JPS60201728A - Coding system of binary data - Google Patents

Abstract

PURPOSE:To obtain a DC free code having a small range of the cumulative charge variance DSV and large suppressing effect of DC variance, by performing an operation of attain DC free every code, i.e., every 10 bits. CONSTITUTION:The integrated value is set at 0 with P set at 0 with a waveform pattern in case H and L levels are set at +1 and -1 respectively for a waveform subjected to underwent NRZI conversion. Then the 8-bit data is converted to a 10-bit code via a ROM table 31 to calculate a coded word digital sum CDS. A logical circuit 43 inverts (non-inverts) the head bit W1 of the 10-bit code W as long as P is set at 0 (1) and the codes are equal to each other between DSV and CDS. Thus DSV=DSV-CDS is satisfied. If the DSV and CDS have different codes, the bit W1 is non-inverted (inverted) to satisfy DSV=DSV+CDS. Then an exclusive OR is obtained between P and the P of the code W' obtained after said inversion (non-inversion).

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention provides an electronic device such as a magnetic disk, an optical disk, etc., when recording binary data on a recording medium or reproducing it from a recording medium. Encoding method of binary data to be converted into code sequence and/or
Or it relates to an electronic device having the encoding method.

Conventionally, in electronic devices such as magnetic disks and optical disks, it is necessary to record a large amount of information, and when recording binary data on a recording medium, it is essential to improve the recording density.

Generally, the encoding method converts m bits of original data to wI&
The number of bits '0' between adjacent bits @1' is converted into an n-bit code limited to a minimum of d and a maximum of n bits, and this code is subjected to NRZ conversion to become a recording waveform pattern. In other words, the recording waveform pattern corresponds to the code bit ""1m inverted and the code bit 10" not inverted.The encoding method is generally (nl +
It is expressed by four parameters: ne 6 e k).

Next, terms using the above parameters will be explained.

T: Data bit interval Tminw-((1+1)) T: Minimum inversion interval Tma
x = -(k-N)T: Maximum reversal interval Tw = -T
:Detection window width (demodulation phase margin) D8V (Digital
8nm Variation, cumulative charge variation): H1gh1@17 of the recording waveform pattern whose sign has been converted to NRZ
Integral value for the recording waveform pattern when @II is +I Low 111761 is -1.

When Dfff can increase without limit, its sign has a DC component. When the variation range of D8V is finite, DO79- is obtained.

CD8 (Ooaewora Digital 8nm)
= nov from beginning to end of one code.

Note that the following description will be made using the above symbols.

In addition, as conventional encoding methods, (1) AoM, Fatal, @Encoder ana
I) Ecoaer for a Dyte-Or 1
entea (0+5)e, "9 QO'le"
+M Technical Disclosure
Bulletin, Vol.B, No.I June
1975 (p248) (2) A+M+Patel, @Oharge-oon
strainea Byte -0rientea (
0,5) Cods”, more M TechnicalDi
Closure Bulletin, Vol. 1? ,
No, 7. December 1974 (p2715) is mentioned. (1) is the (8,9,0,!1) encoding method, (2) is the (1) (a, ?, o, 3) code that is concatenated and then the connected bits are 2 bits. It is a DC-free code by adding, and as a result, DC
It is a 7 Lee (a, 1o, o*a) code.

In method (1), Tm1n = −; T = 0.89
TTmax = −X 4 T = 5.56 T(
#iTmin = t o T = 0.8 using method 2)
"Tmax = -X 4 T = 5.2 T0 and DO7 Lee. However, in this method, after concatenating two codes, after 18 bits, and DC
D because the connection bit to make it 7jj- cannot be added.
Although the fluctuation range of 8V is limited, it becomes large.

The important thing to note about the encoding method is that T
Regarding m1n, it is better for Twin to have a thicker h for rounding that does not include high frequency components and is less susceptible to band limitations. Further, Tw is preferably large so that pulses can be distinguished and K<< does not hold.

Also, make Tmax as small as possible, 'rmin and Tm!
In order to make synchronization easier by reducing the difference in L3C and to reduce low frequency components, the smaller Tmax is, the better b6 or more As explained above, the present invention has a small Tmax (k = 2 (compared to the conventional)) and extremely synchronization. Gatoshasuku.

There are few low frequency components and both Tm1n and Ty are 0. B
Encoding system with large value of T and DO using this
It is an object of the present invention to provide a 7-Lee encoding system, and further to provide electronic equipment such as magnetic disks and optical disks having the above-mentioned encoding system.

The present invention will be described in detail below with reference to the drawings.

FIG. 1 is a block diagram of an electronic device that performs digital modulation of magnetic disks, optical disks, etc. lF!
2 is an information source encoding unit for suppressing redundancy of the information 1rA1. Note that band compression compresses the transmission frequency band in an analog way, and high-efficiency encoding is a digital way to compress the transmission frequency band.
It attempts to reduce the average number of bits per element (sample value), and in that sense it is similar to amplitude compression.r and s are the communication channel and transmission channel channel coding section, and include correction, digital modulation, etc. It will be done. , 4 is a recording/reproducing system for the above-mentioned magnetic disk, optical disk, etc. Further, 5 and 6 are decoding units for decoding the encoded data in the encoding unit 2.3. 7 is an output unit that outputs the information obtained through the above processing.

The igZ diagram is a block diagram showing an example of the recording/reproducing thread 401, and is an example applied to a video disc.

First, let's talk about the signal Q system. For example, a semiconductor laser emits light in a blinking manner based on a drive signal from a signal source 8 based on human data. In addition.

The signal source 8 includes the encoders 2 and 3 shown in FIG. The luminous flux emitted by light #9 passes through collimator lens 1
0] Parallel light beam and grating 11. The polarizing plate 12 passes through an optical probe 13 whose transmission reflectance is polarization dependent. A point image is formed on the perpendicular magnetic recording medium 15 by the objective lens 14 . The semiconductor laser light is approximately P-polarized with respect to the optical probe 13, and the polarizing plate 12 is also installed with the polarization direction in the P direction.

Grating j11i) 9 A beam angle separation is performed to focus a sub-spot for detection on the perpendicular magnetic recording medium 15 using an objective lens 141C.

During the operation of the grating 110, three point images are formed on the recording medium 15. Of these five point images, two point images used for tracking signal detection during reproduction are ±1st-order diffracted light of the grating 11, and the remaining one is undiffracted light (zero-order light). Depending on the setting of the diffraction efficiency of the grating 11, it is easy to record signals using only the point images of undiffracted light without recording signals using these two point images.

The combination of the cylindrical lens 16 and the 4-split detector 17 is 1.
This is to obtain an autofocus signal for adjusting the position of the objective lens 14 in order to focus the point image correctly.

The signal from the 4-split detector 17 is divided into two systems by the signal splitter 1B, one for the autofocus signal and one for the F
i Used for recording signal output and monitoring. Note that this output is sent to the decoding section 5.6 described in FIG. It includes an information output section 7.

Also, during recording, a tracking signal detection detector 19.
The differential signal from 20 is in the OFF state.

Next, the signal reproduction system will be described.

A signal of a constant level is applied from the signal insert 8 to bring the light source 9 into a constant light emitting state. Further, the amount of light at this time is adjusted to about i so that the recorded magnetic domain pattern is not reversed as described above. The light beam transmitted through the collimator 10, grating 11, polarizing plate 12, and optical element 13 is sent to the objective lens 1.
4. Focus three point images on the recording medium. The light beam from the recording medium 15 undergoes modulation of the polarization plane due to the Kerr effect, and is transmitted to a detector 17 by a system of a light splitting optical element 13 and an analyzer 21.
．． At 19.20, the light enters a modulated state of brightness and darkness. The signal from the detector 17 is distributed into two systems, one system serving as an autofocus signal and the other system serving as a reproduction signal.

In addition, the signals from the detectors 19 and 20 are differentiated by the differential AMP 22, and the objective lens is swung left and right using that signal.
Perform tracking.

Note that due to the action of the optical element 13, a bright and dark pattern with high contrast can be detected in the reproduction system.

Note that a 7 Alladay one-turn element can be inserted between the optical element 15 and the DD recording medium 15 as a means for adjusting the light amount eA between recording and reproduction.

A Faraday rotation element is made of, for example, Y-G (yttrium, iron, garnet) crystal or glass doped with rare earth elements, and can rotate the plane of polarization of multiple luminous fluxes by applying a magnetic field. can. The reason for using this Faraday rotating element is as follows.

The polarization direction of the reflected light from the 1lr30 body 15 during recording,
This is different from the polarization direction of reflected light that has undergone Kerr rotation during reproduction. Therefore, the reflected light beam is separated from the incident light beam by the light splitting optical element 13, and the amount of light transmitted through the analyzer 21 is different.

First, during reproduction, the amount of light emitted by the light source must be lower than during recording to prevent the recorded magnetic domain pattern from reversing, so this factor also causes the amount of light that passes through the analyzer 21 to differ between recording and reproduction. different.

If the light intensity of the light beam guided through the cylindrical lens 16 to the four-split detector 17 for detecting recording signals and autofocus signals is significantly different, it becomes necessary to change the sensitivity of the detector 17 between recording and reproduction. occurs.

7A 2-day one-rotation element rotates the polarization plane of the recording light beam by applying an appropriate magnetic field during recording.
The amount of light entering the detector 17 is adjusted by the combination of the detector 17 and the analyzer 21, thereby solving the above problem.

In this example, we talked about video discs, but
There is no need to limit the scope to these, but a workstation, a printer, a host computer, and an encoding system (corresponding to 1 and 2.5 in FIG. 1), which is one aspect of the present invention, will be explained next.

D., T., Tang and L., RoBahl, “Bl.
ock Codes for aClass of C
on5trained No1seless Ohan
tuslg”.

Information and Control, V
According to OL, 17,197G, p436, it has been proven that the number of limited codes of length n bits, where d=0 and k is a finite value, is the following N k (n).

Nk (n) = 2n (0<n≦k) The nine results calculated using this formula are shown in Table 1.

Table 1 From this t41 table, it can be seen that there are #'1504 codes where J) n=10 and *=2 (a=O). However, when these codes are concatenated, as shown in Figure 3, the =2 restriction may be violated at the junction between the codes. If possible, the restriction of =2 will not be violated even by concatenation of signs.

Figure 4 (a) Fi is a code in which the first bit is always 1, the last bit is 1, and the middle 8 bits are limited to 2. According to Table 1, there are 9149 such codes. In FIG. 4(b), the first bit is always 1, the last 2 bits are 10, and the middle 7 bits are limited to 20. There are 81 such items in Table 1. Figure 4 (0
) is a code whose last bit is always 1, and the last 3
If the number of bits is 100, the middle 6 bits are limited to 20. There are 44 of these in Table 1.

As described above, even if the connection is configured as shown in FIG. 4, there are 274 codes for the limit of 5 and 5 for the limit of 2.

Consider separating data into 8-bit units and converting this into a 10-bit code. Then, 8-bit data string 2
”=256 codes exist, which is smaller than the 274 10-bit codes shown in Figure 4. Therefore, 256 codes are randomly selected from among the 274 codes, and these are divided into 256 8-bit codes.
By making a one-to-one correspondence with bit data (m,
It can be seen that the code n, a, k) = (L 10+0t2) can be realized. Next, m4 diagram (a) l (b) l (e)
An example of specific creation of each code will be described. This may be done according to the state transition diagram shown in FIG. Here sl, s2. s3 has three states, and 81 is the initial state. The D in Dlo indicates a data bit, and 0 indicates a sign bit. Figure 6 shows the 4th
A method of creating a restriction code for 6 bits = 2 in Figure (Q) is shown. The same applies to FIG. 4(a) (t)) )ζ. FIG. 6 will be explained in detail. First, 6-bit data (
D,...D, ) are prepared. (D,...D6) is (0...o), (o...1), (o...10).
... and arrange them in increasing order by 1. 6-bit code (
0,...a, ). First (D,...D, )
The case when =(0...0) will be described. The state transition diagram in Figure 5.9D, ==0 and 0. =1 is generated. Next, D2shi 0
02=1 is generated, and thereafter D4=0 and Q=1 are generated in the same manner, thereby completing the creation of the first code. The same goes for the following, but when (D, . . . D4.) = (oooll), the last bit goes to 5top. This is (00011
1) means that a code cannot be created.

Thereafter, codes are created in the same way, and the process is repeated until 44 codes are created. and (D, ~D,)=(
110110) and (0,...04) = (001001
) and all 44 are extracted. Figure 4 (,)
Or in the case of (b), each data (D,...
D, ) (D,...D,) and sign (C,...
O, ) (C, . . . a, ) can be created. By the 274 connections extracted as above, k 1I
We randomly select 256 out of ll unbreakable codes and calculate 2
The (8, 10, 0, 2) encoding method can be realized by making one-to-one correspondence with 56 pieces of 8-bit data. As mentioned above, (m, n, a, k) = (a, 1
o, o, 2). The 256 codes selected are ROM
All you have to do is write it to xoolc up table and make it correspond to the data.

Next, an operation is performed to make the DO free for each code and every 10 bits of tfi, and D8V (Digita1
8um Variation) to reduce the variation range,
This is the so-called creation of a DO7jJ- code that has a large effect of suppressing DC fluctuations.

do clarification. Figure 7 is the first block diagram of the encoding configuration.
It is a block diagram of the structure of the part corresponding to 1, 2.3 in the figure. First, the encoding algorithm performed in the above block diagram will be explained below.

8TJCP1: Set MV==O, p=08
TJIiP2: Converts 8-bit data into a 10-bit code (named W, and the first bit of W (always 1) is named W) using the aforementioned encoding method.

8TffiP3: Calculate OD8.

8-gun sp4: (1) If p=Q and sign D8V
If = SignCD8, reverse the town and set DSV = D8V-OD8.

(11) If p=o and algn DSV % 51g
n0D8 Saraha w, Han inversion Te DSV =+=D8v
Set +ODB.

On) If p=j and Sign DSV = 51gn
CD8 Suraba w, q non-inversion DSV = D8V-CD
Set it to 8.

If p=1 and sign DSV J? 51gn0
If D8, invert W, D8V = D13V + O
Set with DS.

The code W after the operation of 8TjCP4 is completed is named V.

Note that P is 72g, which is 0 if the number of 61'' is even and 1 if it is odd.
Hlgh 1evel (1) of the video signal is +1, and L
ow 1evel (0) is -1, and the total value is CD8 (Oodeword Digital
Bun). However, it is assumed that the process starts from the low 1 level of the 2-video signal converted to Wt-NRZ. Also indicates the polarity of Sign ODB (+or-). However, when ODE is 0, it is 111gn0D8ij+.

BTRP52P! PΦ (P of T) where e indicates exclusive OR.

Output W' as a sign.

8TJCP6: STj! for next encoding! iP2
Jump to.

Based on the above explanation of the algorithm, FIG. 7 will be explained. The data of every 8 bits of fc input from the input terminal shown in Fig.
The corresponding 10-bit code is read from the ROM 51 and input in parallel to the shift register 52 (BTIAPR of the above algorithm).

This 10-bit code is also input to 35 77 registers at the same time.The 10-bit code coded in 635 is passed through 34 MRZX converters and NRZI converted.

The number of @1'' in this NRZ-converted code is 35 CDs.
8 to counter A, and the number of 0'' to 36 CtDll.
i counter B, and the respective values are stored in registers 57 and 38, register B (5Tnps in the above algorithm). The 34 ON RZ converter is constructed as shown in FIG. 8, in which O in FIG. 8 indicates an input terminal and a 59-digit exclusive OR circuit 40 has a 1-bit delay element. Also, the calculation of CD8 is done by subtracter 4 in Figure 7.
By subtracting the contents of register B of S8 from the contents of register S7, the results of requests tb and 3 are stored in register CD8 of 42.

Next, 8Tj! of the above algorithm! :P4 but (1
) ~ 1iv) case classification is (1) Jin・-・zP(Oqeaq)−1(11) Jin=>
P (Oq■aq)=1(+n)<=-shi p(cq
t number + jq) = 1QJ a = - =>1' (Oqe
aq) = ', and this determination is performed by 43 logic circuits. However, the value of pFi Fig. 7 44 is 1 or 0.
It is. It is MBB of the CD date register of Oq.42,
(lq is the MBB of the DBV register 45. The contents of the register are expressed in two's complement, so when MBB is 1 it indicates a negative number, and when it is 0 it indicates a positive number, so the comparison of the polarities of DSV and CD8 is It is only necessary to compare aq and aq.■ indicates exclusive OR.Based on the above four flag signals, the logic circuit of 43 is 460 if the flag of 1) is on.
Invert town bit Sase D8V == DSV - CD8
47 adders/subtractors perform the operation, and the result of the operation is 45
The data is stored in the D8V register. Figure 9 shows the above four 72
A specific example of a programming signal generation circuit is shown below. The details are omitted. As above, 8TJ! iP4 ends. Next 5
TJI! P5, but the code W' is 77 registers 52
■ Outputs from the output terminal of 48 and outputs from the output terminal of 48
Identify whether the number of 1's in T is even or odd in P clip 70gm. #) When the number of 10 pieces is an even number, 48 should be 0, and when it is an odd number, it should be 1. This can be done by setting the initial value of the 48 7 lip 70 lip to 0 and inverting it at the 9th time 1 comes. This content is stored in 490 baths. Then, the exclusive OR circuit section 50 performs an exclusive OR operation on the contents of 49 and P44, and the result is stored in P 44. With the above, 87IP5 is completed, and the process returns to 8TjCP2 and repeats the same process.

As detailed above, the method of the present invention (8°10.0
．． 2) According to the sign, Trnin =;! Q+8T, Tmax=;(2+1
)Tm2.4TTw=-T=0. With BT, in the conventional method, at typical 0 (8t9tOt3), Tm1n = -TwQ, $9 T, Tmax
=H(3+1)=3,56Tw =-Tw0.89
For example, when comparing this method and this method, this method can significantly reduce Tmax with a slight decrease in IiTwin and Tw, and is easier to synchronize and has fewer low frequency components. There is an effect that Note that the same holds true when compared with other methods. Also, a DO7 Lee code is realized using an (8,10,0,2) code. Also.

When selecting 256 codes from the 274 codes shown in Fig. 2, 194 codes in Fig. 4(a), 81 codes in Fig. 4(b), and codes in Fig. 4(0) are selected. 26 from inside It@
By selecting , the number of cases in which the limit is violated in the concatenated part of the code can be minimized, and efficient encoding can be achieved.

Also, the 1-bokuto style is rounding based on the (8, 10, 0', 2) code, and only in the concatenated part of the 10-bit code ==
2 has the possibility of breaking. However, the conventional method also has the effect of creating a DO7RI code that has few low frequency components and is easy to synchronize.

This method uses DO7jJ for every code or 10 bits.
D8V (Digi
This has the effect of creating a DO7!1-code that has a small variation range (tal Sum Variation) and has a large DC variation suppression effect. Note that the present invention is not limited to encoding from 8 bits to 10 bits, but also applies to encoding or decoding of 5 bits, or other bits other than the above, such as bits from 3 bits. Applicable to Moreover, as can be explained more than once, the present invention can provide a slave device that can perform highly efficient recitation and highly accurate data handling.

[Brief explanation of drawings]

Figure 1 is a block diagram of the configuration of an electronic device, Figure 2 is a configuration diagram showing an example of a recording/reproducing system, Figure 5 is an explanatory diagram of connections between codes, and Figure 4 is an explanation of codes in a 10-bit configuration. Fig. 7 is a block diagram of the configuration of encoding; Fig. 7 is a block diagram of the configuration of Fi exclusive OR circuit; FIG. 9 is a diagram showing a specific example of a flag signal generation circuit. 1... Information source 2... Information source encoding 5... Channel encoding 31... ROM table 34... NRZ engineering converter 32.55... Ri7 register 55.36... OD8 counter = 4 (α)! [=ko==]==I==1=2r==[=ko] t 14 Q 4I! 1(b) 1! L=1==[
2U=[=]22r=]ro s t tr#(C)
10 ■ Engineering ■ Rotoo 4441 companies? 14a C-near? i@bit DB 020304 DB 04 □ C1 (4CIC
4C6Cbooooooo xtttt otoooo totttt olooool 101110 f101010010f0 11ot1o ootoor

Claims (1)

[Scope of Claims] A binary data encoding method characterized by using the encoding algorithm shown in Table T1 below as an algorithm for converting 8-bit data into a 10-bit code. In T1 above, P is 0 if the number of '1's is even and 1 if it is odd.
Hlgh 1evel (i) of the value signal is +1 and LO
The value obtained by setting W 1evel (Q) to -1 is CD8 (Codeword Digital 8ua
+ ). However, the binary signal obtained by converting W to NRZIg 114 shall start from Low 1 level. Also, s i gncDs indicates the polarity of CD8 (-1-
or-). However, when CD8 is 0, s i gnCD8
shall be ten.