JPS6240907B2 - - Google Patents

Description

[Detailed description of the invention]

The present invention relates to an improved ghost removal circuit that can automatically suppress reception interference caused by ghost waves in a television receiver when receiving television broadcasts. Ghost waves are generated during the propagation of broadcast waves when reflected waves or diffracted waves from topographical objects or buildings arrive at a reception point later than the desired broadcast waves (direct waves). Due to this delay, the modulation signal of the ghost wave has a delay with respect to the modulation signal of the direct wave, and the carrier wave of the ghost wave has a phase difference with respect to the carrier wave of the direct wave. on the other hand,
Since television signals are vestigial sideband (VSB) signals, waveform distortion occurs when a signal with a different phase from the standard direct wave is detected, and ghosts cannot be removed by simply delaying the detection output of the direct wave. I can't. In addition, regarding ghost removal in a TV receiver, the amplitude ratio, phase difference, and delay time of the ghost to the direct wave are parameters for the ghost removal conditions, but in actual ghosts, the above phase changes from time to time, so the ghost removal conditions must be set. In conventional ghost removal circuits in which ghost removal is manually performed, the viewer must constantly set the ghost removal conditions, which is impractical. From this perspective, a ghost removal method that uses a squared sine wave pulse as a reference waveform has been proposed as an automatic removal method that can automatically set ghost removal conditions. It has the disadvantage that it is rarely used in television broadcasting. A method has also been proposed in which the vertical synchronization pulse portion within the vertical retrace period is used as a reference signal, but with conventional methods, convergence to the ghost removal conditions is slow and even after convergence, large errors occur. It has the disadvantage of having. This invention has been made for the purpose of eliminating the above-mentioned drawbacks, and uses the vertical synchronizing pulse portion within the vertical retrace period actually included in the television signal or the horizontal synchronizing signal as a reference signal for ghost removal. The present invention provides an improved ghost removal circuit that can remove ghosts at high speed and with high precision. Hereinafter, a ghost removal circuit according to the present invention will be explained based on embodiments. The figure is a block diagram showing the configuration of an embodiment of the present invention. In the figure, 1 is a video detection circuit, and 2 is a so-called transversal filter type compensation signal generation circuit composed of a plurality of unit delay circuits and a plurality of coefficient circuits. It generates a compensation signal that cancels the ghost component contained in the output of the detection circuit 1, and the waveform of this compensation signal is configured so that it can be controlled over a wide range by the coefficient values of a plurality of coefficient circuits. ing. 3 is the first adder circuit, 4 is the output terminal,
The output signal from this output terminal 4 becomes a video signal from which ghosts have been removed. 5 is a scanning line extraction circuit;
Reference numeral 6 denotes a sample pulse generation circuit, and the scanning line sampling circuit 5 utilizes the vertical and horizontal synchronization signals included in the output signal to extract the sample pulses only during the scanning line period including the vertical synchronization signal portion within the vertical retrace period.
is operated to generate sample pulses for a predetermined scanning line period. Note that the synchronization of the sample pulses is selected to be equal to the unit delay time within the compensation signal generation circuit 2. 7 is a sampling circuit, 8 is a control circuit, 9 and 10 are first and second delay circuits that synchronously delay the signal by one sample, and 11, 12, and 13 are each applied signal to each circuit's inherent gain. 14 is a second addition circuit, and this second addition circuit 14 is amplification circuit 1.
Add signals 1, 12, and 13. The output of the second addition circuit 14 is sampled by a sampling circuit 7 based on the output of the sample pulse generation circuit 6,
Only the error signal is extracted. The control circuit 8 is configured to control the coefficient group of the compensation signal generation circuit 2 using the sample series of the error signal. The operation of this embodiment will be explained in detail below. Letting the output signal of video detection circuit 1 be x _k ,
This can be written as x _k =a _k +g _k (1) where a _k is the direct wave signal and g _k is the ghost. Here, x _k is a certain time t=
_k is the value of x at T, and similarly in other cases. In addition, the output _k of the compensation signal generation circuit 2 is obtained by inputting the coefficient value from the input side, with the number of coefficient circuits being M.
If C ₁ , C ₂ , ..., C _M is given by The first addition circuit 3 is configured to add the output of the compensation signal generation circuit 2 and the output of the video detection circuit 1. Therefore, the output signal y _k from the first adder circuit 3 is given by equations (1) and (2). becomes. The operation as a ghost removal circuit is expressed in equation (3).

[Formula] Coefficient C _n of compensation signal generation circuit 2 (m=1, 2, ..., M)
is to decide. Now, in television broadcasting, a so-called vertical synchronization section is inserted during the vertical retrace period, and this is
It has a flat part with constant amplitude over 1/2H. If a ghost exists, this portion will no longer be flat, so by using this information, the ghost can be detected without being affected by the direct wave signal and ghost of the actual image signal. That is, by using this vertical synchronizing pulse portion as a reference signal and determining the coefficient value of the compensation signal generating circuit 2 so that the ghost in this portion becomes zero, the ghost mixed in the television signal can be removed. In the figure, the scanning line sampling circuit 5 uses the vertical and horizontal synchronizing signals included in the output signal from the first adder circuit 3 to generate sample pulses only during the 1/2H period after the equalization pulse within the vertical retrace period. circuit 6
to generate sample pulses with predetermined synchronization. The sampling circuit 7 is configured to sample the output signal from the second addition circuit 14 using the sample pulse from the sample pulse generation circuit 6. Hereinafter, explanation will be given focusing only on the sampling time point by this sample pulse, that is, the 1/2H period after the equalization pulse. Since this period is a no-signal period, the output of the first adder circuit 3 during this period, that is, the output terminal 4
The output signal is more of a residual ghost It's getting old. Further, the second adder circuit 14 applies the output signal at the output terminal 4 to the amplifier circuit 11 to obtain an output, and a signal obtained by delaying the signal at the output terminal 4 by one sample time by the delay circuit 9 to the amplifier circuit. 12 and a signal obtained by applying a signal obtained by delaying the output of the delay circuit 9 by one sample time by the delay circuit 10 to the amplifier circuit 13.
Therefore, if the gains of the amplifier circuits 11, 12, and 13 are P, Q, and R, respectively, then the output δ _k of the second adder circuit 14 is δ _k =Pε _k +Qε _k-1 +Rε _k-2 . . . (5). Further, the control circuit 8 utilizes a series of error signals δ k obtained by sampling the output signal of the second addition circuit 14 with the sampling circuit 7, and uses the series of error signals δ _k obtained by sampling the output signal of the second adder circuit 14 to generate a signal proportional to the error signal δ _n at the m-th sampling point. The corresponding coefficient C _n of the compensation signal generation circuit 2 is modified by the value given. That is, when the coefficient value in the first field is C ^(l) _n , the control circuit 8 calculates that the corrected coefficient value in that field is C ^(l+1) _n = C ^(l) _n + δ ^(l) _n = C ^{(l )} _n + (Pε ^(l) _n +Qε ^(l) _n
-1 +Rε ^(l) _n-2 )
...(6) It is configured to perform control such that.
With this configuration, the residual ghost ε _k can be made zero, as shown below, and the ghost can be removed. Now, in equation (4), the value of C _n that makes ε _k = 0 is C
If _n is ₀ and each coefficient value is C ^(l) _n in the vertical synchronization pulse part of the first field, then the residual ghost ε ^(l) _k in this field is is given by Here, △C ^(l) _n =C _n , ₀ - C ^(l) _n ...(8). This ΔC ^(l) _n can be obtained from the residual ghost series ε ^(l) _k using equation (7). Since the band of a television signal is limited, the rising portion of the vertical synchronization pulse portion has a short but certain time. Consider the following as an example of a practical value for this rising portion. The sample value series of the vertical synchronization pulse section, that is, the sample value series with the sample frequency determined from the delay time of the unit delay circuit included in the compensation signal generation circuit 2 is (0.4, 0.8, 1.0, 1.0, ..., 1.0)... ...(i), and by considering the case where the ghost is small and setting x _k a _k , equation (7) can be expressed as follows. Solving equation (9) for △C _n gives us The values of the elements marked with * in the matrix on the right side of equation (10), which yields , are small values that are less than 1/2 of the values of the diagonal elements. Rewriting equation (10), △C ^(l) _n =2.5ε _n −5.0ε _n-1 +3.6ε _n-2 +……2.5ε _n −5.0ε _n-1 +3.6ε _n-2 (m= 1, 2, ..., M) ...(11). As mentioned above, the corrected value of the coefficient value C _n controlled by the control circuit 8 is given by equation (6), where P=2.5, Q=-5.0, R=3.6 In other words, if the gains of amplifier circuits 11, 12, and 13 are 2.5, -5.0, and 3.6, respectively, the second term of equation (6) becomes (11)
As shown in the formula, ΔC ^(l) is approximately equal to _n . As mentioned above, the part marked with * in the matrix of equation (10) is not zero, but takes a small value, so it is not possible to set C ^{(l + 1)} _n = C _n0 with one coefficient correction operation. , all coefficients are reduced to C _n0 (m=1,
2,...,M). Since the coefficient C _n0 is nothing but a ghost removal condition, it is clear that ghosts are removed by the above operation. Note that since the remaining scanning line portions constituting the field determined in this vertical synchronization pulse portion remain unchanged, ghosts are removed from all image signals constituting the field. As is clear from the above explanation, according to the method of the present invention, the coefficient value of the compensation signal generation circuit 2 necessary to remove the ghost from the residual ghost ε ^(l) _k in the vertical synchronization pulse portion of the first field Since values of C _n , ₀ with good approximation accuracy can be directly obtained for all m, this invention has a great practical advantage of being able to perform ghost removal at high speed. Furthermore, each coefficient C _n is controlled by its corresponding error signal δ _n , but as expected from equation (10), this δ _n is influenced by other coefficients C _j (j≠m). This method also has the advantage of being able to perform removal operations with high precision because there are fewer problems. Incidentally, methods have been known that use the vertical synchronizing pulse portion as a reference signal, but the conventional method uses the value of the output signal ε ^(l) _n in equation (9), and this ε ^{(l )} Corresponding coefficient C _n is modified by an amount proportional to _n . In this conventional method, as expected from equation (9), the amount of correction of the m-th coefficient C _n is the residual difference ΔC ^(l) _n of each coefficient.
In addition, the influence of the residual ΔC _j (j<m) of other coefficients is also included in the correction. Therefore, it is necessary to repeat the correction operation many times in order to converge △C ^(l) _n to zero for all coefficients, and even after ε _n = 0 for each m, all △C _n becomes zero. There is no guarantee that it will stay the same. This is a drawback of conventional methods, and the method of the present invention eliminates these drawbacks as described above. Also, to simplify the explanation, we have used x _k
ak , and the sample value sequence of the vertical synchronization pulse portion is as shown in (i) above. However, in general, there may be cases where a ghost overlaps with the reference signal, so x _k ≠ a _k , and the above sample The value series (i) also generally changes. In general, the coefficient control method shown in equation (6) C ^(l+1) _n =C ^(l) _n +δ ^(l) _n =C ^(l) _n + (Pε ^(l) _n +Qε ^(l) _n
By selecting appropriate values for P, Q, and R in _-1 +Rε ^(l) _n-2 ), this correction operation is repeated and ΔC _n can finally be converged to zero. For example, let P, Q, R be a positive number smaller than 1.
You can choose 2.5k, -5.0k, 3.6k, 1, -
It has been experimentally confirmed that convergence is achieved even if 2, 1, or 2, -1, -1 are selected. Furthermore, it has been confirmed that convergence can be achieved even if R=O and P and Q are selected such that P>O and Q>O. In this case, the second delay circuit 10 and third amplifier circuit 13 can be omitted in the figure. Furthermore, as another control method, using the sign of the error signal δ _n and using A as a proportionality constant, C ^{(l + 1)} _n = C ^(l) _n + Asgn (δ ^(l) _n ), the above case does not occur. Almost the same △
C _n can be converged to zero, and ghosts can be removed. Next, a case will be described in which a horizontal synchronizing signal within a vertical retrace period is used as a reference signal for ghost removal. In television broadcasting, there are multiple scanning lines that contain only burst signals in addition to the horizontal synchronization signal within the vertical retrace period, and these have flat portions that span approximately 1H. If a ghost exists, this portion will no longer be flat, so by using this information, the ghost can be detected without being affected by the direct wave signal and ghost of the actual image signal. That is, by using this horizontal synchronizing signal as a reference signal and determining the coefficient value of the compensation signal generating circuit 2 so that the ghost in this portion becomes zero, the ghost mixed in the television signal can be removed. In this case, the scanning line extraction circuit 5 shown in the figure
Using the vertical and horizontal synchronization signals included in the output signal, the next sample pulse generation circuit is operated only during a predetermined scanning line (multiple) period that includes only the horizontal synchronization signal and burst signal within the vertical blanking period. Further, the sample pulse generation circuit 6 is controlled by the scanning line extracting circuit 5, and is configured to generate a sample pulse after the trailing edge of the horizontal synchronizing signal in the predetermined scanning line period. For the rising and falling parts of the horizontal synchronization signal, the same sample value series as for the rising part of the vertical synchronization pulse part can be obtained, so P,
Ghosts can be removed by determining Q and R. Note that when this horizontal synchronization signal is used as a reference signal for ghost removal, the coefficient correction operation can be performed at least six times in one field, so that the convergence time can be further shortened. Furthermore, if a horizontal synchronizing signal is used, the subsequent no-signal time is long, so even ghosts having a delay time of approximately the horizontal scanning period can be removed. In the above embodiment, the output of the video detection circuit 1 and this signal are transmitted to the compensation signal generation circuit 2.
A case has been described in which a feed-forward configuration is used in which the first adder circuit 3 adds the output obtained by applying the signal to the input signal, but the discussion up to now also holds true even if this part is configured as a feedback configuration. Furthermore, in the previous explanation, the first delay circuit 9 and the second delay circuit 10 were used as means for generating the error signal δ _k , but these delays can be reduced by configuring the control circuit 8 as follows. The circuit can be omitted and the output signal of the output terminal 4 can be directly applied to the sampling circuit 7. That is, the control circuit 8 is sampled with a sample pulse from the sampling pulse generation circuit 6, and from the obtained sample value series, the mth sample value is multiplied by P, the (m-1)th sample value is multiplied by Q, and ( Even if the configuration is configured such that the m-2)th sample value is multiplied by R, these values are added, and the coefficient conversion signal is sent to the coefficient circuit corresponding to the mth sample point based on this value, the above discussion will not occur. be satisfied. As described in detail above, according to the method of the present invention, it is possible to quickly converge to the ghost removal conditions and perform the removal operation with high accuracy.

[Brief explanation of the drawing]

The drawing is a block diagram showing an embodiment of the invention. In the figure, 1 is a video detection circuit, 2 is a compensation signal generation circuit, 3 is a first addition circuit, 4 is an output terminal, 5 is a scanning line extraction circuit, 6 is a sample pulse generation circuit, 7 is a sampling circuit, and 8 is a control circuit, 9,
10 are the first and second delay circuits, 11 and 1, respectively.
2 and 13 are first, second, and third amplifier circuits, respectively, and 14 is a second addition circuit.

Claims (1)

[Scope of Claims] 1. Ghost removal by adding the output of a video detection circuit and the output of a compensation signal generation circuit including a plurality of unit delay circuits and a plurality of coefficient circuits with variable coefficient values. , sequentially samples a predetermined period within the vertical retrace period of the added output,
j of the sample value series obtained by this sampling
The above sample values are added by giving appropriate weights to each of the th sample value, the (j-1)th sample value, and the (j-2)th sample value, and the value obtained by this addition is A ghost removal circuit characterized in that the coefficient value of the coefficient circuit temporally corresponding to the j-th sample value is corrected by a value proportional to or a sign of the obtained value. 2. The ghost removal circuit according to claim 1, wherein the weight of the 2(j-2)th sample value is 0. 3. The ghost removal circuit according to claim 1, wherein the predetermined period within the vertical blanking period is a vertical synchronization pulse portion. 4. The ghost removal circuit according to claim 1, wherein the predetermined period within the vertical blanking period is a period of a plurality of scanning lines including only a horizontal synchronizing signal and a burst signal.