JPH04115614A - Automatic equalizer - Google Patents

Abstract

PURPOSE:To obtain a picture excellent in the visual sense by weighting the correction calculation of a tap coefficient so that the distortion removing effect of a tap corresponding to a distortion component whose delay time from a main signal is large is reduced. CONSTITUTION:A microprocessor 25 obtains differential waveform from GCR waveform fetched in an input and an output waveform memories 4,15, and obtains error waveform from the output differential waveform and reference signal waveform stored in a ROM 26, and further, it corrects the tap coefficient by the correlation calculation of the error waveform and input waveform. In the tap coefficient correction calculation, leak quantity is set for every tap, and the leak quantity is weighted. Namely, the leak quantity is made large linearly in proportion as the delay time from the main signal becomes long, and the removal of a ghost is suppressed. Thus, the excellent picture in the visual sense, can be obtained.

Description

[Detailed Description of the Invention] [Purpose of the Invention (Industrial Application Field) The present invention utilizes a reference signal periodically inserted into a received signal to automatically correct linear distortion in a transmission system on the receiving side. Concerning an automatic equalizer that equalizes .

(Prior Art) In recent years, in television broadcasting, GcR (ghost cancello
rreference) It has been decided to insert the signal (Reference 1, "Ghost Removal Method Lecture Materials, April 13, 1989, detailed at the Broadcasting Technology Development Council") into the vertical retrace interval, and this OCR signal is used to A television receiver has been developed that uses waveform equalization to remove ghosts.This OCR signal is composed of a 5inx/X bar waveform and a pedestal waveform, and as shown in Figure 9, The 5inx/x bar waveform and the pedestal waveform are inserted in the 188th and 281H in the blanking section.As shown in FIG. .8 field contains 5inx/x bar waveforms (Sl, S3, 5
6S8) is inserted, and the pedestal waveforms (S2, S4, S5, S7) are inserted in the 2.4 and 5.7 fields.
) is inserted. By performing the 8-field sequence calculation shown in equation (1) below on these signals, the influence of the previous line, horizontal synchronization signal, and colorverse 1~ signal is removed, and the GCR signal component S GCR is It's being taken out.

5GCR-(SI S5) + (S6 S2
) + (S3 -37 ) + (S8 -34
)... 1) FIG. 10 is a block diagram showing a conventional automatic equalizer that performs waveform equalization using this OCR signal.

A video signal subjected to ghost interference is input to an input terminal 1. An OCR signal for removing ghosts is inserted into this video signal. The input video signal is sent to an analog/digital converter (hereinafter referred to as A/D converter) 2
is sampled every unit time T seconds and converted into a digital signal by a transversal filter (hereinafter referred to as TF) 3 with variable tap coefficients and an input waveform memory 4.
given to. Note that the clock generation circuit 16 generates a period T (for example, T - about 70
ns=1/4 fsc (fsc is color subcarrier frequency)
The clock CK is created.

The transversal filter 3 includes a delay circuit group 5 consisting of unit time delay circuits connected in series, a multiplier group 6, an adder 7, and a tap gain memory 8. Tap coefficients Cw- to Cn stored in tap gain memory 8
is applied to each multiplier in the multiplier group 6 to determine the gain of each multiplier. The output of the A/D converter 2 is sequentially delayed by each delay circuit of the delay circuit group 5. The delay time between the taps of the delay circuit is set, for example, to the sampling period (about 70 ns) of the input video signal. Each delayed signal from the delay circuit is applied to each multiplier in the multiplier group 6 and given a tap coefficient. Since the equalization time range is determined according to the number of taps, the total number of taps is determined according to the range of delay time of the ghost to be erased. Note that among these taps, the tap through which the input signal passes with a relative time delay of 0, that is, the tap of the multiplier to which the main signal is input is referred to as the main tap. m before the main tap (relative delay time is negative)
The front ghost ■~ is removed by the n taps, and the rear ghost is removed by the n taps at the rear (where the relative delay time is positive). The outputs of each multiplier are added by an adder 7, and a video signal from which ghosts have been removed is outputted to a pressure terminal 9.

The tap coefficient is corrected every unit time T by the microprocessor 10 performing the calculation shown in the flowchart of FIG. 11 using the input GCR signal from the A/D converter 2 and the output OCR signal from the TF 3. be done. That is,
First, when power is turned on or channel switching is performed in step A1 of FIG.
0 sets the initial state in the next step A2. As a result, tap coefficients c-1 to C2 become zero. The subsequent steps A3 to A8 constitute an equalization loop.

This loose equalization operation will be explained with reference to the waveform diagram of FIG. 12. FIG. 12(a) shows the input/output GCR waveform, FIG. 12(b) shows the difference waveform x, y, and FIG. 12(c)
shows the reference waveform r, and FIG. 12(d) shows the error waveform e.

In step A3, input and output OCR waveforms are captured. The S1 line detection circuit 13 is 1 shown in FIG.
The arrival of the second OCR signal S1 is detected and a detection signal is provided to the acquisition control circuit 14. When the acquisition control circuit 14 receives the detection signal, it instructs the microprocessor 10, input waveform memory 4, and output waveform memory 15 to acquire the GCR signal waveform. Here, the second sample value of the input GCR waveform is denoted by X, and the second sample value of the output GCR waveform is denoted by Yk. In addition, the input and output OCR waveforms are collectively referred to as 5O
CR. The input/output waveform memory 4.15 captures waveforms from the first OCR signal (X+, Yl) to the eighth GCR signal (Xs, Ys). 1st
FIG. 3 is a memory map of the work RAM 12. As shown in this memory map, the microprocessor 10 stores the GCR waveforms captured field by field by the input/output waveform memories 4 and 15 in each area in the work RAM 12. .

In the next step A4, the microprocessor 10 reads the input GCR signals (x+ to Xs) and the output GCR signals (Yl to Y,) from the working RAM 12, and reads the above (
]) and calculate the first
The final OCR signal ()Cac-, Yac shown in Figure 2(a)
! +). Note that the GCR signal XccR, YC;
R is composed of 1 word (1 word - 8 bits) and can be expressed by a sample value series shown in equations (2) and (3) below.

XGCR-(Xcc*, x) (k=0~1023
) −(2)Y OCR= (Y ac++ to
) (k=0 ~ 1023) −・ku 3
) In the next step A5, the microprocessor 10 performs the difference calculations shown in equations (4) and (5) below to obtain the difference waveforms (xi) and (yk) (FIG. 12(b)) and performs the work RA. Store in M12.

Xk −XGCR, +l XGCR,K −
(4) 3/ k= YGCR, +l Yac*, *
-(5) Next, the microprocessor 10:
Step A6) detects the maximum peak of the difference waveform (xk) of the final OCR signal and determines the peak position. Note that the input difference waveform X at the peak position indicates the peak of the main signal impulse.

In the next step A7, the microprocessor 10 obtains the error waveform (ek) shown in FIG. 12(d). R
The reference signal waveform (r=) shown in the 12th [3(C)] is stored in advance in the OM 17, and the microprocessor 1
0 subtracts the reference signal waveform (rk) from the output waveform 1yk) at the peak position (formula (6) below) to obtain an error waveform (ekl) and stores it in the work RAM 12.

e* = y* r'* + (6) Next, in step A8, the microprocessor 10
Proportional method shown in formula (7) below or (8) below
The tap coefficient is corrected based on the incremental method shown in the equation. As shown in equations (7) and (8), the tap coefficient correction amount is determined by a correlation calculation between the input OCR waveform and the error waveform determined from the output GCR signal and the reference signal.

CI IIew: C1, ald (Z
' Σ X k ° e h+1 ''' (
7) CI IIew = C1, oldδ ・
sgn(ΣXk ・ek+1)−(8) Here,
The subscript i is the tap coefficient of the goth 1- with the delay time IT seconds.
new, ol
d indicates before and after modification, respectively. Further, sgn is an encoding function, and α and δ are small positive correction amounts.

The microprocessor 10 uses the obtained tap coefficients as work RA.
It is transferred from the tap coefficient storage area of M12 to the tap gain memory 8. This tap coefficient is applied to each multiplier in the multiplier group 6 to add a coefficient to the delayed signal, and the adder 7 adds the coefficients to perform waveform equalization.

Thereafter, the equalization loop of steps A3 to A8 is repeated. Through these steps A3 to A8, tap coefficients are generated based on the magnitude of the error waveform every unit time T, that is, tap coefficients are generated so that the error signal converges to 0, and ghosts of the input video signal are generated. is removed.

By the way, for waveform equalization or proximity ghost use, T
As F, one having a relatively narrow equalization time range (few taps) can be used.

FIG. 14 is an explanatory diagram showing the correspondence between each tap and the ghost delay time. As shown in FIG. 14, for example,
The number of taps can be limited to about 8 taps before and after the main tap.

Now, as shown by the upward solid line in FIG. 14, the ghost component of the delay time corresponding to 1 (delay time fJ
It is assumed that ghost) 20 is mixed in after T seconds. When this ghost component 20 is removed by the apparatus shown in FIG. 10, as shown by the upward broken line in FIG. will occur. However, since 21 exists in the tap corresponding to this grandchild ghost 21,
This grandchild ghost 21 is eventually removed. The grandchild ghost generated by removing this grandchild ghost 21 is 31 away from the main tap, and there is no corresponding tap, but it is sufficiently attenuated so that it is not noticeable on the screen.

However, from the main tap to n taps ((1/2)
(n<8) IlI may be mixed with a delay time of I to 22 corresponding to the tap. In this case,
The delay time of the grandchild ghost 23 generated by removing this ghost 22 corresponds to 2n taps.

That is, in the example of FIG. 14 in which the number of taps is limited to eight taps before and after the main tap, there are no taps corresponding to the grandchild ghosts 1 to 23, and the grandchild ghost 23 cannot be removed.

According to Document 2 (“A Study on the Quantification of Subjective Evaluation of Multiple Ghost Interference J”, NHK Research Institute of Technology, Shigeru Yamazaki, Journal of the Television Society), it is stated that the short delay time corresponding to the taps near the main tap (approximately 0 to 50μ Regarding the proximity ghost 1- (seconds), it has been shown that the longer the delay time, the greater the influence on the subjective evaluation value.In other words, the shorter the delay time from the main signal, the less noticeable the proximity ghost is on the screen. The closer the ghost with a longer delay time from the main signal, the more noticeable it will be on the screen.In other words, the closer the original closer ghost is, the more noticeable it will be on the screen. The power of removing grandchild ghosts, which have a relatively long delay time, is a serious problem.

Therefore, if the number of taps is limited, even if a close ghost is removed, the grandchild ghost caused by this close ghost cannot be removed, and depending on the delay time of the close ghost, the screen will deteriorate significantly. There was a problem.

(Problem to be Solved by the Invention) As described above, in the conventional automatic equalizer described above, when the number of taps is limited, even if close ghosts are removed, grandchild ghosts with a relatively long delay time still occur. Since the ghost remains without being removed, there is a problem in that the screen deteriorates significantly depending on the delay time of the proximity ghost.

The present invention has been made in view of such problems, and includes:
It is an object of the present invention to provide an automatic equalizer that can obtain visually good images by weighting tap coefficients given to each tap according to the distance from the main tap.

[Structure of the Invention] Means for Solving the Problems) An automatic equalizer according to claim 1 of the present invention comprises an equalization means including three transversal filters with variable tap coefficients, and an output signal of the equalization means. tap coefficient correction means for detecting distortion and correcting the tap coefficient; and placing emphasis on correction calculation of the tap coefficient so that the distortion removal effect is reduced as the tap corresponds to a distortion component with a large delay time from the main signal. In the automatic equalizer according to claim 2 of the present invention, the weighting means includes a tap coefficient for each tap according to the distance from the main tap. In the automatic equalizer according to claim 3 of the present invention, the weighting means multiplies the tap coefficient by a predetermined window function set for each tap. .

(Operation) In the present invention, distortion components due to taps near the frontmost tap and last tap are not completely removed because their distortion removal effect is small. Therefore, by removing this distortion component, the level of the newly generated distortion component is also small. Although there is no tap corresponding to the delay time of this newly generated distortion component and it cannot be removed, it is not noticeable on the screen because the level is small.

(Example) Hereinafter, an example of the present invention will be described with reference to the drawings. FIG. 1 is a block diagram showing an embodiment of an automatic equalizer according to the present invention. In FIG. 1, the same components as in FIG. 10 are given the same reference numerals.

A video signal into which an OCR signal has been inserted is input to an input terminal 1. The A/D converter 2 converts this input video signal into a digital signal and sends it to the TF3 and S1 line detection circuits 13.
, to the clock generation circuit 16 and input waveform memory 4. The TF 3 is composed of a delay circuit group 5, a multiplier group 6, an adder 7, a tap gain memory 8, etc., and the gain of the multiplier is controlled by the tap coefficient to equalize the waveform of the input signal. A clock generation circuit 16 generates a clock within the device. The S1 line detection circuit 13 detects the GCR signal S1 and provides the detection signal to the acquisition control circuit 14.

When the acquisition control circuit 14 receives the detection signal, it outputs the waveform acquisition timing (No. 3) to the input waveform memory 4, the output waveform memory 15, and the microprocessor 25.

The signal whose waveform has been equalized by the TF 3 is outputted from the output terminal 9 and is also given to the output waveform memory 15 .

The input/output waveform memories 4, 15 take in the 5inx/x bar waveform at the timing of the timing signal, and the microprocessor 25 takes in the OC
Find the difference waveform from the R waveform and save the output difference waveform and ROM26.
The error waveform is obtained from the reference signal waveform stored in the input difference waveform, and the tap coefficient is corrected by performing a correlation calculation between the error waveform and the input difference waveform.

In this embodiment, the program for equalization operation stored in the ROM 26 is different from the conventional one, and the microprocessor 25 performs the calculation shown in the following equation (9) to obtain the tsubu coefficient. As shown in equation (9), in the tap coefficient correction calculation, the leakage amount is set for each tap.

c+, new ” (1-β' f(i) ) −c
1. old-α・ΣXh・ek++ −<9)
However, the subscript 1 indicates a tap, f(i) is a function that weights the leakage amount for each tap, and β is a leakage coefficient (
constant).

FIG. 2 is a graph showing the weighting f(i) of the leakage amount for each tap from -s to Ko to 8.

Note that FIG. 2 shows an example where the number of taps is limited to eight taps before and after the main tap. In FIG. 2, the leakage amount for taps from -4 to 4 is 0, and the leakage amounts for taps from -5 to - and taps from 5 to Ks are varied linearly. That is, as the delay time from the main signal increases, the amount of leakage increases linearly to suppress ghost removal.

Changes in the amount of leakage are made according to the state of the ghost. If a ghost corresponding to a tap that is 4 or more taps away from the main tap is mixed in, as mentioned above, there are no taps necessary to remove the grandchild ghost, so processing is performed to reduce the influence of the grandchild ghost. I do. In other words, if the level of the ghost 1 component with a long delay time is relatively large, the level of the grandchild ghost will also be large. The weighting of the leak amount is set to 1 to sufficiently suppress ghost removal. On the other hand, when the level of the ghost component with a long delay time is relatively small, as shown in Figure 2 (b), the maximum value of the weighting of the leak amount is set to 1/2, and the amount of ghost removed is compared. Make the target bigger.

Next, the operation of the automatic equalizer configured as described above will be explained with reference to the flowchart of FIG. 3.

The operations in steps A1 to A7 are the same as in the conventional method. That is, first, when power is turned on or channel switching is performed in step A1, the microprocessor 25 sets an initial state in the next step A2. In the next step A3, the capture control circuit 14 outputs a timing signal based on the detection signal from the S1 line detection circuit 13, and the input/output OCR waveform is captured. The GCR signal is taken into the input/output waveform memories 4 and 15.

In the next step A4, the microprocessor 25 reads the input and output OCR signals from the working RAM 12, and
The final GCR signal (XGcRlYGcn) is obtained by calculating the 8-field sequence of equation (1) above.

In the next step A5, the microprocessor 25 performs the difference calculation shown in equations (4) and (5) above to form a difference waveform (
xk), fyk) and store them in the work RAM 12.

Next, the microprocessor 25 detects the maximum peak of the difference waveform (X, l) of the final OCR signal to determine the peak position (step A6). Next, in step A7, the microprocessor 25 detects the maximum peak of the difference waveform (X, l) of the final OCR signal and determines the peak position (step A6). ) and the output waveform (y), an error waveform (ek) is obtained and stored in the work RAM 12.

In this embodiment, in the next step A9, the microprocessor 25 corrects the tap coefficients based on equation (9) above. Delay time from main signal to tap -4
If it is within the range corresponding to 4 to 4, the weighting of the leakage amount is 0, and the above equation (9) is equivalent to the above equation (7). If the delay time is outside this range, the amount of leakage is increased linearly as the delay time becomes longer, as shown in FIG. 2(a) or FIG. 2(b). Then,
As the delay time becomes longer, the amount of leakage increases and the amount of ghost removal is reduced. That is, the longer the delay time of a nearby ghost, the lower the level of its grandchild ghost.

In this way, in this embodiment, the longer the delay time, the more the leakage amount is increased to reduce the amount of ghost removal, and the level of grandchild ghosts with a relatively long delay time is lowered to reduce the amount of ghosts removed on the screen. I try to keep a low profile.

FIG. 4 is a flow chart for explaining another embodiment of the present invention.

In this embodiment, steps A7 and A in the equalization loop are
A step AIO is provided between step 8 and step AIO in which the error waveform is weighted according to the following formula (]O), and step A1 is provided after step 8 to determine whether or not the process has been repeated a predetermined number of times.
This is different from the embodiment of FIG. 1 in that 1 is added.

eh = ei+ ·a* = (10) The operation of the embodiment configured as described above will be explained with reference to the explanatory diagram of FIG. Fig. 5(a) shows the error waveform 1ek), Fig. 5(b) shows the weighted coefficient (a-), and the 5th
Figure (c) shows the corrected error waveform (e-').

Steps A1 to A7 are the same as in the embodiment of FIG.

Now, as shown in FIG. 5(a), an error waveform e-x is generated by the ghost of the delay time corresponding to -2 at the tap, and an error waveform e6 is generated by the ghost of the delay time corresponding to 6 at the tap. shall be. Error waveform e-2, eb
shall be at the same level. Step AIO in Figure 4
In , these error waveforms and weights are multiplied by a coefficient. A coefficient (ak) for each weight is set for each tap, and in this embodiment, as shown in FIG. 5(b), the values of the counts a-a to a4 are m (positive numbers). and the coefficient a−a
The coefficients a-a and as decrease linearly in the coefficients a-s and the coefficients a4-a8, and the coefficients a-a and as become zero.

By calculating the above equation (10), a corrected error waveform e es' is obtained as shown in FIG. 5<c>. Tap coefficient correction calculations are performed on these corrected error waveforms in the next step A8. In this case, for example, if the weighting coefficient is 1/2, the amount of ghost correction calculated in step A8 is equivalently 1/2.

The weights corresponding to taps from -4 to 4 are coefficients a to a4.
The value of is larger than the other coefficients a-s to a-s and aq to a8, and the amount of ghost correction is large.

On the other hand, the coefficients a-g to a-s and the coefficients as to a
In s, the value of the coefficient corresponding to the tap farther away from the main tap is smaller, and the amount of ghost correction is smaller. That is, the convergence time of the error waveforms corresponding to taps from -4 to 4 is shorter than the convergence times of the error waveforms corresponding to other taps, and corresponds to the taps whose distance from the main tap is large. The more error waveform, the longer the convergence time.

In the next step A11, it is determined whether the number of corrections of the tap coefficients has reached a predetermined number. When the number of corrections reaches a predetermined number, the correction operation is stopped.

This makes it possible to converge only the error waveforms corresponding to - and 4 on the tap, and stop the operation without converging the error waveforms corresponding to the other taps, especially the taps with a long delay from the main tap. can. As a result, the longer the delay time from the main tap is for a closer ghost, the more the ghost component will not be completely removed, and the influence of the grandchild ghost can be reduced.

In this embodiment, as shown in FIG. 6, a group of coefficients may be used for multiple weightings. In FIG. 6(a), the values of coefficients a-a to a4 are m. be. coefficient a-4
The values of the coefficients a4 to a8 decrease linearly, so that the a-series number a-s is O. Similarly, the values of the coefficients a4 to a8 decrease linearly, and the coefficient as is O.

Furthermore, in the weighted coefficient group shown in FIG. 6(b), the values of coefficients a-4 to a4 are m. The values of the coefficients a to a-s decrease linearly so that the coefficient a-s is m/2, and similarly, the values of the coefficients a4 to a8 decrease linearly to the coefficient a8.
is m/2. Moreover, the weighting shown in FIG. 6(c) is for the coefficient group from coefficients a-++ to a! The value of I is m in all cases. This ensures ghost I/removal when no nearby ghost with a relatively long delay time is mixed.

Depending on the state of the ghost, one of these weighting coefficients can be selected to obtain the visually best image. Furthermore, the same effect can be obtained by weighting the correction coefficients α and δ instead of weighting the error waveform (ek).

FIG. 7 is an explanatory diagram for explaining another embodiment of the present invention.

Generally, visually, the screen deteriorates more significantly when a negative ghost is mixed than when a positive ghost is mixed. Even if either the positive ghost or the negative ghost is removed, a grandchild ghost that becomes a negative ghost will be generated. For this reason, the correct goth 1~
If the negative grandchild ghost remains unremoved by removing .

The original ghost removal may further deteriorate the image. Conversely, by removing negative ghosts, even if negative grandchild ghosts remain without being removed, the image quality is improved because the originally high-level negative ghosts are removed.

For this reason, in this embodiment, different weights are given to positive ghosts and negative ghosts. For the positive ghosts corresponding to the taps near the front tap and the last tap, weight is applied so as not to remove them, or
Alternatively, weighting is performed so that the amount of removal is smaller than the amount of negative ghost removal corresponding to these taps. This makes it possible to visually improve the waste image.

Usually, tap coefficients for removing positive ghosts are often negative, and tap coefficients for removing negative ghosts are often positive. Therefore, it can be determined from the value of the tap coefficient whether a positive ghost or a negative ghost is mixed. After determining whether it is a positive ghost or a negative ghost in this manner, the weighting amount for the error signal is set in accordance with the taps, as shown in FIG. 7(a). In FIG. 7, white circles indicate positive tap coefficients, and black circles indicate negative tap coefficients.

Moreover, as shown in FIG. 7(b), 1. The leak amount may be set depending on the tap.

FIG. 8 is a flowchart for explaining another embodiment of the present invention. This embodiment applies ghost removal using a batch calculation method in the frequency domain to ghost removal using the sequential correction method on the time axis described above.

Steps A1 to A4 are the same as in the embodiment of FIG.
In this embodiment, the tap coefficient CI is determined by calculation in the frequency domain in the next step AI2. Regarding the batch calculation method in the frequency domain, see Reference 3 (
The details of the development of the ghost canceller are disclosed by Tanaka, Sasaki, Miyazaki, Kobayashi, Shiki, Kobayashi, Television Society Technical Report). Furthermore, in the next step A13, the tap coefficients are weighted by the calculation shown in equation (1) below.

C+ :C+ + a+ +++ (11
) In this case, the weighting is performed by the coefficient a+ such that ghosts corresponding to the taps near the front tap and the last tap are not completely removed.

As a result, the same effects as in the other embodiments can be obtained in this embodiment as well.

[Effects of the Invention] As explained above, according to the present invention, a visually good image can be obtained by weighting the tap coefficients given to each tap according to the distance from the main tap. has.

[Brief explanation of drawings]

Fig. 1 is a block diagram showing an embodiment of the automatic isolator according to the present invention, Fig. 2 is a graph showing the weighting f(i) of the leakage amount, and Fig. 3 is for explaining the operation of the embodiment. 4 is a flowchart for explaining another embodiment of the present invention, FIGS. 5 and 6 are explanatory diagrams for explaining the embodiment of FIG. 4, and FIG. 7 is a flowchart for explaining another embodiment of the present invention. An explanatory diagram for explaining another embodiment, FIG. 8 is a flowchart for explaining another embodiment of the present invention, FIG. 9 is a waveform diagram for explaining an OCR signal, and FIG. 10 is a conventional OCR signal. A block diagram showing an automatic equalizer, FIG. 11 is a flowchart for explaining the operation of the conventional example), FIG. 12 is a waveform diagram for explaining the operation of the conventional example, and FIG. 13 is a flowchart for explaining the operation of the conventional example. Internal work RAM
FIG. 14 is an explanatory diagram for explaining the problems of the conventional example. 3...Transversal filter, 4...Input waveform memory, 15...Output waveform memory, 2
5...Microprocessor. Figure 11

Claims (3)

[Claims] (1) Equalization means including a transversal filter with variable tap coefficients, tap coefficient correction means for detecting distortion of the output signal of the equalization means and correcting the tap coefficients, and a large delay time from the main signal. An automatic equalizer comprising weighting means for weighting the correction calculation of the tap coefficients so that the distortion removal effect is reduced as the tap corresponds to the distortion component. (2) The automatic equalizer according to claim 1, wherein the weighting means changes the amount of tap coefficient leakage for each tap depending on the distance from the main tap. (3) Claim 1, wherein the weighting means multiplies the tap coefficient by a predetermined window function set for each tap.
The automatic equalizer described in