JPH02159885A - Ghost removing device - Google Patents

Abstract

PURPOSE:To stably execute the removing operation of waveform distortion such as ghost without fail by providing an amplification factor setting means, which sets an amplification factor based on relation between an evaluation function and an evaluation function minimum value, and controlling a variable amplifier by this means. CONSTITUTION:A waveform distortion signal epsilonn' outputted from a synthesizing circuit 24 is supplied to an evaluation function setting circuit 19. In the circuit 19, during a waveform distortion detecting period, the square-average of signal block data used for setting weighing are obtained and defined as an evaluation function E. The obtained function E is supplied to a comparator 20 and compared with a minimum value Emin which is outputted from a minimum value setting circuit 28. The value of the function E is compared with a value betaEm, for which a constant value beta larger than 1 is multiplied to the value of the minimum value Emin. Then, after the condition of E<=betaEmin is confirmed, the value of an amplification factor (a) corresponding to the value of the minimum value Emin is set to an amplification factor setting circuit 29. To the circuit 29, a waveform distortion cumulative inversion epsilonn and a weighing value omega are also supplied from a cumulative adder circuit 27. Then, a variable amplifier 16 is controlled not to pass the signal epsilonn'.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a ghost removal device, which removes ghosts contained in input video signals in various video devices that receive TV video signals, typified by TV (television) receivers. The present invention relates to a device (ghost canceller) that removes ghosts and waveform distortions that occur in waves.

[Conventional technology]

A typical example of a conventional ghost removal device will be described with reference to FIG. FIG. 2 is a block diagram of the conventional ghost removal device 1, in which 11 is a filter section, 12 is a weighting setting circuit, 13 is a waveform extraction circuit, 14 is a peak detection circuit, 15 is an averaging circuit, and 57 is a weighting setting circuit. 18 is a reference waveform generation circuit, and 36 is an amplifier. With this configuration, it is attempted to remove waveform distortion such as ghosts in the video signal at the baseband.

Next, the operation of the ghost removal device 1 will be explained with reference to FIG. 3 and subsequent figures. The input video signal (Xnl) that comes in through the line filter 1 is supplied to the filter section 11.
1 (transfer function: aa(f)) and IIR filter 23 [FIR, filter 22 (transfer function: Gb(f)
)) and a subtractor 56,
The transfer function G(f) is G < f ) = Ga(f
)/(1+Gb(f))...---(1). The F'IR filters 21 and 22 are specifically constructed as shown in FIGS. 4(A) and 4(B), respectively. That is, the FIR filter 21 has a delay time T (for example, T=1/4f), as shown in FIG. 4(A).

C f5cLr3.58HH2) delay port 1ff142 is N
The delay blocks 25 connected to each other, and the N+1 weighting circuits 31 connected to the output terminals of each delay circuit 42,
The weighting consisting of the block 37 and each weighting are carried out by the circuit 3
1 and the main signal output via the line 43. When controlling the gain of the main signal, each weighting controls the gain aN of the circuit 31. , when removing the previous ghost or waveform distortion component preceding the main signal, each weighting is performed as times li! It is used to control the gains aQ to aN-1 of 831.

As shown in FIG. 4 [B], the FIR filter 22 in the IIR filter 23 has a delay circuit 42 with a delay time T of M±1.
The delay blocks 26 and each delay circuit 42 are connected to each other, and M+1 weighting circuits 32 are connected to the output terminals as shown in the figure. This is a transversal filter consisting of an addition/synthesis circuit 30 that adds and synthesizes the signals, and functions to remove delayed ghosts or waveform distortion components added later than the main signal.

The weighting values of the FIR filters 21 and 22 of the filter section 11 configured as described above are controlled by the setting circuit 12. The output (Ynl) of the filter section 11 is taken out as an output video signal from the line 12 and is also supplied to the waveform extraction circuit 13 at the same time.

Among the input video signals (Xn) applied from the line At, a signal train for one horizontal scanning period on which a reference signal for detecting waveform distortion such as a ghost is superimposed is shown in FIG. 5(8). Here, it is assumed that the potential of the blanking level is set to vl. When the weighting value of the filter section 11 is 0, the input video signal (Xn) is taken out from the IT exit line 12 as the output video signal (Ynl).

Note that signal time delays due to processing time that occur in these circuit block diagrams are omitted for convenience of explanation, but will be treated similarly in the following explanation.

In the signal waveform of Fig. 5 (A), the signal section used to detect waveform distortion such as ghosting with reference to the reference signal @ is Td.
This is expanded in the time axis direction and shown in the same figure (B),
The waveform extraction circuit 13 functions to extract the signal in this Td interval. And by clamping the point P slightly before the reference signal @ to zero potential.

It also has the function of setting the blanking level to zero potential. When a ghost or noise is mixed into the signal shown in (8) of the same figure, the waveform becomes as shown in (C) of the same figure. This signal is supplied to the peak detection circuit 14 at the next stage. Peak detection circuit 1
4 is a comparator that compares the level of the input signal with a reference value, a counter that measures the elapsed time from the start point of the section Td,
and a memory (latch) circuit that stores the count value of the peak position, etc., and detects the time at which the maximum value (peak value) in the waveform distortion detection section Td occurs, that is, the peak position. The output of the peak detection circuit 14 is sent to the averaging circuit 1 in the next stage.
5. This averaging circuit 15 is composed of a memory circuit for storing the Td interval, an adder, etc., and the waveform shown in FIG. The signals are synchronously added a predetermined number of times with reference to the peak position (which corresponds to the peak position of the reference signal) and are averaged. Through such averaging processing, noise components that have no correlation with the signal are sufficiently suppressed, and a signal (Yn'1) as shown in (0) in the figure is obtained.The reference signal @ in (B) in the figure is If the original spectral distribution is flat as shown in FIG. 6(^), the spectral distribution of the signal containing the ghost g as shown in FIG. 5(D) will be 0 to 6 as shown in FIG. This characteristic causes fluctuations between 4 MHz, which represents a waveform distortion component due to the ghost g, and this signal (Yn'l) is supplied to the positive input terminal of the subtracter 57.

The reference waveform generation circuit 18 synchronizes with the peak position detected by the peak detection circuit 14 and generates a reference waveform as shown in the waveform diagram of FIG. 5(E) and a reference having an ideal spectral distribution as shown in FIG. 6(^). A waveform signal (γn) is generated. The reference waveform signal (γn) is subtracted from the signal (Yn'!) from the averaging circuit 15 by a subtracter 57 to obtain a waveform distortion signal (ε
. ) (see FIG. 5(F)) is obtained. Next stage amplifier 36
is a signal (ε.') = (αε.) (where α is 1) obtained by multiplying the waveform distortion signal (ε.) by a predetermined multiplication factor α.
ε. ') (see (G) in the same figure) is obtained. The weighting setting circuit 12 detects the amplitude ratio between the peak position, the time width (distance) of the waveform distortion, and the peak value from the supplied waveform distortion signal (ε,'), and calculates the weighting so that the distortion is minimized. Weighting (ω
n) and weighting the value to configure the FIR filter 2122 of the filter unit 11 is performed in blocks 37 and 38.
Each weighting is set in circuits 31 and 32. This weighting requires an arithmetic function in the setting circuit 12, so a microprocessor or microcomputer (n+1cro
ter), etc., and the weighting values are repeatedly calculated and set based on the following equation.

(ωn1h=!ωn) −(ε.'lh...
(2) However, when 1≦, 5m (m: number of processing until convergence): order in the data string (ωnlk+ is weighted by the calculated value (the same figure (H
) Note that the calculated weighting value (ωn)b is set to the corresponding tap in the transversal filter that constitutes the filter section 11.

[Problem to be solved by the invention]

As described above, the method of sequentially updating the detected waveform distortion components with opposite polarity is a theoretically correct method, and the algorithm used in such conventional equipment is ZF (Zera Fo
This is a preferable control method that has the advantage of fast convergence time because control is performed in proportion to the amplitude of waveform distortion.

Moreover, as described above, by weighting the filter section 11 so as to minimize distortion, a video signal from which waveform distortion such as ghosts has been removed should be obtained, but the sequential weighting is If the operation is unstable and falls to a false stable point, or if a disturbance such as noise enters when it approaches the optimal value to a certain extent, it will deviate from the optimal value and move toward another stable point, or the IIR filter 23 will diverge. There was a problem in that it was headed towards the end.

Furthermore, when the root mean square value of the waveform distortion component data in the detection area Td period is viewed as an evaluation function, after the start of control, the value of the evaluation function gradually decreases as the number of processing times for the setting increases, and After reaching the minimum value, it oscillates in the opposite direction and moves away from this minimum value, oscillating slightly around a value different from the original arrival point, and the value of the evaluation function gradually increases and diverges. I often did this. As a result, there were major problems in terms of stability, such as a serious problem in the ghost removal operation, and the disadvantage was that it was difficult to put it into practical use.

[Means to solve the problem]

The ghost removal device of the present invention includes a waveform extraction means for extracting or waveform converting a signal of a predetermined fixed period including a pulsed reference signal for ghost detection from a television video signal to obtain a first signal; a reference waveform generation means for generating a reference waveform signal in synchronization with the signal; a subtraction means for obtaining a second signal by taking the difference between the first signal and the reference waveform signal; and cumulative addition of the second signal. an accumulative averaging means for obtaining a third signal by averaging processing; a synthesizing means for obtaining a fourth signal by adding and synthesizing the obtained second signal and third signal at an appropriate ratio; means for obtaining an evaluation function by calculating the mean square value or the average value of the sum of absolute values of the fourth signal; and a method for determining the relationship between the obtained evaluation function and the minimum evaluation function value obtained before a predetermined time. weighting set in the filter means by controlling the gain of a variable amplifier that controls the value of the fourth signal by the amplification factor setting means; The system is configured so that the value of is updated sequentially, which solves the above problem.

〔Example〕

In the ghost removal device of the present invention, not only the weighting is simply corrected sequentially, but also an evaluation function for evaluating the amount of waveform distortion components and a cumulative average value of the waveform distortion signal are already set. For weighting, values, etc. are introduced as parameters for setting, and under conditions where the evaluation function decreases, weighting makes the setting processing sequential, and weighting is cumulative as a condition for setting processing. The ghost removal operation can be stabilized by performing the weighting setting process only when the average value or the weighting that has already been set is a significant value. An embodiment of the ghost removal device of the present invention will be described with reference to FIG. 1 and the like.

FIG. 1 is a block diagram of a ghost removal device 10 of the present invention. In this diagram, the same components as those of the conventional device shown in FIG. As is clear from the comparison, in the ghost removal device 10 of the present invention, the amplifier 3 in the conventional device 1 is
6 is changed to a variable amplifier 16, and its amplification factor can be controlled externally, and an evaluation function setting circuit 19 is newly introduced to apply control in the direction of decreasing the value of the evaluation function, and to control the value of the waveform distortion signal. Calculate the cumulative average value and add it to the current waveform distortion signal at an appropriate ratio.The weighting that has already been set is the value, and only for data for which the cumulative average value of the waveform distortion signal has a significant value. The most important feature of weighting is that it is configured to operate the setting circuit 12. In order to realize such operation, variable amplifier 16. Evaluation function setting circuit 1
Comparator 20 for 9 songs. Minimum value setting circuit 28. Amplification factor setting circuit 29. Cumulative addition circuit 27. and a combining circuit 24, which are connected as shown in the first figure.

Next, the specific operation of the ghost removal device 10 of the present invention will be described with reference to the waveform diagram of FIG. 5 and the like. NTSC supplied from line AI to filter section 11
When the weighting value of the filter section 11 is 0, the input video signal fXn) is output as is from the output line 12 as the output video signal (Yn). This signal represents one horizontal scanning period in which a reference signal for detecting waveform distortion such as a ghost, which is multiplexed in the vertical blanking period in the video signal, exists.

In this signal (Ynl), the waveform distortion detection section Td such as a ghost is detected by the waveform extraction circuit 13 as shown in FIG. 5(B) or (C).
). This waveform extraction circuit 13 also has a clamping function, and has the function of fixing the reference level (blanking level) to zero potential as shown in FIG. 5(B). In the sixth-stage averaging circuit 15 in which the peak position of the reference signal @ in the output waveform of the waveform extraction circuit 13 is detected by the peak detection circuit 14, noise as shown in FIG. The signals are averaged based on the peak position, and uncorrelated noise is sufficiently suppressed.
A signal (Ynl) as shown in FIG. 3(D) is obtained and output.

The reference waveform generation circuit 18 generates a reference waveform (reference signal γn, see (E) in the same figure) whose phase is synchronized with the peak position in the signal (Ynl).This reference signal (Yn) and the averaging cycle FI@ The signal from 15 (Ynl and subtractor 5
7 and subtracted there to obtain a waveform distortion signal (crystal) (see figure (n)), which is supplied to the cumulative addition circuit 27 and the synthesis circuit 24.The cumulative addition circuit 27 is used during the waveform distortion detection period. It is composed of a memory circuit and an adder circuit for storing the signal (waveform) of Td (see Figure 8), and the incoming waveform distortion signal (E
n) with the peak position as a reference, and produces a signal (waveform distortion cumulative sum value) as shown in (1) in the same figure.
The 0 synthesis circuit 24 outputting (a) converts the waveform distortion signal (ε, ) from the subtracter 57 and the cumulative waveform distortion I En from the cumulative addition circuit 27 into the following equation: (εn' 1=(pε). +
(ip) En l・・・・・・・・・(3)
(However, p is a constant). This composite output (ε,') is multiplied by a predetermined multiplication factor α (however, amplification factor: α multiplied by 1) in the variable amplifier 16, and then (εn″
)=(αεn′ l ), which is a waveform distortion signal (ε.″) (see (J) in the same figure), and the weighting is output to the setting circuit 12.

The setting circuit 12 weights the input waveform distortion signal (ε).

″) to the peak position and waveform distortion (distance)
and the amplitude ratio with the peak value is detected, a weighting value that reduces waveform distortion is calculated, and the weighting of the filter unit 11 is performed by blocks 37 and 38, and each weighting is performed by a circuit 31. The calculation and setting of the zero weighting value, which serves to set the value to 32, is carried out iteratively based on the following equation, for example:

(ωn)b=(ωn) −(ε.−1k・・・・・・
(4) However, (ωnlh) is the calculated value for the second processing (see (H) in the same figure), and this is
FIR filters 21 and 22 constituting the corresponding taps.

Next, the evaluation function processing system, which is one of the greatest features of the present invention, will be explained. The waveform distortion signal (εn') output from the synthesis circuit 24 is supplied to the evaluation function setting circuit 19. In this evaluation function setting circuit 19, the waveform distortion detection period Td
In the above, the weighting is actually done by the signal section Te(
Root mean square value of Ne data (see Figure 5 (F)) B = (Σε.'ri) / N e ......
・・・・・・・・・・・・・・・(5) is obtained, and this is used as the evaluation function E. This evaluation function setting circuit 19 can be configured with a multiplier for performing square calculations, a feedback circuit having a latch circuit that generates a delay of one tarok in the loop, etc. (see Fig. 8). The evaluation function E obtained is supplied to the comparator 20 at the next stage, where the minimum value B
+m (the minimum value during the processing up to that point). Then, by comparing the value βElln, which is the value of the minimum value E1 multiplied by a constant value β larger than 1, and the value of the evaluation function E, the condition of E≦βE≙ (where β is a constant of 1) is satisfied. After this is confirmed, the value of the amplification factor α (α×1) corresponding to the value of Ellll is set in the two amplification factor setting circuits.

This amplification factor α is set as the amplification factor in the variable amplifier 16, and is obtained by multiplying the input waveform distortion signal (ε.′) by α (ε.“
) = (αε, ′). This amplification factor setting circuit 2
9 shows the waveform distortion cumulative addition average value from the cumulative addition circuit 27 (
G), and the already set weighting values (ωnli) are also supplied, and the amplification factor α is maintained at that value only when these values have a significant value compared to the predetermined value. ,
At other times, it sets the amplification factor to 0 and controls the variable amplifier 16 so that the waveform distortion signal (εn' does not pass through. Therefore, the amplification factor setting circuit 29 also serves as a switch. .

On the other hand, the comparator 20 compares the evaluation function E from the evaluation function setting circuit 19 and the minimum value ETLIn from the minimum value setting circuit 28, and the value of the evaluation function E at that time is the minimum value EIII.
When the value is smaller than n, the minimum value Eam set in the minimum value setting circuit 28 is replaced with the value of the evaluation function E, and that value is newly set as the minimum value E1.1lrl. This minimum value setting circuit 28 sets the comparator 2 when E < E atn.
Consists of a latch circuit that temporarily stores the evaluation function at the end of the pulse-like confirmation signal applied from 0 (8th
(see figure).

FIG. 7 is a characteristic diagram showing an example of the evaluation function E in the evaluation function setting circuit 19 constituting the apparatus of the present invention. The horizontal axis represents the number of times the evaluation function E is updated.
・The mark represents the value of E, the broken line represents the trajectory of EllIn, -
The dotted chain line is the value of βEil+1 (in this example, β”Fl,
5) respectively. In this figure, k=
As shown in the figure, the initial value of ENIll that was set at the time of o satisfies the condition E<Eyy until =6, and E+m is updated for each process (therefore, the straight line connecting the From this figure, it can be seen that the operation enters a stable state from 9 onwards until 21, and E+lkl is not updated. Furthermore, the characteristics of βE+m are as follows:
It is β times (constant) the characteristic of the minimum value B11k'l.

The amplification factor α of the variable amplifier 16 is set by the amplification factor setting circuit 29, but when the value of E is less than βEILln in FIG. Weighting is calculated value (each absolute value of ωnlh is a significant value (!Enl>Pa, Iωn1) Pb, however, Pa, P
b is a positive constant), the effective amplification factor α is set, but as in the case of k=11°12.16.17, E
> When βXEILln, the amplification factor α becomes O, and the weighting is configured so that the value (ωn) is not updated. Therefore, the amplification factor setting circuit 29 also serves as a switch. By updating the weighting value (ωnlb) based on such an evaluation function value E, it is possible to avoid troubles that tend to occur with conventional equipment, such as entering the wrong stable point due to the incorporation of noise or abnormal signals. There is no divergence or divergence of operation, and control is constantly directed toward a true stable point, making it possible to stabilize and ensure the operation of removing waveform distortions such as ghosts.

FIG. 8 shows the main components of the device of the present invention, namely the variable amplifier 1.
6. Evaluation function setting circuit 19. Comparator 20° synthesis circuit 24
．． Cumulative addition circuit 27. Minimum value setting circuit 28. 3 is a specific circuit diagram of an amplification factor setting circuit 29 and the like. FIG. Below is the function of the circuit based on this diagram.

Explain the operation.

The waveform distortion signal (ε,) from the subtracter 57 supplied via the line 16 is first outputted from the composite diagram! ! After being given a nine-fold gain by the amplifier 52 constituting the adder 62
The output signal (G) of the cumulative addition circuit 27 is multiplied by (1-p) by the amplifier 53 and added here, (ε.')=(pεn+(1-p)sn)...・・・・・・
......(6) The signal output is as follows.

Note that the output (ε.) of the subtracter 57 is output from the cumulative adder circuit 27.
It is also supplied to a subtracter 58 which constitutes an amplifier 55.
The output from the input signal (ε, ) is subtracted from the input signal (ε, ) and stored in a memory (storage unit) 66, and is output to the amplifier 53 and two amplification factor setting circuits. The memory 66 has a capacity to store data for the waveform distortion detection period Td.
The output of this memory 66, which serves to repeat the read and write operations each time a period is taken, for example, every field or frame, is supplied to a subtractor 59, where the output (ε) of the subtractor 57 is fed to a subtractor 59. ), a gain multiplied by δ is applied by the next stage amplifier 55, and the result is output to the negative input terminal of the subtracter 58. The cumulative addition circuit 27 having such a configuration essentially functions as a noise reduction circuit, and also calculates the cumulative addition average value (E) of the waveform distortion signal (ε.).
n). This output signal (En) is supplied to the amplifier 53 and the A terminal of the amplification factor setting circuit 29.

On the other hand, the output (ε.') of the synthesis circuit is supplied to the delay circuit 49, and the delay time generated in the evaluation function processing system (evaluation function setting circuit 19, comparator 20
．． Minimum value setting circuit 28. The signal is delayed for a predetermined time to compensate for the time required for signal processing in the amplification factor setting circuit 29. Note that the delay time here does not change the form of the input signal, so for convenience of explanation, the output signal will also be written as "ε.'" like the input signal. The output of the delay circuit 49 is sent to a data selector 61 that constitutes the variable amplifier 16.
Enter it in A9 child. Based on the control signal supplied to the C terminal, the data selector 61 outputs either the A terminal input signal or the B@ terminal potential held at zero potential from the D@ terminal. 51 is an amplifier with a gain α, (ε.

′), the output signal is (ε,″)
=(αε.') is obtained, the weighting is transferred to the setting circuit 12, and the weighting is used for control.

Next, the evaluation function will be explained. The waveform distortion signal (εn') from 24 is also supplied to the multiplier 41 constituting the evaluation function setting diagram #119, where the waveform distortion squared signal ((εn') is supplied.

)2) is created. Then, the adder 63 at the next stage and the delay line 43 (1 clock delay time T (T=1/4f, . . . ) included in the feedback loop) are generated.

f → 3.58H1lZ)), the waveform distortion squared signal (
(εn')' The integrated value of l is determined. This integrated value is calculated by the attenuator 65 at the next stage, which is the reciprocal of the integrated number Ne (1/N
e) and the arithmetic mean is taken. The evaluation function E obtained in this way is expressed by the following formula E=Σ(ε,')2/Ne ・・・・・・・・・・・・
It will be given by (7). It should be noted that time adjustments in the operation start timing, initialization, operation end timing, etc. of the integration circuit 45 are omitted for the sake of explanation (the same will be treated hereinafter).

The evaluation function E obtained in the above manner is calculated by the comparator 20.
is supplied to The comparator 46 calculates the value of the evaluation function E supplied to the AyfA child by multiplying the evaluation function minimum value E+lIn by β (>1) by the amplifier 54, and then converting the value to the β value supplied to the B terminal.
Compared with E[n, when E≦βE+lin, the “H” level is set to the C terminal, and in other cases, “L” is set and transferred to the B terminal of the amplification factor setting circuit 29. The determination output from the C terminal of the comparator 46 is set after the evaluation function E is formed by integration processing, but it is updated every time the evaluation function E is updated for each field or frame. It turns out.

In addition, in the comparator 47, the evaluation function E from the attenuator 65 is supplied to its input terminal, and the evaluation function is supplied from the D@ terminal of the minimum value setting circuit 28 to the B#A terminal G of the comparator 47. The function is to compare the minimum value Ellkl and output a pulse-like control signal representing the state from C@ only when the condition E<BPn is satisfied. The minimum value setting circuit 28 is a latch circuit having a so-called preset terminal.

The C terminal is a preset input terminal, and at the same time as the processing operation of the ghost canceller starts, a preset pulse is supplied from line 18 as an initial setting, and the maximum value is set to the output terminal. The E1 value is .

Further, Bf@ is a data input terminal of the evaluation function E, and A@ is a tarok terminal. From comparator 47 E < E +t
When the control pulse signal applied when the condition ln is satisfied ends, the value of the evaluation function E supplied from the B terminal is taken into the minimum value setting circuit 28 and becomes E ain, and the minimum value becomes a new value. Replaced. The output E aln from the D terminal thus obtained is supplied to the comparator 47 and the amplifier 54 as described above. In this way, by using the evaluation function E as a parameter and controlling the evaluation function so that it always decreases, the device of the present invention attempts to stabilize its operation.

Next, the function and operation of the amplification factor setting circuit 29 will be explained. The A@ terminal of the amplification factor setting circuit 29 is supplied with the waveform distortion cumulative average value (gnl) from the cumulative addition circuit 27 as described above.The C terminal is supplied with the weighting setting via the line 19. The weighting that has already been set from the circuit 12 is the value (ωn
The code information of ib is supplied to the B terminal, and if the criterion condition E≦βE filn of the evaluation function from the comparator 20 is satisfied, “H” (High), otherwise “L”
(Low)" information is added. This amplification factor setting circuit 29 is composed of, for example, a comparator, and outputs "H" only when the A and C terminal inputs are significant values and the B terminal input is "H". , otherwise, the output signal of the amplification factor setting circuit 29, which operates to output "shi" from the D terminal, is supplied to the C terminal of the data selector 61.

From the D terminal of the data selector 61, when this C terminal input signal is "H", the A terminal input (ε.') is selected,
When the C terminal input signal is "L", the zero potential input to the B terminal is selected and output.

The above is a detailed explanation of a specific circuit example regarding the control system for the evaluation function E shown in FIG. In this way, the device of the present invention attempts to stabilize the operation by using the evaluation function E as a parameter and controlling it in the direction of constant decrease. .

In addition, in the above explanation, the square value of the waveform distortion signal is used as the evaluation function, but the same effect can be obtained by using, for example, the absolute value of the waveform distortion signal (Σlεn l )/N eE. I can give it to you. Furthermore, the reference signal for detecting waveform distortion is not limited to a pulse-like waveform set in the middle of the horizontal scanning period as shown in Fig. 5 (8); It is also effective in the device of the present invention to use a pulse signal extracted by waveform conversion, such as a waveform obtained by differentiating the falling edge of a bar waveform or the leading edge of a vertical synchronization signal. The present invention also works effectively with an algorithm that obtains a pulsed waveform from such different types of waveforms by waveform conversion and removes ghosts as a reference signal.

〔effect〕

As mentioned above, in the ghost removal device of the present invention, the evaluation function E is introduced to proceed with the processing while constantly monitoring whether or not the ghost removal operation is heading in the convergence direction. Since the weighting that has already been set can be set only when the value takes a significant value, it is possible to set the weighting only when the value takes a significant value. By eliminating the problems (that were not found), masking unnecessary noise, etc., and proceeding with the processing while looking at the correlation with the processing progress up to that point,
The noise resistance and stability have been significantly improved, and the waveform distortion signal (ε
, ) and the waveform distortion cumulative addition signal (En), so-called ZF (2ero Forcing
) method, it has the advantages of being able to compensate for the lack of sensitivity that tends to occur with the method, improving the removal accuracy, and being able to obtain high-quality video signals without waveform distortion.

[Brief explanation of the drawing]

1 and 2 are block diagrams of the present invention and the conventional ghost removal device, respectively, FIG. 3 is a specific block diagram of the filter section in the present invention and the conventional device, and FIG. 4 (^),
(B) is a circuit diagram of each FIR filter constituting the filter section, FIGS. A)
, (B) is a spectrum diagram for explaining the principle of signal processing in the present invention and the conventional device, FIG. 7 is a characteristic diagram showing an example of the evaluation function of the evaluation function setting circuit constituting the device of the present invention, and FIG.
The figure is a specific circuit configuration diagram of the main components of the device of the present invention. 10... Ghost removal device, 11... Filter section,
12... Weighting setting circuit, 13... Waveform extraction circuit, 14... Peak detection circuit, 15... Adding average circuit, 16... Variable amplifier, 18... Reference waveform generation circuit, 19...Evaluation function setting circuit, 20.46.47...
・Comparator, 21.22...FIR filter, 23...
・IIR filter, 24... synthesis circuit, 25.26・
... Delay block, 27... Cumulative addition circuit, 28...
・Minimum value setting circuit, 29... Amplification factor setting circuit, 30.
...addition synthesis circuit, 31.32...weighting is circuit, 3
7.38... Weighting is a block, 41... Multiplier,
42.43...Delay circuit, 45...Integrator circuit, 51
~55...Amplifier, 56-59...Subtractor, 61.
...Data selector, 62.63...Adder, 65.
...Attenuator, 66...Memory. Patent applicant: Japan Victor Co., Ltd. Representative: Kunio Umiki

Claims (1)

[Claims] A ghost removal device having a filter means for removing waveform distortion such as a ghost from a television video signal, which extracts or waveforms a predetermined period of signal including a pulsed reference signal for ghost detection from the television video signal. a waveform extraction means for converting the signal into a first signal;
a reference waveform generating means for generating a reference waveform signal in synchronization with the signal of the first signal; a subtraction means for obtaining a second signal by taking the difference between the first signal and the reference waveform signal; an accumulative averaging means for obtaining a third signal by cumulative averaging processing; a synthesizing means for obtaining a fourth signal by adding and synthesizing the second signal and the third signal at an appropriate ratio; Means for obtaining an evaluation function by obtaining the mean square value or the average value of the sum of absolute values of the fourth signal, and the relationship between the obtained evaluation function and the minimum value of the evaluation function obtained before a predetermined time. and an amplification factor setting means for setting an amplification factor based on the amplification factor setting means, and the weighting set in the filter means by controlling the gain of a variable amplifier that controls the value of the fourth signal by the amplification factor setting means. A ghost removal device characterized in that it is configured to sequentially update the value of.