JPH05266592A - Multivalue recording/reproducing device - Google Patents

Abstract

PURPOSE:To restrain the occurence of a phase error caused by the time axis fluctuation of the recording transmission route of a demodulation signal, to reduce the influence of the time axis fluctuation received on the recording transmission route at the time of detection and to improve a detection discrimination error rate. CONSTITUTION:An input binary data string is converted into blocks at converters 1, 2 and 3, the block data is transformed into PSK mutivalue data whose phase angle is moved linearly dividing PSK plural transmission pulses and after it is modulated at a PSK plural transmission pulses and is modulated at a PSK modulator 5, and is recorded on a recording medium. A reproducing signal from the recording medium is demodulated by a PSL demodulator 9, a pulse signal is detected, the PSK mutivalue data which moves from a detecting signal value linearly by the method of least squares is calculated and discriminated by a linear phase parameter discrimination section 13 and the discriminated mutivalue data is re-transformed into the original binary data by converters 14 and 15.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION The present invention relates to PSK (Phase Shift
BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a recording / reproducing apparatus for recording multi-valued data by using keying), and more particularly to a multi-valued recording / reproducing apparatus for a multi-valued digital signal which efficiently reduces a detection error due to time axis fluctuation from a reproduced signal and improves an error rate.

[0002]

2. Description of the Related Art Generally, a method of converting binary basis digital data into multivalued data and transmitting it by communication is widely used, and a high transmission speed is realized without widening a band. Further, it has been attempted to realize high-density recording by applying this multi-valued data transmission, which has been put to practical use in the communication field, to a magnetic recording medium or the like. For example, as the most effective method, an example in which multi-valued quadrature amplitude modulation (multi-valued QAM) is applied to magnetic recording, [“recording surface density 1 μm ² Examination of a digital VTR exceeding 1 / bit "1991 Annual Conference of the Television Society of Japan" and Japanese Patent Laid-Open No. 1-98167.

In the multi-level QAM system described in these documents, an input binary base digital signal is multi-valued in the amplitude direction, and then orthogonal carrier waves are modulated, combined and recorded. Further, at the time of demodulation, the carrier wave synchronized with the carrier wave at the time of recording is detected again, and if it is detected and determined, it is converted into the original binary base digital signal.

In particular, in the case of digital recording and reproduction of both the binary basis digital and the multi-valued digital, the re-sampling is performed accurately and at an appropriate time so as not to be affected by the fluctuation of the time axis of the recording / reproducing transmission line during reproduction. A detection process that can be realized is required. Therefore, the digital VTR or the like always records the time-series data with channel coating when recording. One of the purposes is to improve the accuracy of the self-clock function for obtaining an appropriate timing for resampling from the reproduced signal.

That is, in the case of binary basis data, the peak data of the reproduced signal and VFO (Variable Frequency Oscillator)
) PLL (Phased Locked) for position synchronization with the clock
Loop) is configured and generally generates a self-clock signal. Therefore, various coatings have been devised and put into practical use so as to avoid a state in which peaks are not frequently obtained, that is, a state in which the same data lasts as long as possible.

In such multi-level digital recording, the self-clocking function is of course necessary, and Q
Since it is accompanied by carrier wave modulation as represented by AM,
Synchronous detection is always required during demodulation. Therefore, a pilot signal for generating a carrier wave that is incidental to the recording signal is recorded at the same time.

[0007]

However, since the multi-valued QAM method for realizing the high-density recording described above is a digital recording for recording multi-level data, it has a high self-clocking ability and a high clocking ability when it is demodulated and detected. Accurate synchronous detection is absolutely necessary. How high the self-clocking ability is and how the synchronous detection can be realized without being influenced by the time axis fluctuation influences the reproduction accuracy of the recorded multilevel data signal, and as a result, the multilevel reproduction data is reproduced. It greatly affects the reduction of the error rate of the binary base data obtained by decoding.

In the recording / reproducing by the multi-valued QAM system, when the multi-valued data recorded at a constant time interval is affected by the time axis fluctuation in the recording transmission line, the reproduced signal has an appropriate multi-valued level. Since the time indicating
If detection is performed at regular time intervals as in the case of recording, the detected multi-value level becomes an erroneous value, resulting in an error rate.

Further, the carrier wave itself included in the reproduced signal also has a phase fluctuation due to the time base fluctuation of the recording and transmission path, and if the fixed frequency carrier wave is detected, the reproduced signal to be demodulated is adjusted to the time base fluctuation. Cause a phase error. Therefore, the signal before modulation cannot be completely demodulated, which deteriorates the error rate.

Similarly in the demodulation of the multilevel PSK system, if the self-clocking ability and the synchronous detection ability are low, it becomes difficult to effectively remove the influence of the time axis fluctuation from the reproduced signal, and the detected value has a large error. Therefore, in the multi-level PSK method, a method for effectively reducing the phase error of the demodulated signal due to the fluctuation of the time axis is required because it causes a judgment error and increases the error rate.

Therefore, the present invention, in the multi-level PSK system, suppresses the phase error due to the time base fluctuation of the recording transmission line of the demodulated signal, reduces the influence of the time base fluctuation on the recording transmission line at the time of detection, and makes a detection decision error. It is an object of the present invention to provide a multilevel recording / reproducing device that improves the rate.

[0012]

In order to achieve the above object, the present invention divides input binary data into predetermined bits,
Multi-valued data conversion means for converting into multi-valued data, a predetermined plurality of phase angles as a unit, the phase angle is a preset phase straight line that changes linearly with time, the number of multi-valued data,
Every time the input binary data of a predetermined bit is converted into multi-valued data by the multi-valued data conversion means, the converted multi-valued data is converted into a corresponding phase straight line, and the phase angle forming the phase straight line is output. A linear phase converting means, a modulating means for angularly modulating a carrier wave with a phase angle output from the linear phase converting means, a recording / reproducing means for recording and reproducing a signal output from the modulating means on a recording medium, Demodulation means for demodulating the reproduction signal reproduced by the recording / reproduction means,
Phase angle detection means for detecting the value of the phase angle from the signal demodulated by the demodulation means, and the predetermined plurality from the detected phase angle detected by the phase angle detection means as a unit,
Judgment selection for selecting a judgment value closest to the linear phase value set at the time of recording by comparing the phase straight line approximated by the calculation means with a phase straight line approximated by the calculation means and a preset phase straight line Means, a multivalued data conversion means for converting the phase straight line judged by the judgment selection means into corresponding multivalued data and outputting the multivalued data, and the multivalued data converted by the multivalued data conversion means into binary data. Provided is a multi-valued recording / reproducing device including a unit for converting and outputting.

[0013]

The multilevel recording / reproducing apparatus having the above-mentioned structure adopts the carrier system of the multilevel PSK system, and in one two-dimensional signal space, within a plurality of transmission pulses having a constant amplitude and a plurality of different phases. , A phase straight line that changes linearly with respect to the time axis is set in advance, and multivalued data is recorded in time series in correspondence with this phase straight line.

In reproduction, an approximate straight line of the phase with respect to the time axis is calculated from the detected value obtained by demodulating PSK using a least square method within a predetermined plurality of transmission pulses, and a preset phase straight line is obtained. The slopes and intercepts of the straight lines are referred to for comparison or threshold determination, and the specified phase straight line is converted into corresponding multi-valued data, and further converted into binary data and output.

Further, the multilevel recording / reproducing apparatus of the present invention reduces the demodulation phase error caused by the time base fluctuation of the recording transmission path, and the detected value of the detected phase straight line is the corresponding phase straight line because of the demodulation phase error. Are detected in a well-balanced manner across the line, and a plurality of detected values are linearly approximated by the method of least squares to obtain the phase straight line.

[0016]

Embodiments of the present invention will now be described in detail with reference to the drawings. FIG. 1 is a block diagram showing the configuration of a multilevel PSK system multilevel recording / reproducing apparatus according to a first example of the present invention.

In this multilevel recording / reproducing apparatus, an input binary data string which is a binary-based digital data string to be recorded is input to the serial / parallel converter 1.
It is divided into blocks by a predetermined number of bits. Here, the block data output from the serial / parallel converter 1 will be referred to as parallel multi-level data B ₁ hereinafter. This parallel multi-valued data B ₁ has the input 2 for each L bit.
If the value data string is divided, it can be represented by L bits 2 ^L
It can be regarded as multi-valued data of individual patterns.

The parallel multi-valued data B ₁ output from the serial / parallel converter 1 is input to the parallel multi-valued data / linear phase parameter conversion unit 2 and converted into a corresponding linear phase parameter. This linear phase parameter is usually 2 ^L that can be represented by the parallel multi-valued data B _1. 2 ^L in order to convert each multi-valued data in a one-to-one correspondence and without omission Individually prepared. Here, the linear phase parameter is a linear parameter that represents a phase transition that is preset so that the phase of a signal point in the two-dimensional signal space of multi-valued PSK linearly transits with respect to the time axis.

As is well known, the modulation method by multi-level PSK corresponds to multi-level data at signal points arranged in a two-dimensional signal space assumed for each transmission pulse of multi-level data and having a constant amplitude and different phases. Then, the coordinate component of the corresponding signal point is used as the transmission signal. The two coordinate components are transmitted by making the amplitude of the basic pulse proportional to the coordinate component value to form a pair of transmission pulses and amplitude-modulating orthogonal carrier waves into a combined signal.

FIG. 2 shows an example of a two-dimensional signal space of multilevel PSK. Here, four signal points represent four values, and two orthogonal axes are one in-phase axis I and the other orthogonal axis Q.
And The phase angle with respect to the I-axis is “positive” in the counterclockwise direction.
And the clockwise direction is defined as “negative” and is represented by “θ”. Conventionally, the signal points set and arranged in the two-dimensional signal space are uniformly determined, and the signal point arrangement does not change for each transmission pulse.

The linear phase parameter is a predetermined number of transmission pulses lined up on the time axis, and is the slope and intercept of the straight line when the phases of the signal points in the multilevel PSK system are arranged so as to line up on the line. Is. The phase of the signal points arranged in this way is not uniform in each signal space within a predetermined number of transmission pulses.

[0022] FIG. 3, the range of the time axis t = t _-2 ~t _2, expressed as one of the multi-level signal space by a set of five signal space arranged in the transmission pulse period interval, the linear phase parameter explain. In FIG. 3, the vertical axis represents the phase angle θ of the signal point that can be represented by multi-level PSK, and the horizontal axis represents the time t. First, linear phase data is set in the multilevel signal space. For example, what is shown in the figure is that θ = nα · t n = ± 1, ± 2, ± 3, ± 4 (1) set on the multilevel signal space.

A straight line represented by is shown. in this case,
The parameter n is a linear phase parameter. Also with this straight line, when the set of values of the phase angle at time t = t _-2 ~t ₂ called linear phase block data, linear phase block data S _n, which is defined on the multi-level signal space, this In the example, S _n = [nαt ₋₂ , nαt ₋₁ , nαt ₀ , nαt ₁ , nαt ₂ ] ... (2) Expanding further, the linear phase data is set in the above-mentioned multi-valued signal space, θ = nα · t + mβ where n and m are integers, α is the slope of the straight line, and β is the straight line of the phase angle of t _0. phase parameter, n, points to the m, the phase angle data words linear phase block data at time t = t _-2 ~t ₂ is

[0024]

[Equation 1] Can be expressed as

The number of messages that can be represented by the linear phase block data S _{n, m in the} multilevel signal space set in this way is (n × m), and the linear phase block data S
The pattern of _{n and m} is the number of messages that can be represented by the block data converted by the parallel / serial converter 1 2 ^L Must be set equal to. That is, n × m = 2 ^L … (5)

[0026] The linear phase parameters n and m of the straight line defined above are integers, but any constant can be selected within a range in which the error rate when demodulated in this method satisfies a desired value.

The parameters n and m converted and output by the parallel multi-valued data / linear phase parameter conversion unit 2
Is converted by the linear phase parameter / linear phase block data conversion unit 3, and the component of the linear phase block data S _{n, m} determines a signal point on the PSK signal space for each transmission cycle, and the orthogonal axis component I , Q are output. The series of processes described above is as shown in FIG.

The output from the linear phase parameter / linear phase block data converter 3 is input to the PSK modulator 5 via the hold circuits 4a and 4b that hold the transmission pulse period.

In the PSK modulator 5, the two-channel baseband signals, the I-axis component signal and the Q-axis component signal output from the hold circuits 4a and 4b are cut above the band of the transmission line to obtain the I-axis component signal. each multiplied by the sin .omega _c t to Q-axis component signal carrier cos .omega _c t the same carrier with different [pi / 2 phase, summed and input to the transmission path 7 via the recording amplifier 6.

Further, the reproduction signal output from the transmission line 7 is amplified to a predetermined amplitude via the reproduction amplifier 8 and then P
It is input to the SK demodulator 9. In this PSK demodulator 9,
Splits the reproduced signal into two, a carrier cos .omega _c t, respectively multiplied by sin .omega _c t to Q-axis component signal of the same carrier wave by changing a [pi / 2 phase, I-axis component signal and the Q-axis is a baseband signal Separate into component signals. The I-axis component signal and the Q-axis component signal thus separated are equalized by the equalizers 10a and 10b.
Each is pulse shaped and output. Here, with reference to FIG. 5, the basic principle of recording / reproduction by the multi-level PSK method will be described.

First, the recording signal converted into the multilevel signal is the time series signals d _i (t) and d _q (t) of the I axis component and the Q axis component of the multilevel QAM signal space and the basic pulse. rectangular pulse p of the pulse width T _B (t) is convolved, since a signal obtained by that is multiplied with further cos .omega _c t and sin .omega _c t synthesized, expressed by the following equation. If this signal is m (t),

[0032]

[Equation 2] Becomes

The reproduced signal depends on the frequency transfer characteristic of the transmission line, but in order to briefly explain the basic principle, m (t)
Is assumed to be reproduced by receiving only the linear deformation of the basic pulse on the transmission line. Furthermore, m (t) is the time axis fluctuation ψ in the transmission line.
If (t) is received, m (t) reproduced will be

[0034]

[Equation 3] It is represented by.

Then, the reproduction signal m (t + ψ (t)) which has received the time axis fluctuation ψ (t) is next subjected to quadrature detection processing. The reproduction signal m (t + ψ (t) ) is, cos .omega _c t
, And sin ω _c t are multiplied together, pass through the LPFs respectively, and are generally subjected to synchronous quadrature detection processing to remove high frequency components.

The signals separated into two by the quadrature detection process are input to the equalization filter and pulse-shaped. By this pulse shaping processing, the reproduction signal m
The basic pulse generating (t) is a pulse h that satisfies the Nyquist first criterion of zero intersymbol interference from the preceding rectangular pulse.
(T). The two signals after the pulse shaping process become predetermined baseband signals, and these are I _d (t + ψ
(T)) and Q _d (t + ψ (t)),

[0037]

[Equation 4] It is represented by. Here, it should be noted that I _d (t) and Q _d (t) cause so-called fluctuations in time due to time-axis fluctuation ψ (t), which results in asynchronous quadrature detection that changes with time. Further, the time axis fluctuation ψ (t) is the reproduction signal I _d (t), Q.
_Make important assumptions about the impact on _d (t).

First, since the pulse width of the shaping basic pulse h (t) is sufficiently smaller than the period of the time axis fluctuation ψ (t), the time axis fluctuation ψ is within the time of the pulse width in which one basic pulse is defined. It is approximated that (t) is almost constant. Therefore, in the subsequent development of the formula, attention is paid to an arbitrary basic pulse width on the time axis in the reproduction signal m (t). According to the equation (1), k: an integer,

[0039]

[Equation 5] Assuming that the time axis fluctuation ψ (t) at is constant,

[0040]

[Equation 6] And I _d (t + ψ (t)), Q _d (t + ψ
(T) can be approximated and expressed as follows.

[0041]

[Equation 7]

When I _dk (t) and Q _dk (t) are sampled by the self-clock, the above equation (12) is used.
Then, it can be expressed by setting t = t′−ψ _k and further rearranging t = t ′. Therefore, _X'k and _Y'k are

[0043]

[Equation 8] Therefore, when X ′ _k in equation (8) is _replaced with X _k and Y ′ _k is _replaced with Y _k ,

[0044]

[Equation 9] Becomes Further, the waveform-shaped h (t + ψ _k ) is also
In the same way

[0045]

[Equation 10] And is resampled at the proper time, ie, t = kT _B. Since h (kT _B ) = h (0) (16), the detected sampling values I _k and Q _k can be calculated by the equations (14) and (16).

[0046]

[Equation 11] Given in. If h (0) = 1, then equation (17) becomes

[0047]

[Equation 12] Becomes

The I _k and Q _k thus detected represent the I-axis component and the Q-axis component which are the coordinates of the detection signal point on the k-th two-dimensional signal space. When this two-dimensional signal space is treated as a general complex plane, the k-th detection value m _k on the time axis obtained by asynchronous quadrature detection is

[0049]

[Equation 13] Expressed as Complete synchronous detection is performed, if there is no time base _{fluctuation, ζ k = 0 ... (20} ) So, if similarly represented recording signal points, which as a _{s k, s k = (x} k / 2 ) + J (Y _k / 2) (21) That is, the _k -th detection value m _k obtained by the quadrature detection asynchronously is m _k = m _k · exp [−jζ _k } (22)

When the time axis fluctuation ζ _k is received, the detection signal point represented by the detection value m _k demodulated by this method is the phase angle ζ _{k with} respect to the recording signal point when the time axis fluctuation is not received.
Produces an error phase angle of The recording signal point s _k (a _k d _k ) is deviated by the error phase angle according to the time-axis fluctuation amount, and the detection signal point m
It will be the _{k (a 'k b' k} ).

Next, the linear phase parameter / linear phase block data conversion unit 3 is converted from the demodulated signal subjected to the phase fluctuation.
A method of detecting and determining the linear phase block data converted in step 1 will be described. The reproduction signal of the preceding linear phase block data obtained by conversion at the time of linear phase recording is subjected to the above-mentioned fluctuation of the phase angle. FIG. 6 shows how the linear phase block data is reproduced. The reproduced linear phase block data is detected by phase fluctuation dispersion from the phase straight line at the time of recording. A linear approximation formula is obtained from the detected phase values obtained by dispersion using the method of least squares. If the obtained linear approximation formula is θ = n′α · y + m′β (23) where n ′, m ′ are integers, α is the slope of the straight line, and β is the phase angle of t _o , it is detected. The linear phase parameter is
n ', m' indicates the phase angle data words linear phase block data at time t = t _-2 ~t ₂ is

[0052]

[Equation 14] Indicated by.

As the linear phase parameters n'and m'detected in this way, those which are determined by a predetermined threshold value determination and are closest to the values of the linear phase parameters n and m set at the time of recording are selected. It becomes a judgment value. The linear phase parameter is converted into parallel multilevel data by the inverse conversion of the conversion performed in the parallel multilevel data / linear phase parameter conversion unit 2, and the original input binary data is restored by parallel / serial conversion. Output binary data.

The reproduced linear phase block data is dispersedly detected from the phase straight line at the time of recording by the phase fluctuation amount. However, the detected phase value obtained by dispersion is within the range of the block data B _i , and the error is If the lines are well-balanced and randomized and distributed, the lines obtained by the method of least squares have few errors. When the linear phase block data is captured in a signal space having the number of dimensions (five dimensions in this embodiment) of the number of constituent elements, the data of the phase that changes linearly is an arbitrary direction in this signal space. It is a signal point vector indicating one point on a straight line having a vector. The slope of the phase straight line corresponds to the signal point position on the straight line in the signal space. For example, the signal points that can be represented by the five phase angle data of the detected block data of this embodiment are:

[0055]

[Equation 15] However, the direction vector d ₀ = (1,1,1,1,1) and the direction vector d = (− 2, −1,0,1,2).

The first vector of the equation (25) is a slant of the detected signal point vector S'n _{, m} onto the straight line of the direction vector d _o , and the next vector is the detected signal vector S '.
It is a slant on a straight line of the direction vector d of _{n and m} . Correct signal point vector S _{n, m} and detected signal point vector S ′
_{When the} error vector S'n _{, m-} Sn _{, m} representing the error of _{n, m} is shaded on the straight line of each signal space, if the error vector does not deviate in the direction of the direction vector, that is, the phase straight line is If the detected values are distributed in a well-balanced manner as the center, the probability of occurrence of the error vector indicating the same direction as each of the above-described direction vectors becomes small.

It can be said that the error vector, which is caused by the time-axis fluctuation of the transmission line to be solved by the present invention, has a well-balanced dispersion as referred to in the present embodiment when observed in the range of the cycle of the time-axis fluctuation or more. Therefore, the determination probability is improved by determining the shaded error vector on the straight line of the signal space. Returning to FIG. 1, the equalizers 10a and 10b.
The following configuration will be described.

The I-axis component signal and the Q-axis component signal which are pulse-shaped and output by the equalizers 10a and 10b are respectively branched into two, one of which is input to the clock recovery circuit 11 and the other of which is a linear phase parameter. Input to the detection unit 12.

The clock reproduction circuit 11 extracts a clock signal from the reproduced I-axis component signal and Q-axis component signal, reproduces the clock signal, and outputs the clock signal. The linear phase parameter detection unit 12 receives the reproduced I-axis component signal and Q-axis component signal to detect the phase block data first, and then detects the linear phase parameter n ′,
Calculate m'using the least squares method.

Then, the calculated linear phase parameters n'and m'are judged by the linear phase parameter judging unit 13 and n and m are output. N and m output from the linear phase parameter determination unit 13 are linear phase parameters /
It is input to the parallel multi-valued data conversion unit 14 and performs the inverse conversion of the conversion process performed with the parallel multi-valued data / linear phase parameter 2 at the time of recording.

The parallel multi-valued data output from the linear phase parameter / parallel multi-valued data conversion unit 14 is
Since this is binary block data, this is converted into serial binary time-series data by the parallel / serial data converter 15, and output binary data is output.

In this embodiment, when the input binary data is divided into blocks, the continuous binary data is used as the block data. However, since the period of time axis fluctuation is considerably larger than the transmission pulse period, the preset discrete data is set. May be one block data. By doing so, the dispersion of the error in the linear phase block of the above-mentioned phase straight line that fluctuates due to the time base fluctuation received on the transmission line becomes well balanced, and the detection error due to the time base fluctuation in this embodiment is efficiently reduced. This can be done well, and one block data need not be large.

As described above, the multilevel recording / reproducing apparatus of the present invention adopts the multilevel PSK carrier system, and in a two-dimensional signal space, a plurality of transmission pulses having a constant amplitude and a plurality of different phases. In this, a phase straight line that linearly changes with respect to the time axis can be set in advance, and multivalued data can be recorded in time series in correspondence with this phase straight line.

The signals at the time of recording are set signal points arranged in a predetermined two-dimensional signal space of multi-valued PSK. With respect to the set signal point, the detected detection signal point includes an error phase component that fluctuates according to the fluctuation of the time axis, and becomes an asynchronous detection in which the degree of out-of-sync changes with time.
If the reproduced signal does not fluctuate on the time axis at all, the phase difference between the carrier wave at the time of recording and that at the time of reproduction is constant at any time, so the error phase angle between the set signal point and the detected signal point vector is also It is constant with respect to the time axis.

However, it is not possible that the time axis fluctuation does not occur in the above-mentioned transmission path. Especially, in the recording / reproducing transmission path accompanied by the mechanical drive, the time axis fluctuation called "jitter" greatly affects the detection error. To do. Therefore, even if it is simply called asynchronous detection, the phase shift is not constant with respect to the time axis.

However, it is generally assumed that the average fluctuation width of the time base fluctuation is equal to zero, and the mean value of the error phase angle caused by the time base fluctuation can also be equal to zero. Therefore, if the signal point distances of the detection signal point and the setting signal point corresponding to this error phase angle are taken over a plurality of signal points and the average is taken, the setting signal point having the smallest value is the plurality of detection signal points compared. Can be said to have the highest probability close to. This is because the error phase angles due to time axis fluctuations are averaged and canceled out.

In the reproduction, the approximate straight line of the phase with respect to the time axis is calculated from the detected value obtained by demodulating the PSK by using the least square method in a predetermined plurality of transmission pulses, and the preset phase straight line is obtained. The slopes and intercepts of the straight lines are referred to for comparison or threshold determination, and the specified phase straight line is converted into corresponding multi-valued data, and further converted into binary data and output.

Therefore, the multilevel recording / reproducing apparatus of the present invention reduces the demodulation phase error caused by the time axis fluctuation of the recording transmission path, and the detected value of the detected phase straight line corresponds to the demodulation phase error. This phase straight line is obtained by linearly approximating these plural detection values by the least squares method, which is detected with a good balance in a straight line. Further, the present invention is not limited to the above-described embodiments, and it goes without saying that various modifications and applications can be made without departing from the scope of the invention.

[0069]

As described above in detail, according to the present invention, a multi-valued PSK system is used by using a transmission line such as a magnetic recording / reproducing system or an optical disk recording / reproducing system which is greatly affected by time axis fluctuations. When recording, the detection error due to loss of synchronization at the time of demodulation due to the influence of time axis fluctuations on the transmission path during playback is reduced, and the detection error rate is effectively reduced by accurately determining playback multi-level data. Thus, it is possible to provide a multilevel recording / reproducing apparatus which realizes high density recording / reproducing while maintaining an error rate that does not hinder practical use.

[Brief description of drawings]

FIG. 1 is a block diagram showing the configuration of a multilevel recording / reproducing apparatus as a first embodiment according to the present invention.

FIG. 2 is a diagram showing definition of a phase angle of a signal point in a PSK signal space in the first embodiment.

FIG. 3 is a diagram showing an example of setting of linear phase data performed in the first embodiment.

FIG. 4 is a diagram for explaining a recording principle of linear phase modulation.

FIG. 5 is a diagram showing a flow of a signal reproduced by PSK and explaining the principle of phase error generation by asynchronous detection.

FIG. 6 is a diagram for explaining the reproduction principle of linear phase modulation.

[Explanation of symbols]

1 ... Serial / parallel converter, 2 ... Parallel multi-valued data / linear phase parameter converter, 3 ... Linear phase parameter / linear block data converter, 4a, 4b ... Hold circuit, 5 ... PSK modulator, 6 ... Recording Amplifier, 7 ... Recording / reproducing transmission line, 8 ... Reproducing amplifier, 9 ... PSK demodulator, 1
0a, 10b ... Equalizer, 11 ... Clock recovery circuit, 12
... Linear phase parameter detection unit, 13 ... Linear phase parameter determination unit, 14 ... Linear phase parameter / parallel multi-valued data conversion unit, 15 ... Parallel / serial conversion unit.

Claims (1)

[Claims] 1. A multi-valued data conversion means for dividing input binary data into predetermined values and converting it into multi-valued data, and the phase angle linearly changes with time in units of a predetermined plurality of phase angles. Phase lines corresponding to the number of multi-valued data are set in advance, and every time the input binary data of a predetermined bit is converted into multi-valued data by the multi-valued data conversion means, the converted multi-valued data is associated with the phase. A linear phase conversion means for converting into a straight line and outputting a phase angle forming a phase straight line, a modulation means for angularly modulating a carrier wave with a phase angle output from the linear phase conversion means, and a signal output from the modulation means On a recording medium,
Recording / reproducing means for reproducing, demodulating means for demodulating the reproduced signal reproduced by the recording / reproducing means, phase angle detecting means for detecting a phase angle value from the signal demodulated by the demodulating means, and the phase angle detecting means Computation means for linearly approximating the phase in units of the predetermined plurality from the detected phase angle detected by the means, and comparing the phase straight line linearly approximated by the computation means with a preset phase straight line, and recording Judgment selection means for selecting a judgment value closest to the linear phase value set at time, multi-value data conversion means for converting and outputting the phase straight line judged by the judgment selection means to corresponding multi-valued data, A multi-valued recording / reproducing apparatus comprising: a unit for converting multi-valued data converted by the multi-valued data conversion unit into binary data and outputting the binary data.