JPH10208036A - Signal processor and signal processing method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To enhance color sharpness of a color signal without producing ringing. SOLUTION: A median logic circuit 30 samples three signals A to C, selects a value from the three signals and generates an output signal that has a steep color edge. A signal A is an input signal that is not processed, a 2nd signal C is an input signal that is delayed twice, and a 3rd signal B is differentiation of a signal that is delayed one time. This method can ideally be used by an SVP (scan line video processor), needs only the smallest number of instructions but can be realized in a few gates on an ASIC or an FPGA.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates generally to a signal processing circuit including a color processing circuit, and more particularly, to a color processing circuit and a method for increasing a gradient of a transient portion of a color signal.

[0002]

2. Description of the Related Art All image signals, including NTSC, PAL and SECAM type signals, have luminance and chrominance signals containing image information. The bandwidth of the chrominance signal is narrower than the bandwidth of the luminance signal. Due to the limited frequency bandwidth of the chrominance signal, the transient portion of the chrominance signal is relatively slow. In other words, the gradient of the transitional portion indicating the color edge is only a gentle gradient, and when this transitional portion is displayed, the transitional portion falls off from a sharp color boundary. For example, since the transition of a typical color signal from green to magenta is not very steep, a smear phenomenon in which the color trails after the color edge is observed.

[0003] Television systems that perform digital signal processing differ from conventional television systems that perform analog signal processing in that the color signal includes color information for discrete pixels. Digital signal processing is generally performed on these digital chrominance signals to increase the sharpness of the color transitions, especially between adjacent pixels. One such technique is:
"Reduction of color noise and improvement of transients" is disclosed in U.S. Pat. No. 4,935,806 to Rabbie. This technique provides a color signal in which undershoot and overshoot characteristics are controlled and transient characteristics are improved. However, this technique is somewhat complicated.

FIG. 1 shows another known technique.
In this technique, two bandwidth optimization paths are provided so as to detect the position of a color transition portion in an input color signal.
The exact location of the color transition is calculated by detecting the corresponding zero crossing of the second derivative of the two color signals. Low-pass filtering (LPU, LPV, LPUV) is performed to prevent the influence of noise.
The width of the transient portion is obtained from the threshold detection signal.
The width of this transition indicates the area around the detection point where the first derivative of the color signal exceeds a programmable threshold (THRESH). This parameter THRESH is DCTI
It changes the sensitivity of the circuit. A large value of THRESH only improves large color transitions, while small color transitions remain unchanged. The width of the detected transient can be limited by a programmable parameter TRAWID, which performs an adaptation to the bandwidth of the input chrominance signal. By using a large TRAWID value for a signal having a narrow bandwidth of a color signal (for example, a signal of a video recorder),
Optimizing performance. An input signal having a wide color signal bandwidth is processed with a small TRAWID value. A more detailed description of this method can be found in the book "Integrated 9-Bit Triple DA for Enhanced TVs" written by Siemens.
Digital Display Processor with C "in a paper published in the IEEE Transaction on Consumer Electronics. This method is also quite complicated and expensive.

[0005]

It is desirable to provide a digital color transient improvement (DCTI) algorithm that can process color signals with steep color edges without ringing. The improved algorithm must have a simple architecture for performing logical operations and must be carefully considered so that it can be implemented by digital signal processing. This algorithm is a scanline video processor (SVP) with few instructions
, Or a small number of gates on an ASIC or FPGA.

[0006]

SUMMARY OF THE INVENTION The present invention provides the technical advantage of a digital color transient improvement (DCTI) circuit that produces sharper color edges without ringing. The present invention has a relatively simple architecture for performing logical operations. This circuit converts the input signal to the first
And a second delay circuit that delays the input signal by a second time longer than the first time. A differentiating circuit generates a second derivative of the input signal delayed by a first time. This derivative is an approximation of a mathematical derivative. The median logic circuit is the input signal, the first
By selecting either the differential signal of the input signal delayed by time or the input signal delayed by the second time, an output signal having a sharpened color edge without ringing is generated. Preferably, at the color rising edge, the median logic circuit selects the value of the input signal delayed by the second time until the value of the differential signal is greater than the value of the input signal delayed by the second time. The median logic circuit then selects the value of the undelayed input signal when it is determined that the value of the differentiated signal is greater than the value of the unmodified (not delayed) input signal. The output signal processed in this way is a sharp step signal without overshoot or undershoot, whereby a differentiated signal is selected for sharp transients. When using this sharpened edge in processing the color signal, the sharpened edge is provided to achieve an improved color transition. A derivative of the input signal delayed by a first time is obtained, preferably using a bandpass filter, to produce an output signal having a sharp falling edge, the value of this differential signal being delayed twice. The median logic circuit selects the value of the input signal delayed twice until the value becomes smaller than the value.
The median logic circuit then selects the value of the differentiated signal until the value of the differentiated signal is less than the value of the undelayed input signal, and the value of the differentiated signal is less than the value of the undelayed input signal. Select the value of the input signal that is not delayed when

[0007] The method of the present invention includes the steps of delaying an input signal by a first time and further delaying the input signal by a second time that is longer than the first time. The derivative of the input signal delayed by the first time is also obtained. The median logic circuit generates an output signal by selecting any value of an undelayed input signal, an input signal delayed by a second time, or a derivative of the input signal delayed by a first time. I do. First, at the rising color edge, the input signal delayed by the second time is selected until the second derivative of the input signal is greater than the value of the input signal delayed by the second time. The differentiated signal is selected until the value of the differentiated signal becomes equal to the undelayed input signal, and when equalized, the undelayed input signal is set as the output signal.
The same applies to the signal selected for the falling edge in order to sharpen the falling edge by selecting the appropriate signal. The processed output signal from the median logic circuit thus has a steeper slope than the input signal, and these slopes are very sharp with no undershoot or overshoot. The median logic circuit is 3
Just select the value from one of the two signals,
It can be realized with only a minimum number of SVP instructions, and ASIC
Alternatively, it can be realized by a small number of gates on the FPGA. Therefore, the DCTI algorithm of the present invention is relatively simple and inexpensive.

[0008]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS For the purpose of describing and clarifying the present invention, the present invention will be described with reference to processing of color signals having a limited bandwidth. It should be understood, however, that the present invention is well suited for processing various bandwidth limited signals, and that the present invention has application to a wide variety of signal processing. It is. When using the present invention for processing color signals, the present invention can be part of a digital receiver (not shown), and for clarity of one of the uses of the present invention, The invention will be described with reference to the processing of color signals.

Next, reference is made to FIG. Here, a graph of a representative color signal 14 showing a color edge in which the slope of the input color signal is small (the rising time is long) due to the limited bandwidth of the input signal is shown. To increase the color sharpness of the image to be displayed, it is desirable to make the slope of the color edge steeper as indicated by the dotted line 18. Here, it should be noted that it is desired to make the slope of the signal 18 steep and eliminate undershoot or overshoot in this signal.

Referring now to FIG. Here is a digital color transient improvement (DCTI) of the preferred embodiment of the present invention.
The entire block diagram of the device and algorithm is numbered 10
Indicated by Circuit 10 receives an input signal at input terminal 12, for example, a color signal indicated by numeral 14 in FIG. 2, and an output signal at output terminal 16, for example, FIG.
Outputs a desired color signal having a sharp edge as shown by numeral 18. By outputting the output signal 18 having no ringing and having a sharp color edge at the terminal 16, the color sharpness is improved. This improved color transient signal is generally provided to a digital television receiver (not shown) that can include the present invention.

The digitized color signal, represented as the U or V component of the video signal, is applied to input terminal 12. This input color signal looks like signal 14 in FIG. First time delay circuit 20 delays the input signal by a first time T, preferably about 0.148 microseconds (sampling frequency fs = 13.5 MHZ), and outputs the signal delayed by this time to node 22. I do. The second time delay circuit 24 further delays the signal from the node 22 by the time T, and outputs the signal twice as long as the signal C to the median logic circuit 30.

The once-delayed signal from node 22 also
It is sent to a band pass filter 32 as shown. The bandpass filter 32 generates a derivative of the input signal delayed once. This derivative approximates a mathematical derivative, which is applied to amplifier 36. This amplifier 36 generally has about twice the gain to control the slope of the transient section. The amplified signal thus amplified is then applied to coring circuit 40 to minimize noise effects and small signal transients. This coring circuit 4
The transfer function of 0 is shown in FIG. The output signal of the coring circuit 40 is added with the output signal delayed once by the adder 42 from the node 22. This adder 42
Are clipped by the clipping circuit 44 to eliminate underflow or overflow in the signal computation. The output of the clipping circuit 44 is applied to the median logic circuit 34 as a differential signal B. The unprocessed (non-delayed) input signal 14 from input terminal 12 is applied as signal A to median logic 30 as shown. Median logic circuit 3
A 0 selects the value of one of the input signals A, B or C and outputs the value of the output signal 18 to the output terminal 16, as will be described in more detail below.

Referring to FIG. 5A, there is shown input signal 14 applied to input terminal 12. The median logic circuit 13 samples the value of the input signal 14 as the signal A at the displayed data sampling point. Referring to FIG. 5B, the processed derivative (corresponding to the differential signal B) of the input signal delayed by a first time, generated by the bandpass filter 32 and its associated signal processing circuit, is shown. It is shown. The value of the differential signal B is also sampled by the median logic circuit 30 at the same time as the sampling of the input signal 14 as shown. Referring to FIG. 5C, there is shown a twice delayed input signal C, the value of which is also sampled by the median logic circuit 30 at the indicated sampling point. The median logic circuit 30 obtains a median value between the values of A, B and C from these three signals. The sampling points of the black circles are the values selected by the median circuit 30, and constitute the output signal 18. As shown, the signal 18 available at the output terminal 16 of the median logic circuit 30 has sharp rising and falling edges without undershoot or overshoot, the slope of which is much greater than the slope of the input signal 14. It is steep.

Referring to FIG. 5C, during time t ₁ , median logic circuit 30 selects the input value of input signal C, which has been delayed twice, as the value of output signal 18, which is the differential signal B values are selected until it is determined that becomes larger than the value of the signal C shown between times t _2. Until it is determined that the value of undelayed input signal A is less than the value of input signal B indicating overshoot of signal B, the value from signal B is selected and median logic circuit 30 is selected.
Is output as an output signal 18. During time t ₃ , the value of input signal A, which has not been delayed, is selected as output signal 18 to median logic circuit 30 and output terminal 1
6 is output. Until the median logic circuit 30 detects that the value of the signal A has become smaller than the value of the signal C, the value of the input signal A that is not delayed is selected.

During time t ₄ , the output value of signal C is selected. The value of signal C is selected until it is determined that the value of differential signal B has become smaller than the value of signal C. Next, during the time t ₅ , the output signal 18 having a slope with reduced sharpness is output.
The value of signal B is used to output Lastly, the value of the signal B is for a time t _6, which is lower than the value of the signal A, the value of the signal A is selected.

The value selected by the median circuit 30 and output to the output terminal 16 as the signal 18 is indicated by a black dot in FIG. 5C. Output signal 18 especially during transient
It should be noted that the sharpness of the slope and the undershoot or overshoot are gone. Also, the algorithm is such that only three signal values are sampled by the median logic circuit 30 and only two of the signals are processed to obtain a clean output signal 18 output to the output 16. It should also be noted that is fairly simple, straightforward and progressive. Further, such processing is performed by processing the differential of the raw input signal, the input signal delayed twice, and the input signal delayed once.

Referring now to FIG. 6, there is shown an electrical block diagram of a basic 15-tap FIR horizontal bandpass filter that can be used as filter 32 to obtain the derivative of the input color signal delayed once. It is shown. The amplifiers of the signals output from each tap are denoted as K0 to K14, and the constants of these amplifiers are shown on the right side of the block diagram. This filter 32 outputs a differential signal of the input signal that has been delayed once, which is an approximation of a mathematical derivative. The output of this filter is an addition signal of the signal amplified and output from the tap.
6 is added.

Referring to FIG. 7, the frequency response and gain of the horizontal bandpass filter 32 are shown, the sampling frequency fs is 13.5 MHZ,
The center frequency is about 1 MHZ. The filter characteristics are selected to enhance the high frequency components in the chrominance signal that are limited by the process of limiting the bandwidth of the chrominance signal to 0.5 MHZ.

Referring now to FIGS. 8A and 8B, there is shown the processing of the chrominance signal V and chrominance signal U by SVP, and Table 1 below shows this processing in more detail. The transition shows a color bar transition between green and magenta.

[0020]

[Table 1]

Referring to the above table, there are shown the digital input values of the color signal V and the color signal U, the data points of which are shown in FIGS. 8A and 8B, respectively. Identified in the annotation. Y, U,
The standard digital video format for the V video signal is 4: 2: 2, the Y signal is sampled at 13.5 MHZ, and the U and V signals are sampled at 6.75 MHZ. The standard method is to perform the subsequent processing (ie, C
TI) and conversion to RGB space is 13.5M
The use of a complementary filter to double the U, V sampling rate by filling in the missing values, as done at the HZ rate. The columns of V _in and U _in show color data in 4: 2: 2 format, and the columns of VITP and UITP show color data complemented in 4: 4: 4 format. The output values of the color signals U and V processed by the color transient improvement (CTI) algorithm of the present invention are shown in Table 1 under the heading VCTI.
And under UCTI, these values are shown in FIG.
And 8B are also plotted and identified by annotations.

[0022] Referring first to FIG. 8A, when the inclination of the input data V _in respect to the color signal V and the interpolated color signals VINT from this figure is quite gentle, the output signal V
The slope of the CTI is steep, from about 56 units to 199.
It can be understood that it changes into a unit in a short time.

Similarly, referring to FIG. 8B, for the input value U _{in of the} color signal U and the interpolated color signal UITP plotted so as to have a gentle slope, the CTI
The slope of the output signal UCTI obtained by the algorithm is also steep, and the output value increases rapidly from about 100 units to 155 units.

Next, referring to FIG.
A representative input color difference signal BY and a color difference signal R before I
-Y is indicated. Referring to FIG. 9B, here C
The respective signals of FIG. 9A after the TI are shown, with the corresponding signals having sharpened edges on both rising and falling as shown. FIG. 9A
9B correspond to the SVPDOR trace data which is a full color bar test pattern.
See Table 2 below. Shown here is a set of instructions for SVP2 to implement the algorithm of the present invention.

[0025]

[Table 2]

[0026]

[Table 3]

As shown, the algorithm of the present invention requires only 176 SVP instructions. Therefore, the present invention is a fairly simple architecture for performing logical operations, and is considered to be realized by digital signal processing. The invention can also be easily implemented with a small number of gates on an application specific integrated circuit (ASIC) or field programmable gate array (FPGA). In a typical SVP2, a discrete processing element (P
E) is provided. Thus, each PE of the SVP is programmed with the instructions in Tables 2 and 3. The instructions shown in FIGS. 2 and 3 are those of the preferred embodiment, but are not meant to imply a limitation on these instructions.

In summary, the CTI method and apparatus of the present invention
The installation is fairly simple, linear and progressive, cheap
Such as those used in digital signal processing televisions
Well suited for processing bandwidth limited color signals
ing. Preferably a microprocessor based circuit
Is a simple median logic circuit 30 that has three input signals
Sample all of A, B and C and see FIG. 5C.
As shown and described, of these three signals
Choose a value from one of The resulting output signal 1
8 controls the rising and falling portions of the signal.
It has a sharp inclination. The value of the input color signal delayed twice
Time t₁Used during time t_TwoAnd t _FiveDuring the differential signal
The value of signal B corresponds to a sharp, that is, steep, slope
Use this as it is. Time t_TwoAnd t_FiveBetween
By using the value of signal B only for
The dark or overshoot parts are not used.
Not. The input color signal which has not been delayed is converted to a time t._ThreeAmbassador
And the input signal C delayed twice is used for a time t._FourUse during
And the value of the undelayed input signal A is₆Ambassador
To use. In summary, the median logic circuit 30 has three inputs.
The values of the force signals A, B and C are sampled,
Select values from two signals and have a steep transition slope.
Output signal 18 is generated.

The present invention is applicable to digital processing of a signal having a limited bandwidth, for example, a color signal having a limited video. While the invention has been described with reference to a particular preferred embodiment, many variations and modifications will become apparent to those skilled in the art upon reading this specification. Therefore, the appended claims should be interpreted as broadly as possible with respect to the prior art to include all such variations and modifications.

With respect to the above description, the following items are further disclosed. (1) (a) input means for receiving an input signal;
(B) first input means for delaying the input signal by a first time and responsive to the input means to generate a first signal; and (c) delaying the input signal for a second time. (D) second delay means responsive to the input means to generate a second signal, and (d) differentiating means responsive to the first delay means to generate a differential signal of the first signal; e)
A median receiving three signals including the input signal, the second input signal, and the differential signal, selecting from the three signals, and generating an output signal having a steeper slope than the slope of the input signal. A signal processor comprising a filter means. (2) The signal processor according to (1), wherein the second time is longer than the first time. (3) The first means, wherein the differentiating means includes a band-pass filter.
A signal processor according to claim 1. (4) a gain amplifier for receiving the differential signal;
A signal processor according to claim 3. (5) The signal processor according to (4), further comprising a coring circuit for processing the differential signal. (6) The signal processor according to (5), further comprising an adding circuit for adding the differential signal from the coring circuit and the input signal delayed by the first time. (7) The signal processor according to (5), further comprising a clipping circuit for processing the differential signal from the coring circuit.

(8) (a) an input terminal for receiving an input signal; (b) a first delay circuit connected to the input terminal for delaying the input signal by a first time; A) a second delay circuit connected to the input terminal for delaying the input signal by a second time longer than the first time;
(D) a differential circuit connected to the first delay circuit for generating a differential signal of the input signal delayed by the first time; and (e) the input signal delayed by a second time. A median logic circuit for selecting from the input signal and the differentiated signal and generating an output signal having a steeper slope than the slope of the input signal.

(9) (a) sampling an input signal; and (b) applying the input signal to a first time and the first time.
Delaying by a second time longer than
(C) generating a differential signal of the input signal delayed by the first time; and (d) selecting from the input signal, the input signal delayed by the second time, and the differential signal. Generating an output signal having a steeper slope than the input signal. (10) The method according to (9), wherein the differential signal is obtained by processing the input signal using a bandpass filter. (11) To generate a steep transient portion of the output signal,
The method of claim 9, further comprising the step of selecting the differentiated signal. 12. The method of claim 11, including selecting the differentiated signal as the output signal when the value of the differentiated signal is greater than the value of the input signal delayed by the second time. (13) selecting the input signal delayed by the second time as the output signal when the value of the differential signal is smaller than the value of the input signal delayed by the second time. Item 13. The method according to Item 12. 14. The method according to claim 12, comprising the step of selecting said input signal as said output signal when the value of said differential signal is greater than the value of said input signal. The method of claim 11, further comprising the step of selecting the differentiated signal as the output signal when the value of the differentiated signal is less than the value of the input signal delayed by the second time. (16) selecting the input signal delayed by the second time as the output signal when the value of the input signal is smaller than the value of the input signal delayed by the second time. 16. The method according to claim 15, wherein 17. The method of claim 15, further comprising the step of selecting the input signal as the output signal when the value of the input signal is greater than the value of the derivative signal.

(18) A digital color transient improvement (CTI) method and apparatus for enhancing the color sharpness of a color signal by increasing the steepness of a color edge without causing ringing, wherein a median logic circuit (30) is provided. The three signals A, B, C are sampled and values from these three signals are selected to generate an output signal having steep color edges. The signal (A) is the unprocessed input signal, the second signal (C) is the input signal delayed twice, and the third signal (B) is the derivative of the signal delayed once. The present invention is ideally used with SVP and requires a minimum number of instructions, but could be implemented with a small number of gates on an (ASIC) or (FPGA).

[Brief description of the drawings]

FIG. 1 is an electric block diagram of a conventional DCTI circuit.

FIG. 2 is a waveform diagram of a representative baseband chrominance signal waveform with gradual color edges and a desired chrominance signal with steep color edges shown by dotted lines but without ringing.

FIG. 3 shows a median logic circuit generating an output signal by selecting a value from one of three signals to obtain a steep color edge as shown by the dotted line in FIG. 2; FIG. 2 is a block diagram of an electric algorithm according to a preferred embodiment of the present invention.

FIG. 4 is a diagram illustrating a transfer function of the coring circuit of FIG. 4;

FIG. 5A is a diagram showing an input signal A; B is a diagram illustrating a differential signal of the input signal delayed by a first time corresponding to the differential signal B. C is a diagram showing three input signals A, B and C selected by the median logic circuit, wherein the composite output signal of the median logic circuit is represented by a signal having a series of black dots.

FIG. 6 is an electrical block diagram of a typical 15 tap FIR bandpass filter with fifteen coefficients of amplifiers K0-K14 shown on the right and the filter having f (sampling) equal to 13.5 MHZ.

FIG. 7 is a diagram showing the amplitude of the frequency response of the 15-tap band-pass filter shown in FIG. 6;

8A is a diagram showing input and output of a color signal V. FIG. B is a diagram showing the input and output of the color signal U.

FIG. 9 is a diagram showing color difference signals BY and RY before and after a CTI algorithm.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 10 Digital color transient part improvement circuit 12 Input terminal 14 Color signal 16 Output terminal 18 Output signal 20 First time delay circuit 22 Node 24 Second time delay circuit 30 Median logic circuit 32 Band pass filter 36 Amplifier 40 Coring circuit 42 Adder 44 clip circuit

Claims (2)

[Claims] (A) input means for receiving an input signal; and (b) first input means responsive to the input means for delaying the input signal by a first time and generating a first signal. (C) second delay means responsive to the input means for delaying the input signal by a second time and generating a second signal; and (d) differentiating the first signal. Generating the signal
And (e) receiving three signals including the input signal, the second input signal, and the differential signal, selecting from the three signals, and determining a slope of the input signal. And a median filter means for generating an output signal having a steeper slope. 2. A step of: (a) sampling an input signal; (b) delaying the input signal by a first time and a second time longer than the first time; Generating a differential signal of the input signal delayed by a first time; and (d) selecting from the input signal, the input signal delayed by the second time and the differential signal, Generating an output signal having a steeper slope than the input signal.