JPS59207401A - Reproduction signal processing system of recording disk - Google Patents

Abstract

PURPOSE:To eliminate completely the scratch noises by detecting and correcting the scratch noise after converting the reproduction signals into digital data and then delivering the digital data after converting it into analog signals. CONSTITUTION:An arithmetic circuit 6 and an RAM7 sample the analog reproduction signal of a recording disk to obtain a digital data Xn and calculates a smoothed data X'n of data X(n-1) and X(n+1) which are positioned beofore and after the data Xn. The data (an) corresponding to the difference between data Xn and X'n is compared with the prescribed reference data. When the data (an) is larger than the prescribed data, the data Xn corresponding to the data (an) and the data before and after the Xn are corrected by a prescribed correct- ing method. These data are converted into analog signals by a D/A converter 8 and delivered. As a result, the impulsive noises due to the flaws, attachment of foreign matters, etc. on a recording disk are eliminated completely. Thus, the quality of the reproduced music is improved.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a reproduction signal processing method for a recording disk, and in particular to a reproduction signal processing method for removing so-called scratch noise caused by scratches on the recording disk. This may damage the tone groove in which music signals are recorded. In addition, it is common to experience foreign matter such as dust adhering to the device. When such a disc is played on a disc player, impulse-like so-called scratch noise occurs.4. In particular, if there is a scratch, it often spans several sound grooves, and as the pickup traces uneven spots each time the disc rotates, the noise periodically mixes into the reproduced sound, making it extremely difficult to hear. This has become a fatal flaw in that it becomes a bad thing and also damages the V( of the music played).

SUMMARY OF THE INVENTION An object of the present invention is to provide a reproduction signal processing system for a recording disk that can remove impulse-like noise caused by damage to the recording disk or foreign matter landing on the recording disk.

The playback signal processing method according to the present invention samples an analog playback signal from a recording disk and converts digital data [nu]
Smoothed data X' 1 of j2-xn-1,xl+1
1, and this smoothed data x'n and original data x 1
Compare the data El n corresponding to the difference from 1 with the standard footer and find that this data
The present invention is characterized in that data before and after data including data X11 corresponding to J1 are corrected by J2 using a predetermined correction method, and then converted into prelog credit after being processed.

The present invention will be explained below with reference to the drawings.

FIG. 1 is a block diagram of an embodiment of the invention.

In the figure -C11 [- and 1R are records -j゛Isk]
When playing a record (referred to as a record) on a record player A7, L'? This is a manual amplifier for amplifying the R channel, Vl channel, and P signal to an amplitude level suitable for Δ-1 (analog-to-digital) conversion. 2L and 2R are band-limiting low-pass filters for preventing generation of aliasing sound during sampling. J: As is well known, this low-pass filter cut and
The off frequency is set to 1/2 the sampling frequency. In the case of normal audio delta frequency band processing, the sampling frequency is about 40 i < kl '/:, so the cutoff and off frequencies of these l-1-pass filters are 20 K l-l z (d 3L and 3R are the time required to sample the analog signal 8 and perform A-D conversion. Since the signal of the system is divided into time and processed by one A/D converter, l
, R channel sampling is performed alternately with a phase shift.

4 is a multiplex circuit that supplies the lc data alternately sampled by the lean pull bold circuit to the A/D converter. 5 is an A/D converter for quantizing the amplitude value of the sampled data, and as mentioned above,
In this example, two systems of signals are processed by one A/D converter, so the A/D converter is twice the sampling frequency, about 80KH2.
To perform a conversion.

6 is an arithmetic circuit that performs arithmetic processing that will be described later; 7 (Random access memory 1 for storing data necessary for data correction that will be described later) (RAM)
It is.

The arithmetic circuit 6 and the RAM 7 are connected by an address bus for accessing the address of this RAM, a chip select for data write/read control, a data bus for inputting/outputting data. ing. 8 is an I)/△ converter for D-A converting the data that has been subjected to the arithmetic processing described later. In this example, one D/A converter converts two systems of signals into D-△. Therefore, like A/D converter 5, this D/A converter also performs D-A conversion at about 280 Kl-1z.
cormorant. 9 (This is a demultiplex circuit that distributes the two systems of signals that have been input into the R″ft channel. Also, here, the D/A
Converter 8 (Koni'C5E Ushizuru glitch excluded, time I
+l; Make it into a discrete RAM wave above, and then fill in the next [1-bus fill ([supply].
This is a low-pass filter that filters the waves and restores them to the original analog signal, and the cut-off frequency of this filter is also set around 20 Kl-1z. 111-
and 11R are output amplifiers that amplify the analog signal output of the low-pass filter. 12 is a system clock generation circuit that generates a system clock for controlling this system, from which a sample hold circuit control clock, a multiplex circuit control clock, a △/D converter control clock, circuit control clock,
Demultiplex circuit control [cocks are supplied separately.

Next, the operation of the present invention will be explained. The feature of the present invention is to first convert the playback signal of the record player A7 into digital data, and detect scratch noise on this data by the operation of the arithmetic circuit 6 and RAM 7 shown in FIG. The purpose is to correct the noise portion, restore the analog signal, and output it.Therefore, the operations of the arithmetic circuit 6 and IRΔM7 will be mainly explained with reference to FIG.

(Figure 2 (Δ) is music containing scratch noise).
- played on the record player 1: Yu (ko 1111
This is a sample of the filthy river I signal. The scratch noise part is very steep compared to the movement of the waveform before and after it.
= (...,XQ,XI,X2',X3゜×
1. ...) (1 is the sampling interval), x
The sample value after the phase of n is 0'-xn-i.

×rl]1; IJ, x'n - (xn-1+xn+ 1)/2, cl,
Data string X-'(nt,) = smoothed sample value
(-, x' o, x' 1.

X'2. x'3. x'4. ...). Data sequence a(nt) of the square of the difference between this smoothed data x'n and the original data X11 = (-, (Xo - x'o) 2° (XI
'+)', (X2-X'2)21 (X3-X'3)'.

......) = (..., ao, a+, a2, a3・
...) When I play 81 rounds, a(nt) becomes very large near the peak of the scrap noise. an == (xn-x
'11) The purpose of squaring the equation 2 is to emphasize the magnitude of the numerical value and to make the value of an a positive value to simplify subsequent processing. Poly a(nt) in Figure 2<13).

Therefore, add λj to this a(+'+t), set the threshold value C]t1 for scratch discrimination in advance, and use J3, an>
It is sufficient to detect the CTH portion as scratch noise. In this case, since the scratch noise is band-limited, even something like () can span multiple +) samples. For general scratch noise, 40 K l-I Z
When sampled at , contamination (ρ often occurs) over around 20 samples. Therefore, if the date aII>C'r l-1 is detected, an will reach its maximum over about 20 samples after that. All you have to do is find a spot, make it the center of the scratch noise, and correct the 10 samples before and after it in some way.

As a correction method, if the main focus is on concealing scrap noise, a sufficient effect can be obtained by using a linear interpolation method as shown in FIG.

When scratch noise occurs in the range indicated by the dotted line in Figure 3, it is corrected to a value such that each sample data exists on a straight line connecting the first data and the last data in this range. It is.

This linear interpolation is used in PCM (Pulse Code Modulation) digital audio equipment when an error can be detected but cannot be corrected. , is widely used together with the average value interpolation methods.

In this case, in digital audio equipment that uses general error detection codes, the interpolation range is at most several consecutive points, whereas in the present invention, the number of data points for which continuous interpolation is performed is large and/or large. j, so it is necessary to prevent discontinuous parts caused by interpolation from becoming audible.Therefore, compared to methods that have a DC component in a certain interval, such as prefix and postfix retention, Although linear interpolation is more advantageous, it is not limited to this.Furthermore, a method may be considered in which the number of data equal to the number of data to be interpolated is taken from the immediately preceding part i1) and replaced.

Linear interpolation itself is an extension of average value interpolation, so it is based on the so-called least squares method (linear approximation).When data in a certain interval is represented by a straight line using linear interpolation , its corrected error curve is expressed by where xm is the true number of data points, a is the data position where linear interpolation starts, b is the data position where linear interpolation ends, and B is the slope of the straight line. , xa is the data value at the point where linear interpolation starts.

FIG. 2(C) shows a waveform corrected by the linear interpolation method. FIG. 3 shows each calculation process of scratch nose r portion detection and linear interpolation using a chart 70. This processing process is repeated until there is no more input of the original waveform x(nt). still,
As a result of the processing of the arithmetic circuit 6 (Y△-D change 1!),
r Smile data to RAN17 once, then forward and backward j
-ri 4:, j no 11 calculation! 1 piece of strainer - 77
f., to Schiff 1〈÷2 ((1 hit), flat d゛1
The original take x [) where X'
From r1, find J, ゛(-20) in the identification (and multiplication circuit),
I'll compare A1 white C1-(by comparing and detecting scrap-noise ■'f, and draw the above-mentioned aII line between sleeves<:r-) ”'C1D/△Transformer ε Sea urchin take,
lξSupply C Yatmu IJ , j: Ikoto [2, this arc η[〈△1\11? j"jl doji 7(ha, 1st line? When performing the interval (this is required history 1 'r4', ”, 7-ri Il
l/)17'l'J 1. = +2;? i'lsa"
((It'(, (, jl, i: Ino t, Il child pip I shi 2 ni 'I ('i to shikichi'b i,
I-< 2 tunnel minutes'' (I gokata 7!'(')
0ha completed 1・, light bjj)ne'14-child 200ha(
:” +', i) l'-i, t'?) 411;I
:), between +15 and where J K.? -7'j Erotic.

1'ν'' 2, the present invention (due to the scratches and dirt on the VT+-de 'X9's I-''),
4-abbreviation, completely eliminate, ・J thing ・ bJ own pido / fru no -(:l; ・i:::', ', 1] [Good'jj'', j
: ! Tl 't j:fl or 111''. Also,
Other than i, j+j minutes when scrap oil is stopped (J△, /D
, 1]/' Each change to 1 or 1 prefecture is carried out ('J Tsuruno-(', there is a drawback that signal deterioration is bad.

] In the example of the fruit color of the fork,
Change to one (case (ko, two J for 11, “1 minute 1'l'l
7)\, this (maI)raro/υ, 1-1 A-D converter is installed for each 1-τ channel. ?At the time of writing, it was written, but at this time, (Table 1-9)
In addition, in the case of 11゜A・channel record 1~, IJ L (b, J, 9) can be similarly extended and ζ becomes -(.

[Brief explanation of the drawing]

Fig. 1 (Block diagram of the embodiment of the present invention, Fig. 2)
i') 3 [Figure (, L hundred line complement (shows an example of pJ V t**,
Q'I41 ZilJ Tree Invention Embodiment Frobe i/-l
- is. Main part 118 β (light F)...Δ, /D converter (3...) νi arithmetic circuit 7... RAM 8...D/Δ converter Volume 7・'Figure bottom 2 Prisoner (C) y Stem 3 figure hair 4 figure

Claims (1)

[Claims] 1) Digital data xn (first sample data, where n is an integer) is obtained by zumbling the analog playback signal from the recording disk, and the data x11 is sequentially and sequentially Calculate smoothed data X' rl of data X l-1, X T', +1, and calculate data E according to the difference between the data Xr1 and the smoothed data X'n
i fixed group 1 ([If the data an is greater than j compared with the data, then the data X r corresponding to this data an
A processing method for a recording disk, which is characterized in that the data before including 1 is corrected by a predetermined correction method, and then converted to an analog signal. (2) The above correction 11. The reproduction method according to claim 1, wherein the method is a linear interpolation method.
method.