JPS6010892A - Digital acc circuit - Google Patents

Abstract

PURPOSE:To improve the picture quality by providing an error processing circuit applying processing to an error signal being a difference between an amplitude of a color burst gain-controlled by an ACC signal and an object amplitude and outputting the result as a control signal to be inconsipicuous lateral stripe luminance noise. CONSTITUTION:A chrominance signal 109 inputted to a digital ACC circuit 110 is multiplied with an ACC signal 409 by a multiplier 401, the gain is adjusted and outputted newly as a chrominance signal 126. The chrominance signal 126 is inputted to a burst amplitude detection circuit 402, where a color burst amplitude 403 of the chrominance signal 126 is obtained. The value is subtracted from an object amplitude value 404 at a subtractor 410 and an error signal 405 is outputted. The error signal 405 is inputted to an error processing circuit 406, where the nonlinear processing is applied and the result is outputted as a control signal 407.

Description

Detailed Description of the Invention [Technical Field of the Invention] The present invention relates to a digital /l/ACC (Automatic Co1or C) system applied to a digital television receiver that performs baseband video signal processing digitally.
ontrol ) circuit.

[Technical background and problems]

Conventionally, all signal processing in television receivers has been performed by analog signal processing, but there have been problems that need to be improved, particularly in signal processing after the video stage, as described below.

In other words, in terms of performance, it is a problem caused by processing performance on the time axis, which is considered to be a general weakness of analog signal processing.Specifically, it is a problem caused by the processing performance on the time axis, which is considered to be a general weakness of analog signal processing. These include separation performance, various image quality improvement performance, and synchronization performance.

On the other hand, as a cost and manufacturing problem, even if the circuit is made into an IC, there are many external parts and parts to be adjusted.

In order to solve these problems, it is being considered to completely digitalize the signal processing from the video stage onwards, including primary color signals and color difference signal recovery.

By the way, in television receivers, the gain of the color signal is controlled so that the color burst in the color signal always has a constant amplitude despite changes in the amplitude of the color signal due to the characteristics of the transmission path from the transmitting station to the receiver. An ACC circuit is required.

The ACC control procedure is normally the same.

This is the difference between the following. First, the input color signal is multiplied by a signal (referred to as an ACC signal) for controlling the gain of this signal. Next, the amplitude of the color burst in the gain-controlled color signal is detected, and an error signal that is the difference between this magnitude and a separately determined target value is calculated. Finally, the error signal is passed through a loop filter having an appropriate low-pass characteristic, and its output is used as the ACC signal, which is again multiplied with the color signal.

Through the above control loop, the gain of the color signal is controlled so that the amplitude of the color burst converges to the target amplitude. This control procedure is the same for digital television receivers. However, in digital television receivers, analog video signals can be digitized with a finite word length,
Control problems arise due to the truncation of the lower bits in subsequent operations. FIG. 1(a) is a diagram for explaining this. FIG. 1(a) shows the state of convergence when an error occurs at time 2 in the amplitude of the color burst, which is controlled by the method described above. The error at time ■ decreases to within 10% at time ■, and the time during this time is defined as convergence time ts. In analog control, this change is of course continuous, and after time ■ it converges toward the target amplitude without limit, but in digital control, it is accompanied by a finite word length error, so after time ◎ it is in convergence mode. to the steady vibration mode. The vibration width △ at this time depends on the operation word length, and the vibration period tosc is the loop gain of the control loop,
It depends on the convergence time ta. This color burst
Furthermore, the steady amplitude of the color signal appears as a change in color saturation on the screen, degrading the image quality. Furthermore, since the ACC control is performed in units of one horizontal period, this color change becomes horizontal striped color noise on the screen. Normally, the vibration width Δ is set so that this color noise is not noticeable, and the operation word length is determined accordingly. Naturally, reducing the vibration amplitude Δ requires a much longer operation word length, which increases the circuit scale and requires a faster element speed. Furthermore, the vibration width of the color noise appearing on the screen is considerably expanded by Δ for the reason described below, and therefore the calculation word length and element speed required to cancel this are further increased.

That is, the color signal subjected to ACC control is demodulated into I and Q signals, and further subjected to a matrix operation together with the luminance signal (Y) to be converted into an RGB signal. Now, if we focus on colors with the same amplitude as the color burst, the same amount of noise will be generated in the spot color where Δ noise was generated in the color burst due to ACC control, and this is a problem.

Even when demodulated into Q signals, each has approximately Δ noise. In the subsequent matrix calculation, for example, B (blue) is given by B == Y -1, II + 1.7Q, so especially for blue and yellow, 1. ! Since =Q has different signs, the noise of the B component of these colors is expanded to 2.8·Δ. Conversely, for R (red), R = 0.95 I
+0.62 Q, so the noise of R components such as cyan, magenta, and red where I and Q have the same sign is 1.57 @
It is enlarged to Δ. Furthermore, since a highly saturated color has an amplitude 2 to 3 times that of the color burst, the noise will also be magnified to 2 to 3 times the above-mentioned value, and the impact on the screen will be significant. Of course, almost the same expansion occurs with normal noise, but in this case, the noise is in the form of horizontal stripes, so it is visually noticeable and significantly degrades the image quality.

In order to suppress the amplitude of such noise, in addition to increasing the arithmetic word length in the ACC control loop and the associated element speed as described above, the quantization word length itself in the A/D converter must also be increased to It is necessary to reduce the word length error. Therefore, the burden placed on the hardware is enormous. By the way, what determines whether this color noise is visually noticeable or not is the amplitude △ in Fig. 1(a) mentioned above.
In addition, there is a vibration period tosc, which corresponds to the repetition period of horizontal striped color noise on the screen. Furthermore, tosc depends on the loop gain in the ACC circuit. Therefore, in order to visually suppress color noise, the quantization word length and calculation word length for the amplitude Δ and the vibration period tosc
, it is necessary to consider both the loop gain for .

However, another factor that determines the loop gain is the convergence time ts, which is generally one of the most important parameters in a feedback control system and determines the stability of the system and the noise performance of the control. Therefore, the loop gain is usually determined from the requirement of convergence time t$, and the oscillation period tosc,
No measures will be taken for this. On the other hand, as mentioned above, the vibration width Δ increases due to the processing after ACC control, so
In order to suppress this to a detection limit, a significant additional circuit and element speed are required for the ACC circuit and A/D converter.

[Purpose of the invention]

The present invention has been made in view of the above-mentioned points, and its purpose is to change the repetition period of horizontal striped color noise in a steady state without changing the convergence time in an ACC circuit of a digital television, and to visually display this noise. This makes it less noticeable and improves image quality.

[Summary of the invention]

The present invention provides a digital ACC circuit in which an error signal, which is the difference between the actual color burst amplitude and the target amplitude, is
A when it is within the specified range and when it is not based on zero
The main point is to change the loop gain of the CC loose.

[Embodiments of the invention]

FIG. 3 shows one of the digital television receivers according to the present invention.
It is a figure showing an outline of an example. The television signal received by the antenna 101 is sent to the tuner 102 and the IP stage 103.
．． The baseband video signal 123 passes through the detection stage 104.
is converted to The video signal 123 is fed to a synchronization circuit 106 where synchronous playback is performed and horizontal and vertical drive pulses 117, 118 are output and fed to a deflection circuit (not shown). Further, the video signal 123 is an A/D
The signal is input to a conversion circuit 105, where sampling and quantization are performed, and a digital video signal 124 is output.

The sampling frequency f$ is fm = 41 mc, and the sampling phase is synchronized with the color burst earthwork axis, ±Q axis.

A sampling pulse 115 that provides sampling timing is generated by a clock generation circuit 127, A,] conversion circuit 105. It is supplied to the digital signal processing circuit 122 and the synchronization circuit 106 and used as a reference for circuit operation. Further, among the sampling pulses from the clock generation circuit 127, +I axis and +Q axis ones are sent to the color detection circuit 11.
1 as a detection pulse 116 that provides a color detection axis. The digital video signal 124 is transmitted to the YC separation circuit 10
The luminance signal 108 is input to 7 and passed through a digital filter.
and a color signal 109. The chromaticity signal 109 is inputted to the digital ACC circuit 1105, and after the gain is adjusted, it is supplied to the color detection circuit 111. Incidentally, a burst gate pulse 125 indicating the position of the color burst is input to the digital ACC circuit 110 from the synchronization circuit 106. This provides timing for color burst integration in the digital ACC circuit 110 to detect the size of the color burst. Note that the burst gate pulse 125 is equal to 4 cycles of color burst, t, D
It has a width of 16 samples. The color detection circuit 111 selectively extracts the optical and Q axis phase signals from the ACC-controlled color signal 126 using the detection pulse 116, thereby generating the I and Q signals 113, respectively. 11
I got 4. Luminance signal 108. I and Q signals 113.
114 is input to a matrix circuit 112 and converted into a fi RGB signal 119 by a predetermined matrix operation. This is a digital analog RGB signal from the D/A converter 120.
It is converted into a signal 121 and operates the CRT via an output circuit. The above is an overview of digital television receivers. Next, one embodiment of the digital ACC circuit 110 of the present invention will be described in detail.

FIG. 4 is a diagram showing a block configuration of the digital ACC circuit 110. The color signal 109 input to the digital ACC circuit 110 is multiplied by the CC signal 409 in the multiplier 401, the gain is adjusted, and a new color signal 126 is output. The color signal 126 is input to the color detection circuit 111 in FIG.
02, where the color burst amplitude value 403 of the color signal 126 is obtained. This is subtracted from the target amplitude value 404 in the subtracter 410, and the error signal 405
is output. The error signal 405 is an error processing circuit 406
Here, the signal is subjected to nonlinear processing as shown in FIG. 7, and is output as a control signal 407. The high-frequency components are attenuated by the loop filter 408, and the A
It is converted into a CC signal 409 and input to the multiplier 401. According to this control procedure, the color signal 126 is gain controlled so that the color burst has a target amplitude. More than,
2 is a schematic diagram of the entire digital ACC circuit 110. Next, the characteristic circuits among these will be explained in detail.

FIG. 5 is a diagram showing a specific configuration of the color burst amplitude detection circuit 402. This operation will be explained below. Color signal 12 input to color burst amplitude detection circuit 402
6 is input to the AND gate 501 together with the burst gate pulse 125 supplied from the synchronization circuit 106. Since the burst gate pulse is 1'' during the 16 samples in the color burst and is 1'' at other times, the 16 sample values of the color burst are extracted as the output signal 513 of the AND gate 501. This color burst 513 is
It is divided into a sign bit 514 and a numerical value bit 512, both of which are input to an exclusive OR (Ex-OR) gate 502. By performing an Ex-OR between the sign bit and numerical bit of the data expressed in two's complement, the absolute value of the data can be determined. Therefore, the output signal 515 of Ex-OR gate 502 is the color burst 513
is the absolute value of However, the absolute value of the negative number obtained using this method is smaller by the size of its LSB (δ), so
Needs correction. This is corrected during subsequent calculations. In other words, the absolute value signal 515 of the 16 color burst samples is sent to the adder 505 and the data latch 507.
At this time, the color burst 513 is input to the carry input terminal of the adder 505.
514 is input, and δ is corrected for negative numbers.

On the other hand, the data latch 507 has the burst gate pulse 1
The latch operation is performed by the sample pulse 115 only during the period when 25 is ``1'', and the output signal 508 is used during the period when 25 is ``1''.

Rear. Through this operation, the integrated waveform of the absolute value signal 515 is obtained as the signal 508. Continued data latch 5
11, a latch operation is performed at the falling edge of the burst gate pulse 125. This adds adder 505 and data latch 5.
11 can latch the value at the time it completes integrating the absolute value signal 515.

This value is multiplied by 1/16 by a coefficient multiplier 509,
It is converted into an average color burst amplitude 403 and output.

This operation provides highly accurate color burst amplitudes even in the presence of noise. The above is a detailed description of the color burst vibration detection circuit 402.

Next, the detailed configuration of the error processing circuit 406 will be described using FIG. Error signal 40 input to error processing circuit 406
5 is the sign bit 601 and the numerical bit 611Ex-OR
To calculate the absolute value of the error signal 405 expressed in two's complement format as described above is to calculate the absolute value of the error signal 405 expressed in two's complement form. In this case, the error of the LSB (δ) with respect to the absolute value of the negative number poses no practical problem. The error absolute value 603 obtained in this way is the value d
605 is input to the comparator 604. Here, the magnitudes of the two are compared, and when the value d605 is larger, the comparison signal 606 is ``1'', and when it is smaller, it is O''. This is input to the selection terminal of the data selector 610.Data selector 610 contains the absolute error value 603 and a signal 612 obtained by multiplying this by 9174 by the coefficient multiplier 607.
is input.

The absolute error value 603 itself appears in the output signal 608 of the data selector 610 when the comparison signal 606 is "i" and when the signal 612 is "0". When the absolute error value 603 is smaller than the value d605, a signal 612 which is 1/4 times the absolute error value signal 603 becomes the signal 608, and when it is larger, the absolute error value signal 603 itself becomes the signal 608. Therefore, the signal 608 is the control signal 4 shown in FIG.
Corresponding to the absolute value of 07, from now on the control signal 40
To make 7, Ex-O of signal 608 and sign bit 601
Just take R. This is done with Ex-OR gate 609. With the above circuit, the relationship between the error signal (e) 405 and the control signal 407 (C) is as shown in FIG. becomes 1. wife! lld:)e)-d
+δ, otherwise the relative loop gain is 1/4
become.

The reason why the error signal 405 is converted into the error absolute value 603 and processed in order to obtain the characteristics of the error processing circuit 406 (FIG. 7) is because the amount of circuitry is small.

For example, even if the error signal 405 is 10 bits, the error processing circuit 406 can be constructed with about 150 gates.

Next, the shield loop filter 408 will be explained with reference to FIG. The input control signal 407 is multiplied by 1/32 by a coefficient multiplier 801, and then sent to an adder 802 and a data rack f804.
The result is the ACC signal 40
Output as 9.

[Other embodiments of the invention]

Although one embodiment of the present invention has been described in detail above, the basic circuit configuration is not limited to the above embodiment, and various modifications are possible. For example, the input/output characteristics of the error processing circuit 406, which is a characteristic circuit of the present invention, are discontinuous at e=d and -ti+δ as shown in FIG. The advantage of this is that it can be realized with a simple circuit configuration as described above, but when the error signal 405 that frequently crosses these discontinuous points is input, the control characteristics become discontinuous. , which may have an adverse effect on the screen. FIG. 9 shows an embodiment in which this problem is solved by making the input and output car characteristics continuous.

The difference between this and FIG. 6 is that an adder 613 is added. Tsumashi, to make the characteristics in Figure 7 continuous, e
) At d, add 1τd to the output, then -d
At +δ, there is a discontinuity of 7δ, but this poses no practical problem. ) In order to realize this circuit-wise, it is sufficient to add -7d to the error signal 603 converted into an absolute value, as shown in FIG. As a result, the input/output characteristics shown in FIG. 10 are obtained, and the operational problem is improved.

〔Effect of the invention〕

As mentioned above, digital television has many advantages over analog television, but IC
Regarding the e circuit, this generates horizontal striped color noise due to finite word length error, which distorts the image.

It was deteriorating the quality. Moreover, there is a limit to changing the return period of this noise to a place where the visibility is low due to a balance with the convergence time.

However, according to the present invention, by simply adding a small amount of circuitry, it is possible to change the repetition period of the horizontal striped color noise without changing the convergence time, thereby making it visually less noticeable. . Specifically, an error signal (e) calculated by the difference between the amplitude of the color burst whose gain is controlled by the ACC signal and the target amplitude is input, and the error signal (e) is processed and output as a control signal (C). AC processing circuit
This is achieved by adding it into C loose. FIG. 2 is an example schematically showing the input/output characteristics of this error processing circuit.

FIG. 2(a) is equivalent to the conventional example since c==e and no interpolation processing is performed, and the state of convergence is shown in FIG. 1(a). In FIG. 1(b), processing is performed to reduce the differential coefficient of input and output where e is small, and the state of this convergence is shown in FIG. 1(b). Figure 2 (C), on the other hand, performs processing to increase the input and output differential coefficients where e is small.
The state of this convergence is shown in FIG. 1(C). Note that 0) is substituted for the change in Figure 2 (b), and (C) is substituted for the change in Figure 1 (C).
), similar effects can be obtained.

The circuit configuration in this case is slightly simpler than those in (b) and (C). As shown in FIG. 1, in the present invention, the error signal e
By changing the loop gain where the value of is small, it is possible to significantly change the vibration period tosc in the steady vibration state while keeping the convergence time ta approximately the same. It is possible to convert the return period to a place with low visibility. Further, the error processing circuit added for this purpose can be constructed of, for example, a simple comparator, a data selector, etc., and requires only a small amount of circuitry. In addition, the vibration is 100
There is no problem in practical use as long as it is selected in the upper range. Generally tons
e is easily noticeable when it is 10 to several tens of times the horizontal period, and many methods of increasing tosc are used to prevent this. This is because a method is used to reduce the loop gain near zero of the error signal in consideration of noise performance.

As described above, according to the present invention, horizontal striped color noise, which has been a problem in conventional digital ACC circuits, can be made less noticeable by using only a small squared circuit, and image quality can be significantly improved.

[Brief explanation of the drawing]

Fig. 1 is a waveform diagram showing how the amplitude of the color burst converges to the target amplitude, Fig. 2 is an input/output characteristic diagram of the error processing circuit, Fig. 3 is an overall configuration diagram of the digital television receiver, and Fig. 4 The figure is a block diagram of a digital ACC circuit which is an embodiment of the present invention, FIG. 5 is a block diagram of a color burst amplitude detection circuit which constitutes a part of the present invention, and FIG. 6 is a block diagram of a color burst amplitude detection circuit which constitutes a part of the present invention. FIG. 7 is an input/output characteristic diagram of the error processing circuit shown in FIG. 6, FIG. 8 is a configuration diagram of a loop filter that constitutes a part of the present invention, and FIG. A block diagram of an error processing circuit used in the digital ACC circuit of another embodiment, FIG. 10 is an input/output characteristic diagram of the error processing circuit of FIG. 9. 109... Color signal, 110... Digital ACC circuit, 126... Color signal, 401... Multiplier, 402...
. . . Color burst amplitude detection circuit, 403 . . . Color burst amplitude value. 404...Target amplitude value, 41o...Subtractor, 4o5
...Error signal, 406...Error processing circuit, 4o7.
...Control signal, 408...Loop filter,
4o9...Ace i issue. Agent Patent Attorney Kensuke Chika (and 1 other person)] Figure 1 Figure 2 (0) (b”) (C) Figure 4 Figure 8 Figure 5 Figure 7

Claims (1)

[Claims] After the analog video signal is digitized, a color signal separated from the digital video signal and an ACC signal, which is an amplitude control signal, are input to a digital ACC circuit applied to a digital television receiver that performs various signal processing. a multiplier; an error calculation circuit that subtracts the amplitude value of the color burst in the output signal of the multiplier from a set target value and outputs it as an error signal; and the error signal is a value within a certain range that can be arbitrarily set. an error processing circuit which multiplies this error signal by a constant and outputs the resultant signal, and if the value of the error signal falls outside the set range, adds a constant number to the error signal and outputs the result; A digital ACC circuit, comprising: a loop filter that receives an output signal from the above, attenuates high frequency components thereof, and outputs the ACC signal as the ACC signal.