JPH0290871A - Digitized ghost removing device - Google Patents

Abstract

PURPOSE:To reduce the variation of tap gain and a residual ghost by providing plural coefficient multipliers for which coefficients corresponding to the respective sizes of ghost distortion are set at the output sides of respective transversal filters. CONSTITUTION:A negate signal generating part 17 is provided with variable delay circuits 181, 182,...18n to which a signal from a selector switch 12 is supplied in parallel and whose output signals are supplied to the transversal filters 191, 192,...19n respectively. Then, both the delay time and the coefficient of each delay circuit and each transversal filter are calculated and set by an arithmetic operation control part 14 according to the delay quantity or the size, etc., of a ghost. Then, the output signal from each transversal filter passes through the coefficient multiplier 211, 212,...21n, and is summed by an adder 20, and becomes a ghost negate signal, and is subtracted from a video signal containing the ghost by a subtracter 13. Thus, the variation of the tap gain and the residual ghost are reduced.

Description

[Detailed Description of the Invention] [Object of the Invention] (Industrial Field of Application) The present invention relates to a ghost removal device for automatically removing television ghosts, and particularly to a ghost removal device for automatically removing television ghosts. The present invention relates to a digital ghost removal device for removing ghosts.

(Prior Art) A device for automatically removing television ghosts using a digital equivalent circuit is disclosed in, for example, Japanese Patent Application Laid-open No. 15217-1983.
As shown in Japanese Patent Application No. 1, it has been known for a long time.

FIG. 8 shows an example of such a conventional digital ghost removal device, in which a digital video signal containing a ghost is input to the input terminal 11, and this input video signal is transferred to the feedforward switch of the changeover switch 12. (
FF), and the addition input section of the subtracter 13,'
The signal is supplied to the a calculation control section 14 and the timing generation section 15, respectively. The timing generator 15 generates a clock signal (CK) necessary for each part of this device.
The frequency of the clock signal is generally 3 fsc or 4 fsc (however, f sc is the color subcarrier frequency - 3,579
545MHz) is selected.

The output signal from the subtractor 13 is sent to the output terminal 16 of this device.
It is also supplied to the feedback (FB) input of the changeover switch 12. This changeover switch 12 selects the input video signal that becomes the FF input or the FF input.
This selects one of the output signals inputted by the switching circuit 12, and the signal selected by the switching circuit 12
7.

The cancellation signal generating section 17 includes n variable delay circuits 181 to +
8n, and the signal selected by the changeover switch 12 is supplied in parallel to the delay circuits 181 to 18n. The variable delay circuits 181 to 18n are each constructed of a combination of registers or a RAM, and the delay amount of each of the delay circuits 181 to 111n is set by the arithmetic control unit 14 in accordance with the position of the boss. be done.

The output signals from each of the variable delay circuits 181 to 18n are transmitted to transversal filters (TF) 191 to 191 each having the same configuration of (2lc +1) taps.
1 is input. Then, the output signal from the transversal filters 191-19n is the addition 2: '320
The ghost cancellation signal is then added to the subtraction side terminal of the subtracter 13, and this cancellation signal is added from the input video signal including the coast. The video signal from which the ghost has been removed is output from the subtracter 13 and output from the output terminal 1G.

The lower limit of the number of taps (2 + 1) for each of the transversal fills 191 to 19n is determined by the number required to eliminate a single ghost. Further, the upper limit is determined by the number of taps that can be integrated into one chip, and here, as an example, the number of taps is -21c +1-21, that is, "k=10".

Each of the transversal filters 191 to 19n includes (2k + 1) coefficient units, and is configured in an input weighted type in which input signals are supplied to these coefficient units in parallel. . Then, the coefficient multipliers each have CI (i-1, 2,...2 k + 1
), and the output from these coefficient multipliers is T(T-1/
The signals are synthesized and output by a latch circuit and an adder at intervals (clock frequency).

FIG. 9 shows the state of the coefficient multiplier 19a constituting this transversal filter.
It is composed of multipliers using M or random logic. The number of bits of the coefficient Ct of this coefficient unit 19a is (
a) As shown in the figure, if 0 to 1 are n bits and the dynamic range is -1 to 1, the LSB becomes -2-71 as shown in (b) of the figure.

In order to calculate the tap gain fluctuation and remaining ghost that are allowed in practice, i'-#, the number of bits of the coefficient is n, and 10
is known to be necessary (Haka Hiroyuki Iga, “Effect of Noise on Ghost Cancellers”, Journal of the 1980 National Conference, 30th Anniversary of the Foundation of the Television Society, p. 138).
, 1980). Increasing this number of bits will make it more stable and reduce the amount of residual ghosts, but the circuit scale of the coefficient unit is very large compared to the other circuit scales of the transversal filter. , the number of bits of a coefficient becomes almost proportional to its cost, making it difficult to practically increase the coefficient.

Furthermore, as shown in Japanese Patent Application Laid-Open No. 59-225681, it has been considered to insert a sampling means into the output of the transversal filter, but this alone can reduce the dynamic range of all taps, in other words, the tap coefficients. You can only change the equivalent number of bits to a total number of taps.

(Problems to be Solved by the Invention) This invention has been developed in view of the above points, and has been developed in such a way that the number of bits of all the coefficient units of the transversal filter is substantially and equivalently the same. Another object of the present invention is to provide a digital ghost removal device that can equivalently increase the number of bits of an optimal coefficient in response to ghost conditions, and can effectively reduce tap gain fluctuations and residual ghosts. It is something to do.

[Structure of the Invention] (Means for Solving the Problems) In the digital ghost removal device according to the present invention, an input video signal is processed by a plurality of delays each having a delay time set corresponding to the location of a ghost. The output signal from each of the delay circuits is also supplied to a plurality of transversal filters, and the output side of each of the transversal filters has a plurality of coefficients set corresponding to the magnitude of each ghost distortion. This is to set the coefficient unit of .

(Function) In the ghost removal device described above, the coefficients of the coefficient unit are controlled and set according to the size of the generated ghost. For example, the smaller the ghost, the more the output of the transversal filter is reduced. You will be able to do it. That is,
The coefficient dynamic range of the coefficient multiplier is widely utilized and the number of bits of the coefficient is equivalently increased, thereby reducing tap gain fluctuations and residual ghosts.

(Example) Hereinafter, one embodiment of the present invention will be described with reference to the drawings.

FIG. 1 shows its configuration, in which the digitized video signal input from the input terminal port is transferred to the selector switch 12.
The FB large input signal is supplied to the subtracter 13, and is also supplied to the addition side terminal of the subtracter 13.

This input video signal is further supplied to an arithmetic control section 14 and a timing section 15, and the timing section generates a clock signal (CIC) used in this device and a control timing signal necessary for the control section 14. The output signal from the subtracter [3 is used as a feedback (FF) input of the changeover switch 12, is supplied to the output terminal IO, and is further used as another input of the operation control section 14. be done.

The changeover switch 12 selects either the input video signal input to the FB and FF or the output signal from the subtracter 13, and this selected signal is transmitted to the cancellation signal generator 1.
7. This cancellation signal generating section 17 includes a plurality of variable delay circuits 181.1g2, which are supplied with signals from the changeover switch 12 in parallel as in the conventional case.
... 18n, and this delay circuit fill
, 182, . . . 18n, the output signals from each of the transversal filters 191, 192
, ... 19n, respectively. Then, delay circuits 181, [12, ... Ign, and transversal filters 191, 192, ... 19n
Both the delay time and the coefficient are controlled and set by the calculation control section 14.

In this embodiment, the output signals from the plurality of transversal filters 191, 192, .
・The output signals from each of the coefficient units 211, 212, ..., 21n are added and combined by the adder 20, and then sent to the subtracter
0 (supplied) on the subtraction side of .

FIG. 2 shows an example of the configuration of the transversal filter 19 used in the above device. This transversal filter is called an input weighted type, and has (2k +1) coefficient units 501, 502, . . . Equipped with... Then, the outputs from each of the coefficient units 501, 502, . The outputs of 512, . . . are output for T seconds (T-
1/clock frequency) interval.

Here, the coefficient units 501, 502, . . .
...2 k +1) times. This coefficient CI is determined by the calculation control section 14, and this coefficient CI
is stored in the coefficient memory 53.

FIG. 3(a) shows the pulse response state of a video signal containing ghosts. Such a pulse response is obtained from a vertical sync signal difference and a ghost cancellation reference signal (OCR).

In this pulse response, three ghosts gl, g2
, g3 exist, and the delay time, U/D ratio, and phase of the first ghost g1 are 25 T, -6 dB, and 0', respectively. Further, it is assumed that the delay times of the second ghost g2 and the third ghost g3 are 75T and 110T, respectively, the U/D ratios are -12dll and -18dl'3°, and the phases are 180m and 90''.

The arithmetic control unit 14 obtains the pulse response as described above from the input video signal, calculates the delay amount, size, etc. of this ghost, and sets the delay amount of the variable delay circuit 181 to (25-k). )T, and the delay amounts of variable delay circuits 182 and 183 are set to (75-k)T and (75-k)T, respectively.
110-k) Set to T.

In this way, the variable delay circuit till, 182,...
By setting the delay amounts of . . . , the transversal filters 191, 192, .
2, remove g3 (25-1c)T, (75-1()
T, (110-1()) are controlled to generate cancellation signals starting from T. Means for setting the coefficients of the transversal filter for performing such control is a calculation using an input video signal. This is carried out by the control unit 14, and its control method is described, for example, in the literature (Junzo Murakami et al. "Digitalized Ghost Automatic Eraser" IEICE Technical Research Report EMCJ78-37.1978 1).
Since it is well known in this technical field, its explanation will be omitted here.

And this transversal filter 1911192
,... The output signal from 19n is sent to the coefficient multiplier 211.2.
12, ... 2 inches, are added in an adder 20 to become a ghost canceling signal, and are subtracted from the video signal containing the ghost in a subtracter 13, and the video signal from which the ghost has been removed becomes the output signal. It is something that is exposed as H.

As described above, ghosts gl, g2, and g3 are associated with transversal filters 191, 192, and 193, respectively. ghost gt here
is -6dB, 0', so its amplitude value is equal to the main signal "
When it is 1", it becomes 0.5". Therefore, the coefficient of the coefficient unit 211 to which the output of the transversal filter 191 corresponding to this ghost is supplied is set to "0.5".

Transversal filter 19 in such a state
The relationship between the coefficient multiplier 501 and the coefficient multiplier 211 is as shown in FIG. 4(a). This is a one-tap model to explain the overall characteristics of the coefficient multiplier, and the dynamic range of tap coefficient 01 is 1 to 1, and the number of bits is 0 to 1.
l is n bits.

Since the coefficient of coefficient 211 is 0.5, the maximum value (maxy) of the total output is “
0.5". Therefore, the minimum step size (L
SB) becomes 2-(n+1), which means that the total number of bits of the coefficient is increased by 1 bit compared to the conventional one.

Similarly, since ghost g2 is -126'', 180', its amplitude value is 0.25''. Therefore, the coefficient 2
:4212 coefficient is set to 0.25°.The one-tap model of the coefficient multiplier at this time is shown in Fig. 5(a), and the maximum value of the total output (a+; txy) is
As shown in (b) of the figure, it becomes 0.25''. Therefore, the LSB of the output y becomes 2-(n+2), and the total number of bits increases by 2 bits compared to the conventional one.

Furthermore, since the ghost g3 is -18 dll and 90 degrees, its amplitude value is less than "0.25". therefore,
The coefficient of coefficient 4213 is set to "0.125". The total number of bits of the coefficient at this time is increased by 3 bits compared to the conventional method.

In the above description, the coefficient units 211, 212, . . . , 21
Although n is used as an attenuator in the description, it can also be used as an amplifier. For example, as shown in FIG. 6(a), normally a coefficient multiplier 50n is used to correspond to a small boss, and the dynamic range of 01 taps is -1 to 1, n+2 bits, and the output y is -012 bits.
It is designed to be 5 to 0.25. And larger ghosts, e.g. U/D ratio -16dB (-0,5
), the coefficient of the coefficient multiplier 21n is set to 2'. As a result, as shown in Figure 6(b), the dynamic range of the overall coefficient multiplier is 10.5 to 0.5.
Then, the LSB of the output y is 2-(n+1), which is the same as the example shown in FIG.

The coefficients of the coefficient units 211 to 21n are determined by approximate values of the distortion amounts of the corresponding ghosts, and do not need to be particularly precise in practical use. Therefore, this coefficient may be a power of 2, and in this case, the coefficient multipliers 211 to 2in can be constructed not from multiplication circuits but from bittonft circuits and selection circuits.

FIG. 7 shows another embodiment. In this embodiment, n variable delay circuits 181, 182, . . . 1
8o are connected in series and a signal from the changeover switch 12 is supplied to one end thereof. The variable delay circuits 181.1g2, . . . 18n are connected in series.
Transversal filter 1 outputs the output signal from each
91.192,... Make it input to 19++.

Therefore, "2 variable delay circuit 181.11112
, . . . 18n is set equal to the delay amount of the variable delay circuit 181 shown in FIG. The sum of the delay amounts of the variable delay circuits 181 and 182 in this embodiment is made equal to the delay amount of the variable delay circuit 182 shown in FIG. Similarly, the sum of the delay amounts of the variable delay circuits full to 18n in this embodiment is set equal to the delay amount of the variable delay circuit 18n shown in FIG.

The transversal filter 191 of this embodiment
, 192, . . . , 19n are controlled in the same manner as in the embodiment shown in FIG. Coefficient multiplier 211.212,・
..., 2In, and this coefficient multiplier 21
1.212, .

The purpose of this invention is to increase the equivalent number of bits of the tap coefficient according to the ghost distortion m.
Therefore, as a limit to using a coefficient multiplier at the output section of each transversal filter, the present invention also includes using a coefficient multiplier for that purpose at the output section of each coefficient multiplier of the transversal filter.

Further, the present invention does not limit the specific means for determining the amount of delay of the variable delay circuit.

According to the gist of the present invention, at least the output signal of the original tap coefficient unit of the transversal filter is multiplied by a coefficient corresponding to the amount of ghost distortion.
In the embodiment shown in the figure, ii) variable delay circuits 181 to
18n are transversal filters 191-1, respectively.
It may be connected to the output of 9n or the output of coefficient multiplier 2 to 2in.

Further, it is also possible to apply coefficients corresponding to the distortion amount to the sum of the tap coefficient units of adjacent stages of the output weighted transversal filter.

[Effects of the Invention] As described above, the digital ghost removal device according to the present invention can cope with the conditions where ghosts occur by simply adding very sophisticated circuit elements to the conventional ghost removal device. Thus, the number of bits of the tap coefficient can be equivalently increased optimally, and tap gain response and residual ghost reduction can be effectively realized.

[Brief explanation of drawings]

FIG. 1 is a block diagram for explaining a ghost removal device according to an embodiment of the present invention, FIG. 2 is a block diagram for explaining a transversal filter used in the above device, and FIG. 3 is a block diagram for explaining a transversal filter included in a video signal. (a) of each of FIGS. 4 to 6 is a diagram illustrating the relationship between a coefficient unit of a transversal filter and a coefficient unit connected to this transversal filter. , Similarly, (b) is a diagram showing the output state, and FIG. 7 is a diagram explaining another embodiment of the present invention (1). Figure 8 is a configuration diagram explaining a conventional ghost removal device. The figure is a diagram explaining the coefficients and output states of the transversal filter in this conventional device. l2... Changeover switch, 13... Subtractor, 14...
... Arithmetic control section, 17... Cancellation signal generation section, 181-
18n... Variable delay circuit, 191-19n... Transversal filter, 20... Adder, 211-2
1n...Coefficient unit.

Claims (1)

[Claims] A plurality of delay circuits to which a digitized input video signal is supplied and delay times corresponding to the positions of ghosts are set, and output signals from each of the plurality of delay circuits. A plurality of transversal filters each generate a cancellation signal corresponding to a ghost position, a signal from each of the plurality of transversal filters is supplied, and a coefficient corresponding to the size of the ghost corresponding to each filter is supplied. A digitized ghost signal generator, comprising: a plurality of coefficient multipliers configured to set the number of coefficient multipliers; output signals from each of the plurality of coefficient multipliers are combined to remove ghosts from an input signal. removal device.