JPS59219034A - Automatic equalizer - Google Patents

Abstract

PURPOSE:To attain equalizing operation in following to the change in waveform distortion by processing so that the correlating value between an error signal of the output waveform of a transversal filter and the prescribed shape of wave-form in a reception signal goes to zero. CONSTITUTION:A signal in which the prescribed form of waveform exists periodically is inputted to a transversal filter 20 whose tap gain is variable and a tap gain of the filter 20 is corrected sequentially by a tap gain correcting and operating circuit 47 based on an error signal ek relating to the error code of an output waveform corresponding to the prescribed form of waveform so as to attain automatic equalization of the reception signal. The correlating value between the error signal ek and the prescribed form of waveform in the reception signal is operated by an operating circuit 55 in this case, the leakage is added in the direction that the correlating value is reduced to zero and the result is added circulatingly, and the result is stored in a circulating adding memory 51 via a changeover device 70. The change in the waveform distortion of an output waveform is decided by a deciding device 53 from the values of the reference memory 50 and the memory 51, a changeover device 60 is changed over by a signal from a switching signal generator 54 in response to the result of discrimination so as to start or stop the correcting operation of the tap gain.

Description

[Detailed Description of the Invention] [Technical Field of the Invention] The present invention relates to an equalizer that equalizes a received signal waveform using a transversal filter with variable tap gain, and particularly relates to an equalizer that equalizes a received signal waveform using a transversal filter with variable tap gain, The present invention relates to an automatic equalizer that automatically removes linear distortion in a transmission system on the receiving side by using a waveform with a predetermined shape.

[Technical background of the invention and its problems]

A ghost canceling device for a television receiver is known as one application example of an automatic disease transmitter. FIG. 1 shows a known example of a ghost canceling device using a transversal filter with variable tap gain, in which vertical synchronization 1? (Reference document: Murakami et al. ``Digital automatic ghost erasing device'' Institute of Electronics and Communication Engineers Technical Research Report EMCJ 7B-37, November 1978) Month).

In FIG. 1, 20 is a transparsal filter with variable tap gain, which includes a tagged delay element 21, a loading circuit 22, and an adder circuit 23. Delay element 2 with tap
The delay time T between taps is selected to be smaller than the reciprocal of twice the highest frequency of the input signal, for example, 0.1 μs.

The number of tags is determined according to the range of delay (advance) time of the ghost to be erased. For example, the total number of taps is 1
If it is set to 00, a time range of 10 μs can be covered. Among the taps, the tap to which the maximum load value (tap gain) is set is called the main tap, the tap with the shorter delay time is called the front tap, and the tap with the longer delay time is called the rear tap. For example, 2 out of 100 tags
If the 0th tap is selected as the main tap, lead ghosts up to 2 μs and delay ghosts up to stts can be eliminated. The weight circuit 22 attached to each tap is a multiplication circuit, and its coefficient is the tap gain. The tap gain of the main tap is expressed as eQ, the tap gain of the front tag is C-u~C-4, and the tap gain of the rear tap is C1~c.
Let's represent it as N. co usually has a value of about 1,
The absolute values of the other tap gains c i (+=-MN) are smaller than 00.

In this powerful transversal filter 20, if the tap gain (Ci) (C-k ~ (!o'''- (!N series is expressed as (cl)) is set to an appropriate value, the input terminal 10 The ghost components (including waveform distortion caused by filters, etc.) that existed in the output terminal 30 are substantially eliminated at the output terminal 30.This tap gain can be automatically controlled to minimize all the ghost components in the output. can be done as follows.

First, from the input video signal applied to the input terminal 10,
Under the control of the timing circuit 44, only a predetermined length of the leading edge of the vertical synchronization pulse of interest is extracted and stored in the input waveform memory 41 via the differentiating circuit 40. On the other hand, only a predetermined length of the output video signal from the output terminal 30 at the same time is extracted, and after passing through the differentiation circuit 42 and the reference waveform subtraction circuit 43, the comparator 49 extracts 1 bit related to only the sign of the error. It is converted into data and stored in the error signal memory 46 as an error signal. Here, the reference waveform supplied to the reference waveform subtraction circuit 43 is generated by the reference waveform generation circuit 45 under the control of the timing circuit 44. In this way, the input waveform memory 41
The waveform stored in the sampling interval 0.1 μS ()
The sample value series for each lance 7 (same as the tap interval of the monkey filter 20) is expressed as (Xk). Similarly, the output wave of the differentiating circuit 42 (yk), the reference waveform generated by the reference waveform generation circuit 45 (r,), and the comparator 49
The error signal sound (ek) (ek=yk-rk) which is the output of
It is written as. That is, the error signal (ek) is stored in the error signal memory 46.

Next, (Xk) and (ek]k are read out from each of these waveform memories 41 and 46 using a clock of an appropriate frequency, and a correlation calculation is performed as shown below.As a result, the correlation range CP,
Q] is usually set to a value of approximately p==-2M and Q=2N.

The physical meaning of di is delay time 1T (T is tag interval)
This is the approximate size of the ghost.

On the other hand, the tap gain memory 48 stores the tap gain (c,) of each tap, and its initial value is 0o=l
tcM~c-4=O, cl-cN=0.

Each time the calculation of the first equation is involved for one i out of i = -IVi -H, the tap gain memory 48 calculates the tag gain c1'f! : Read, for this c, , n
ew==c-told-ad-12)+
After making the correction expressed as 1 1 (a is a positive minute value), tap gain memory 4 is applied again.
Return to 8. It is expressed by the equations (1) and (2).
All i (i=-M-N) between n and one field
The tap gain correction calculation circuit 47 executes this.

The above calculation is repeated each time a new reference waveform is received (that is, once per field).

By continuing this, the output waveform (yk) approaches the reference waveform (rk), and finally (C0) becomes a certain value (cho'
It converges to opt, but the output shape (yk) at this time is E=Σ(yk-rk)2 +++++・+ (3)
The residual error defined by k=J' is made small (see the above-mentioned literature).

If the tap gain correction is repeated according to equations (1) and (2) above, the tap gain will theoretically be (c,). 1
．． converges to. However, in reality, the tap gain is not necessarily (ci) because the frequency characteristics of the transpersal filter 2o are not ideal. After starting the sequential correction operation, (C1) changes toward (c1].1 until a certain time at the beginning of drawing, but as time passes beyond that, it changes to (C3).
may gradually diverge.

Also, in order to simplify the tap gain correction calculation circuit, CC11ne20-
IOld-ae0 song...Old-(4) A method of modifying the tap gain is also known, and is called the cellophoring method.

-d o In the case of the ossing method, even if the frequency characteristics of the transpersal filter are ideal, (C1) may diverge in principle depending on the shape of the input waveform (c).

In order to prevent such tap gain divergence, there is a method of stopping tap gain modification after a certain period of time. However, since both the fd and waiting times of the television signal (booth) change, when the ghost changes, a new ghost will occur unless the correction of the tap gain is restarted. Therefore, a method has been proposed in which the tap gain is modified intermittently. (Unexamined Japanese Patent Publication No. 56-104586 "Television Ghost Removal Device") In this case, an automatic isoelectric detector using an error signal related only to the sign as shown in Fig. 1 determines the change in ghost (waveform distortion). Therefore, after a certain period of time has elapsed after tap gain modification was stopped, tag gain modification was started regardless of changes in the ghost component. Therefore, there is a problem in that tap gain correction is started even though the ghost has not changed, or tap gain correction is not resumed even though the ghost has changed.

[Purpose of the invention]

An object of the present invention is to start or stop a tap gain correction operation in response to a change in waveform distortion of an output waveform in an automatic equalizer that corrects the tap gain of a transversal filter using an error signal related only to the sign. The object of the present invention is to provide an automatic equalizer that can perform the following steps.

[Summary of the invention]

The present invention performs cyclic addition of an error signal regarding only the sign of the error in the output waveform of a transpersal filter or a correlation value between this error signal and a waveform of a predetermined shape in a received signal by adding leakage in the direction toward zero. The change in waveform distortion of the output waveform is then determined from the cyclic addition value, and the tap gain correction operation is started or stopped in accordance with the determination result.

That is, for example, cyclic addition is performed while the tap gain correction operation is stopped, and when the waveform distortion increases beyond a certain value, the tap gain correction operation is restarted. Furthermore, when the tap gain is corrected by cyclic addition and the waveform value decreases below a certain value, the operation of the tap gain may be stopped.

[11th Effect] According to the present invention, since changes in waveform distortion of the output waveform can be easily determined from the error signal related only to the sign, equalization is performed using the optimal tap gain to follow the change in waveform distortion. It has the advantage of being flexible.

[Embodiments of the invention]

An embodiment of the present invention is shown in FIG. In FIG. 2, parts K having the same functions as those in FIG. 1 showing the structure of a conventional automatic isolator are given the same reference numerals and explanations of their operations will be omitted, and the added parts will be explained in detail.

That is, in FIG. 2, the configuration of FIG. 1 includes a reference value memory 50 for detecting waveform distortion, a cyclic addition value memory 51, a subtraction circuit 52, a determiner 53, a switching signal generator 54, an arithmetic circuit 55, and a switching circuit. 60.70 have been added. The operation will be explained step by step below. f,) 4 of the tap gain of the Lance Parsal filter 20
1- until a certain period of time has elapsed, the first switch 60 is switched off as shown in FIG. , terminals 61 and 62, and terminal 63 and 64, respectively. All terminals of the second switch 70 are open. Therefore, in this state, the automatic equalizer performs exactly the same operation as the automatic equalizer shown in FIG.

Next, after a certain period of time has elapsed, as shown in FIG. The gain is not modified. Therefore, a signal corresponding to the tap gain value in this case is output to the output of the transversal filter 20.
In the switch 70, terminals 71 and 73 and terminals 74 and 76 are coupled, respectively. In this 174th case, the reference value memory 50
For the value T, (f%0) held in , set the initial value to 0(T
The initial value of o is 1), and if the value of T,, old is corrected a certain number of times, as in By adding β in equation (6) and t) in equation (6), the correlation value d corresponding to the first tap is added cyclically.

Here, equation (5) is a method for providing a leak proportional to Ti, and equation (6) is a method for providing a constant leak regardless of T1. The calculation of d and the cyclic addition of di are performed by the arithmetic circuit 55, and the result of the cyclic addition is held in the reference value memory 50.

After that, as shown in FIG. 3(c), the switching signal generator 5
In the second switch 70, terminals 72 and 73 and terminals 75 and 76 are coupled, respectively, in response to a signal from 4. and,
The value U held in the cyclic addition value memory 5, (i lol 0) (
The initial value for C is 0 (U.

From the initial value of 1), a serial leak is added using the same method as Ti, and the correlation value d is added cyclically a certain number of times, and is stored in the cyclic added value memory 51.

Next, all terminals of the second switch 70 are left open, and finally the reference value memory 50 outputs the reference value T1, which is a value obtained by cyclically adding the correlation values corresponding to each tap, and cyclically adds the correlation values corresponding to each tap. The added value memory 5 outputs a value Ui obtained by cyclically adding the correlation values corresponding to each tap. A subtraction circuit 52 performs subtraction between T and U, and obtains a subtracted value g1 corresponding to each tag. Then, the determiner 53
A value X is calculated at - (-M, N are tap numbers). Here, if there is no change in waveform distortion, the difference g1 between T and Ul is the transversal filter 2
It is determined by the S/N of the output of 0 and the variations in each 7 and parameters of the circuit. Letting yl be a value determined by the variation in the parameter, a determination criterion Y of a constant value is set, and when X becomes Y or more, it can be determined that the waveform distortion has increased by a constant value or more. The effect of adding leakage in the cyclic addition of the correlation values di will be described below.

S will be described when cyclic addition is performed until the cyclic addition value reaches a predetermined maximum value, for example, the dynamic range of the memory. If the ν-th noise component corresponding to the i-th tap and added to the correlation value d1 is nl, then, for example, when adding a leak as shown in equation (5), TI2. +1=(1-β)T,, 10d, +n,, +,
−・・−−−−−・<9) is repeated to perform cyclic addition of leakage addition, and if the number of additions until the dynamic range of the memory is reached is 1, then the final The cyclic addition value Ti is as follows. Next, in formula 01, the root mean square of T is calculated as follows. ``Lj are unrelated to each other.Next, when leakage is eliminated, consider adding until the dynamic range of the memory is reached.If the number of additions is k k 2, the final cyclic addition value T, ke, and T, 2-rr, 2 to σn2...
1 Inasu・So◆. Here, (10, according to H formula, in the final state of the cyclic addition of correlation values, the magnitude of the signal components is equal, so the relationship tends to be, and the relationship is obtained. Then, comparing the noise power, (11
) + (141tαQ Using the relationship, it becomes.

From the αη formula, add the leakage amount β(
It can be seen that the noise power becomes smaller in relation to 0<β<1) and the number of additions kl. As a result, S
Since /N is improved, it becomes easier to determine changes in waveform distortion. Regarding equation (6), the S/N has a similar healing effect.

Now, if the values of equations (7) and (8) are determined, and the determiner 53 determines that the waveform distortion mentioned above has changed (increased), the signal from the switching signal generator 54 changes, and the first switching 60 enters the state shown in FIG. 3(,), and the tap gain correction operation is restarted. In addition, if X is smaller than Y (/J), it is determined that the waveform distortion has not changed, and the configuration shown in FIG. Repeat until it is determined that the waveform distortion has changed. In this way, intermittent operations are performed in response to changes in waveform distortion, and equalization operations are always performed with the optimum tap gain.

Note that the leak addition method is not limited to equations (5) and (6), and there are various other methods. For example, during cyclic addition, the leak may not be added the first fixed number of times, and then the leak may be added a fixed number of times, and any leak adding method may be used as long as it suppresses the influence of noise on the output signal.

Fig. 4 shows a second embodiment of the present invention, which is the same as the embodiment shown in Fig. 2 except that only the error signal is cyclically added with a leak, and the effect is also the same. .

FIG. 5 shows a third embodiment of the present invention in which the configuration of FIG. 2 is simplified, in which the reference value memory 50 and subtracter 52 in FIG. 2 are removed, and the switch 7θ is replaced with a switch. 80.

In FIG. 5, when the modification of the tap gain of the transpersal filter 20 is stopped, if the waveform distortion of the output waveform is small, the reference T1 (1#O) is all 0 (
The fact that it can be approximated as To=1) is used.

Also, the tap gain of the transversal filter 20 is changed for a certain period of time (in the positive state of i5, the terminal 61 of the switch 60
and 62, and terminals 63 and 64 are coupled, respectively. on the other hand,
All terminals of switch 80 are open. after that,
After a certain period of time has elapsed, all the terminals of the switch 60 are opened, and the switch 80 connects the switching signals from the switching signal generator 54 to the terminals 81 and 82 and the terminals 83 and 84, respectively. Then, after a certain period of time, the cyclic addition value memory 5)
In the same way as in Fig. 2, leakage is added, and the correlation value d
, are added cyclically. Then, the final cyclic addition value U1 is output to the determiner 53, and (-M, N are tap numbers) is calculated. Is this value Z the S/N and each node of the circuit? If you keep it above a certain value V determined by the variation of the parameter,
It is determined that the waveform distortion has changed, and a signal is sent from the determiner 53 to the switching signal generator 54.
1 and 62, 63 and 64 are combined respectively. Further, all terminals of the switch 8o are opened, and correction of the tap gain of the transversal filter 2o is restarted. When 2 is smaller than ■, the addition of d and the determination of the change in waveform distortion are repeated. 2) Similar to the embodiment shown in FIG. 2, intermittent operation is performed in response to waveform distortion.

Note that the present invention can be implemented with various other modifications.
For example, various processes of the tap gain modifying means can be performed by software using a microprocessor.

Further, the present invention is effective even when the transpersal filter has a feedback configuration, and the present invention is also effective regardless of the algorithm used to modify the tap gain of the transpersal filter.

Also, instead of the value X in equation (7), i=-M'
1=-M It is also possible.

Furthermore, when modifying the tap gain of the transparsal filter, leak is added every few times and the correlation value or error signal is cyclically added, and 2 of the α-peak equation is a constant value! :l, when the waveform distortion (ghost) decreases below a certain value, correction of the tap gain may be stopped.

For example, when using the flat part of the output waveform following the falling edge of the vertical synchronization signal as an error signal, the reference waveform (rk) in Fig. 2 is not necessary, and the output waveform is differentiated and a comparator is used to generate 1-bit data. The converted signal is used as the error signal (tuna), but the present invention is also effective when using such an error signal generation method.

[Brief explanation of drawings]

Fig. 1 is a block diagram of a conventional automatic equalizer, Fig. 2 is a block diagram showing an embodiment of an automatic equalizer of the present invention, and Figs. 3 (a), (b), and (c) are diagrams of the same embodiment. Figures 4 and 5 for explaining the operation are configuration diagrams showing other embodiments of the automatic equalizer of the present invention. 20... Transversal filter, 46... Error signal memory, 48... Tag gain memory, 49...
Comparator, 50...Reference value memory, 51...Cyclic addition value memory, 53...Determiner, 55...Arithmetic circuit for cyclic addition.

Claims (3)

[Claims] (1) A variable tap gain transversal filter to which a received signal in which a waveform of a predetermined shape periodically exists is input, and an error in the output waveform of this transversal filter corresponding to a waveform of a predetermined shape in the received signal. In the automatic equalizer, the automatic equalizer is fully equipped with a tap gain modifying means for successively modifying the tap gain of the transversal filter based on an error signal regarding only the sign of the transversal filter. cyclic addition means for cyclically adding a correlation value with a waveform of a predetermined shape in the received signal by adding a leak in a direction toward zero; and a waveform distortion of the output waveform from the cyclic addition value obtained by this means. 1. An automatic equalizer comprising: determination means for determining a change in the change in tap gain; and switching means for controlling start or stop of a tap gain correction operation in accordance with the determination result of the means. (2) The cyclic addition means performs cyclic addition when the tap gain correction operation is completely stopped, and the switching means corrects the tap gain when the determination means determines that the waveform distortion has increased beyond a certain value. The automatic equalizer according to claim 1, wherein the automatic equalizer restarts operation. (3) The cyclic addition means performs all the cyclic addition during the tap gain correction operation, and the switching means stops the tap gain correction operation when the determination means determines that the waveform distortion has decreased below a certain value. An automatic equalizer according to claim 1, characterized in that the automatic equalizer is: