axis is determined by the recording system / reproducing system used. Is determined by the total transmission characteristics of the
It is represented by a half-width Wh at the 50% level or a width Wb at the base level of substantially 0%.

f (t) = 1 / {1+ (t / to) ² } (1) When a plurality of magnetization reversals are continuously present at a constant interval,
If the recording density is low, there is no interference between adjacent pulses at the time of reproduction, and the reproduction signal is simply a series of the above-described isolated pulses alternately inverted.

The recording density has increased considerably, as shown in FIG.
The interval between adjacent pulses is equal to the pulse width Wb at the base level.
When the width is reduced to 1/2, the skirts of the pulses adjacent to each other overlap, and the waveform of the reproduced signal is considerably different from the isolated pulse.

As can be seen from the figure, at this stage, the peak value information of each pulse is stored without any loss, and although there is inter-waveform interference, no inter-symbol interference has occurred.

When the recording density is higher than the state shown in FIG. 12, the peak value of the reproduced signal decreases and nonlinear intersymbol interference (peak shift) occurs in which the interval between peak positions increases.

When the recording density further increases, for example, as shown in FIG. 13, when the interval between adjacent pulses is reduced to 1/4 of the pulse width Wb at the base level, the reproduced waveform of the pulse becomes closer to a sine wave, The peak value also drops significantly. Also in this case, intersymbol interference occurs that impairs the information of the peak value of each pulse.

Incidentally, a partial response (Partial Response) method is known as one that utilizes intersymbol interference.

In this system, for example,
As shown in FIG. 14, there is an advantage that the frequency spectrum is limited to the Nyquist bandwidth and high frequency components are not required.

The characteristics shown in FIG. 14 correspond to the class 4 of the partial response (modified duobinary).
It is expressed as the following equation (2).

Pr (1,0, −1) = sin (2πf / fo) ‥‥ (2) However, since the above-described various modulation codes themselves do not assume intersymbol interference, even if the partial response method is applied. However, there is a problem that the advantage cannot be fully utilized.

Also, for some modulation codes, Tmax is infinite,
Even if the partial response method is applied, there is a problem that functions necessary for a system configuration such as overwriting or clock reproduction cannot be realized.

In view of the above, an object of the present invention is to reduce the error rate at the time of detection while using the current recording medium and recording / reproducing device, and to reliably detect converted data recorded at high density. It is intended to provide a method for detecting possible data.

E. Means for Solving the Problems The present invention provides a data detection method for detecting converted data from a reproduction signal of a medium on which N-bit converted data is recorded, so as to allow n inter-symbol interferences. Set the transmission characteristics of the reproduction system of the medium and make the data to be detected in N-bit units. Compare the signal pattern of the reproduction signal corresponding to the data to be detected with a plurality of predetermined patterns limited by the converted data. Then, an approximate pattern similar to a signal pattern is selected from the plurality of predetermined patterns, and data is detected by referring to the selected approximate pattern when comparing a signal pattern of a subsequent reproduced signal with the plurality of predetermined patterns. This is a method of detecting data.

F. Function According to the present invention, the error rate at the time of detection is reduced while using the existing recording medium and recording / reproducing device, and the converted data recorded at a high density is reliably detected.

G. Embodiment An embodiment of a data detection method according to the present invention will be described below with reference to FIGS.

G1 Configuration of One Embodiment FIG. 1 shows the overall configuration of a magnetic system to which one embodiment of the present invention is applied, and FIG. 2 shows the configuration of the main part thereof.

In FIG. 1, reference numeral (10) denotes a recording system in which an analog signal such as an audio signal and a video signal from a terminal IN is
The data is supplied to a data generation circuit (12) via an A / D converter (11), and recording data conforming to a system format is generated. (13) is a data conversion circuit having a ROM table in which conversion codes as shown in Table 1 below are stored. The output of the generation circuit (12) is the data conversion circuit (RO
M) The symbol data supplied to (13) and output from this conversion circuit (13) is passed through a recording amplifier (14),
It is supplied to the magnetic head (1) and recorded directly on the magnetic tape MT.

A reproduction system (20) is a waveform equalization circuit comprising a signal (differential waveform) reproduced from a magnetic tape MT by a magnetic head (2) via a reproduction amplifier (21) and comprising an integrator and a reduction filter. (22). The output of the equalizing circuit (22) is supplied to an A / D converter (23), where the output is converted into, for example, 8-bit data. ) Includes a PLL as a synchronization signal
The output of the circuit (24) is provided.

(30) is a data detection circuit, the detailed configuration of which will be described later. The output of the A / D converter (23) is supplied to the data detection circuit (30), and the detected symbol data is supplied to the decoding circuit (24), decoded into the original data, and derived to the output terminal OUT. Is done.

In FIG. 2, a series of pattern data is supplied from an input terminal (30i) of the data detection circuit (30) to the synchronization detection circuit (3).
The signal is directly supplied to 1), and is commonly supplied to a plurality of subtracters (33a) to (33m) via a buffer (32). Reference numerals (34a) to (34m) denote reference value ROMs in which the waveform values of the respective code patterns selected as the reference are stored. The output of the synchronization detection circuit (31) is output from each ROM (34a) to
(34m), and the outputs of the ROMs (34a) to (34m) are supplied to the corresponding subtracters (33a) to (33m).

(35a) to (35m) are pattern distance calculation circuits, and (36) is a minimum value selection circuit. The outputs of subtracters (33a) to (33m) are supplied to the calculation circuits (35a) to (35m), respectively. And
The outputs of the arithmetic circuits (35a) to (35m) are supplied to the minimum value selection circuit (36), and the pattern data of the shortest distance to the waveform value of each code pattern is derived to the output terminal (30o).

Further, in this embodiment, the output of the minimum value selection circuit (36) is supplied to each of the reference value ROMs (34a) through (34) via the delay circuit (37).
(34m).

For code patterns and pattern distances,
This will be described in detail in the next section.

G2 Operation of One Embodiment Next, the operation of one embodiment of the present invention will be described with reference to FIGS.

In this embodiment, the relationship between the maximum recording repetition frequency fo and the pulse width Wb of the isolated pulse at the base level is set as in the following equation (3) for vector encoding as described below, and the pulse width is Wb becomes equal to four times the sampling period, and as shown in FIGS. 3 and 4, three sampling points are included in the pulse width Wb of each isolated pulse.

Wb = 1/4 · fo ‥‥ (3) In this specification, the above-mentioned state is referred to as a state where three inter-symbol interferences are allowed, and n states within the pulse width Wb of each isolated pulse. Is referred to as a state in which n intersymbol interferences are allowed.

In this embodiment, the data conversion circuit (RO) of the recording system (10) is used.
M) In (13), as shown in Table 1 below, 2-4 conversion is performed to convert original data in 2-bit units into converted data (modulation code) in 4-bit units.

The conversion data shown in Table 1 has the same number of "0" and "1" in the normal 16-bit 4-bit data pattern, and has [DC component] and [0]. Is selected.

First, assuming that the number of intersymbol interferences allowed according to the characteristics of the transmission path is n, a weighting coefficient distributed linearly decreasing from the center is prepared as shown in the following Table 2 and FIG. .

Next, using this weighting coefficient, weighting is performed on n consecutive codes among the N codes of the replacement data in which each conversion data “0” in Table 1 is replaced with “−1”. Done.

Weighting coefficients w31, w32, as shown in Table 2 and FIG.
Using w33 (w31 = w33), the four codes bi1 to b of the replacement data respectively corresponding to the ith and jth conversion data in Table 1
Of i4, bj1 to bj4, three consecutive codes are weighted.

The three weighted codes are added as shown in the following equation (4), and the first and second codes Ui1, Ui2; Uj1, Uj2 of the i, j-th intermediate sequence in Table 1 are respectively obtained. It is formed.

As is clear from Table 1, the intermediate sequences are different from each other, and the code distance Vij representing the degree of dissimilarity is expressed as the sum of the absolute values of the differences between the k-th codes of the pair of intermediate sequences Ui and Uj. , Are defined as in the following equation (5).

In this embodiment, conversion data in which the code distance Vij is twice or more the reference value for weighting is selected as the modulation code in Table 1. The reference value corresponds to the peak value of the impulse response according to the total transmission characteristics of the recording system (10) and the reproduction system (20).

On the other hand, in the reproduction system (20), the equalization characteristic of the waveform equalization circuit (22) corresponds to the partial response class 2 so as to cope with three intersymbol interferences. The following equation (6) and the one shown in FIG. 6 are selected.

Pr (1,2,1) = cos ² (πf / fo) ‥‥ (6) Thus, at the output of the waveform equalization circuit (22),
As is well known, the number of levels of the reproduced signal is five. Then, a reproduction waveform (code pattern) unique to the data pattern of each conversion data in Table 1 is obtained. This code pattern is as shown in FIG. 3 by itself, and is shown in FIG. 4 by continuous example. When the level of an isolated pulse at a sampling point is A,
At each sampling point in the detection period Td in which four isolated pulses coexist, [2A], [A], [0], [-A],
Each of the five values of [−2A] can be taken.

The pattern distance Dpq representing the degree of dissimilarity between the code patterns P and Q ‥‥ represents the reproduction output level at the sampling point corresponding to each k-th code of the pair of patterns P and Q by Spk and Sqk, respectively. Is defined as the following equation (7).

In the two consecutive code patterns exemplified in FIG. 4, the preceding code is the same but the succeeding code in the detection period Td is different, and the solid 1100/1100 code pattern and the broken line 11
In the code pattern of 00 · 0011, the pattern distance of the subsequent code is It is represented by

As shown in FIG. 7, in this embodiment, the pattern distance D between the four code patterns respectively corresponding to the conversion data in Table 1 corresponds to the recording system (10) and the reproduction system (20).
Of the peak value A of the impulse response of the transmission line composed of More than double.

In the data detection circuit (30) of FIG. 2, the code pattern corresponding to each of the conversion data as described above is selected as a reference and stored in each of the reference value ROMs (34a) to (34m). In the subtracters (33a) to (33m), this m (=
The 4) code patterns are compared with 4-bit input patterns from the buffer (32). On the basis of this comparison output, in the arithmetic circuits (35a) to (35m), the pattern distance between the input pattern and each code pattern is calculated in accordance with the above equation (7).

In the minimum value selection circuit (36), each arithmetic circuit (35
During the output of a) to (35m), one of the m code patterns and one pattern data at the shortest distance Dmin are selected, and the maximum likelihood (Maximum Likelyhood) detection data, that is,
It is led to the output terminal (30o) as the most likely detection data.

In this embodiment, the detection data of the previously determined preceding pattern is supplied to each reference value ROM via a delay circuit (37) having a delay time Td corresponding to an input pattern in units of 4 bits.
(34a) to (34m) are fed back, and the subsequent input pattern is detected with reference to the preceding pattern.

As exemplified in FIG. 4 described above, in the continuous reproduction waveform, the solid line 1100/1100 code pattern and the chain line 0011/1100
In the case where the preceding code is different as in the case of the above code pattern, the level of the succeeding code pattern greatly varies depending on the preceding code pattern even if the succeeding code in the detection period Td is the same.

In this embodiment, at the time of detecting the subsequent input pattern,
By referring to the preceding code pattern, the influence of the level fluctuation as described above can be removed, the succeeding code pattern can be reliably detected, and the error rate at the time of detection can be reduced.

In this embodiment, as shown by a solid line in FIG.
At a recording density of 120 kbpi, the symbol error rate is about 5 × 10 ^-4 , and as shown by the chain line in FIG.
Compared to T (8-10 conversion) having the same error rate at a recording density of just over 70 kbpi, the error rate at a high recording density is significantly improved.

H. Effects of the Invention As described above in detail, according to the present invention, in a data detection method for detecting conversion data from a reproduction signal of a medium on which conversion data of N bits is recorded, n inter-symbol interference The transmission characteristic of the reproduction system of the medium is set so as to allow the data to be detected, the data to be detected is set in N-bit units, the pattern of the reproduction signal corresponding to the data to be detected, and the code pattern group corresponding to each conversion data Compare with
Since the approximate pattern is selected and the selected approximate pattern is referred to when comparing the signal pattern of the subsequent reproduced signal with a plurality of predetermined patterns, the error rate at the time of detection is reduced, and the current medium is reduced. It is possible to obtain a data detection method capable of reliably detecting converted data recorded at a remarkably high density while using the recording / reproducing device.

[Brief description of the drawings]

FIG. 1 is a block diagram showing an entire configuration of an embodiment of a data conversion and detection method according to the present invention, FIG. 2 is a block diagram showing a configuration of a main part of an embodiment of the present invention, FIG. FIG. 4 is a waveform diagram for explaining the operation of one embodiment of the present invention, FIG. 5 is a diagram for explaining the operation of one embodiment of the present invention, and FIG. 6 is one embodiment of the present invention. 7 is a schematic diagram illustrating the operation of one embodiment of the present invention, and FIG. 8 is a schematic diagram illustrating the operation of one embodiment of the present invention. FIG. 9 is a diagram for explaining conventional digital magnetic recording, FIG. 10 is a diagram showing characteristics of the conventional example, and FIGS. 11, 12, and 13 are diagrams for explaining the present invention. FIG. 14 is a diagram showing characteristics of another conventional example. (10) is a recording system, (13) is a data conversion circuit (ROM), (2)
0) is a reproduction system, (22) is a waveform equalization circuit, (30) is a data detection circuit, (34a) to (34m) are reference value ROMs, (36) is a minimum value selection circuit, and (37) is a delay circuit. It is.

Claims (1)

(57) [Claims] 1. A data detection method for detecting conversion data from a reproduction signal of a medium on which conversion data in units of N bits is recorded, wherein a transmission of a reproduction system of the medium so as to allow n inter-symbol interferences. In addition to setting the characteristics, the data to be detected is set in N-bit units, and the signal pattern of the reproduction signal corresponding to the data to be detected is compared with a plurality of predetermined patterns limited by the converted data. An approximate pattern similar to the signal pattern is selected from the predetermined pattern, and data is detected by referring to the selected approximate pattern when comparing the subsequent signal pattern of the reproduced signal with the plurality of predetermined patterns. A method for detecting data, characterized in that: