JPH0319492A - Color blurring correction device - Google Patents

Abstract

PURPOSE:To attain highly accurate color blurring correction over the entire pattern by selecting the number of relay grating points to be 4 times the number of representative grating points or over and to be 3 or over for each horizontal blanking period and interpolating the color blurring correction to remaining points based on a stored data so as to adjust a few representative grating points. CONSTITUTION:Signals XU, YU are fed to a source correction function generating means 11 to generate 25 kinds of correction functions fij(XU, YU). Thus, the permutation of (i,j) is 25 ways and the total number of permutation of the (XU, YU) is 256 and they correspond to grating points being the division of the screen into 16X16. Then the function value of fij(XU, YU) is inputted to a weight storage means 12 and an output defined by SIGMAkjifij(XU, YU) is obtained by the combination of 10-bit variable weight constant kji being an output of a weight coefficient generating means 13. Then the constant kji is easily and definitely decided depending on the quantity of the color blurring correction of 25 representative grating points. Thus, a few representative grating pints are used to correct the color blurring of the entire screen.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a high-definition projection display, and more particularly to a color shift correction device using a rational construction method of a digital convergence circuit.

[Conventional technology]

Conventional digital convergence [g! Jji! As shown in Fig. 2, color blur correction is performed by providing approximately square lattice points of about 100 points divided almost equally on the screen, and for each lattice point among the 100 points, R * G * B This method corrected the positional deviation of the five colors.

[Problem to be solved by the invention]

Although the above-mentioned prior art has good adjustment accuracy, there is a problem in that the time required for adjustment is extremely long. In order to overcome this problem, various efforts have been made to reduce the number of lattice points, but no satisfactory solution has been obtained.

The reason for this is that the method of interpolation between grid points, for example,
As described in Japanese Patent No. 55310, the problem lies in relying on a straight line approximation or curve approximation method that focuses solely on the vicinity of lattice points.

The present invention achieves this by adjusting only a small number of lattice points.
A digital convergence circuit suitable for so-called multi-scan displays, which uniformly, highly precisely and effectively reduces color shift across the entire screen, and can support signal sources with different horizontal and vertical scanning line frequencies. An object of the present invention is to provide a color shift correction device.

[Means for Solving the II Problem] The above purpose is to provide a means for generating a plurality of source correction functions consisting of polynomials of at least fourth degree regarding two-dimensional position coordinates on the screen, and means for determining weighting coefficients for the plurality of source correction functions from the required correction amount of color shift for the following representative grid points; means for storing and accumulating the plurality of weighted source correction functions; This is achieved by providing means for calculating a correction 1k depending on the scanning position coordinates of the display based on a source correction function, and means for performing an auxiliary deflection of the electron beam depending on the output of said calculating means. .

[Effect]

The multiple source correction functions provide correct color shift correction t't only for a small number of representative grid points. Therefore, each point on the remaining screen includes an interpolation error.

However, the above-mentioned interpolation error is practically negligible due to the physics of the projection optical system and the physics of deflection of the electron beam. Therefore, good convergence performance is ensured over the entire screen, and color shift can be corrected.

〔Example〕

Embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 is a block diagram showing an embodiment of a color shift correction device according to the present invention, in which single-line arrows indicate analog signals or single-line arrows.
In the pit digital signal, the D1 double arrow indicates a multi-bit digital signal.

In the figure, 1 is a horizontal retrace pulse input signal generated by a separately known horizontal deflection circuit, 2 is a vertical sawtooth wave input No. 48 generated by a separately known vertical deflection circuit, and 3 and 6 are, for example, monochrome signals. The delay elements are multi-bibrator, and the delay times are τ, , τ2, respectively. 4 is a phase detector, 5 is yco
(tIE pressure controlled oscillator), 7 is an 8-bit counter.

The phase detector 4.vcos oscillation frequency f0 is 256 times the horizontal frequency. The PLL section of VCOS, delay element 6, and counter 7 operates.

8 is a reproduction horizontal pulse output from the counter 7, and 9 is a reproduction horizontal pulse delayed by T2.

FIG. 4 is a signal waveform diagram for explaining the operation of the PLL section, and the symbols attached to the waveforms are the same as those of the first vA. Typically, i, is set to (L 5 T r@t, where 'l'ret is the horizontal retrace period.

The phase detector 4 in FIG. 1 compares the falling υ part of the signal y with the falling υ part of the signal 9, and the PLLLoop works to match both ties. As a result, the reproduction horizontal pulse 8 has a timing ξ advanced by τ from the center of the input retrace pulse 1. It is assumed that τ2 has a value approximately equal to the delay time of the amplifier 17.

An 8-bit signal is obtained from the digital output of the counter 7. Let this be X, its upper 4 bits be X(1, and its lower 4 bits 'kXt). These can be considered to correspond to the horizontal coordinate of the screen scanning position.

Hereinafter, specific numerical values will be explained assuming application to a high-definition projection display with a display pixel count of approximately 1M pixels.

A 12-bit digital signal is obtained from the AD converter 10. Let this be Y, and its upper 4 bits t” Y
g sT 8 bits} IM. These can be considered to correspond to the vertical position of the screen scanning position.

Hereinafter, XsY will also be referred to as an X address signal or a Y address signal.

Signal X. , YO are applied to the source correction function generating means 11;
25 [generate a correction function fiJ (Xus Yo ) of the type.

fzJ(Xa − Yo)K>(1.j)0 There are 25 permutations of Ab1, which corresponds to 25 types.

The total number of permutations of (XO*Yo) is 256, which corresponds to each grid point obtained by dividing the screen into f 6×1 6K. However, these are not representative grid points for coherence adjustment, but relay grid points for convenience of calculation processing.

tiJ(Xa+YO)O seki 1s [[t 8 k'
7 } Expressed in O precision, source correction function generation means 1
1 is a ROM of 16×1 4×8 bits=256 bytes in terms of specific hardware.

This function value is then input to the weight storage means 12 and combined with the 10-bit variable weighting constant k, which is the output of the weighting coefficient generation means 15, to obtain the 51 Σksjt. We obtain the output defined by J(Xg t Yo) where Σ is performed for 25 permutations of (i,j). kLJ is easily and uniquely determined according to the magnitude K of the color shift correction amount of the 25 representative grid points.

As the specific hardware of the weighted accumulation means 12, a micro combinator or a well-known product-sum processing circuit is used.
The processing output is stored in 112FROM as a 12-bit number. ]it2FROM(2) Required storage capacity is C
1 RT per 10,000 p, 16 x 16 x 12 bits = 30
It is 72 bits.

Usually, CRTs are R, G. Since three bits B are used and the correction directions are 20,000 horizontal and vertical, the total is 3072 bits x 5 x 2 = 18452 pits. If there are many types of signal source formats, the above storage capacity is used for each 7 formats.

Since the output of the weighted accumulation means 12 is a function of only (Xσ*Yt+), it is input to the interpolation means 14 according to the address of Xg+YH.

The interpolation means 14K is also the lower 4 bits of the counter 7! , and ^ Y1, which is the lower 8 bits of the output of the D converter 100, is input to the latch circuit 15 based on the reproduction horizontal pulse 8K.
The latched output YX, is applied.

The output F(X, Y) of the interpolation means 14 is expressed by the following equation.

・・・・・・・・・・・・・・・・・・・・・・・・■
Here, F (X g * Y g) = Σki1
fi4(XgtYo)......■ Formula ■ is a linear interpolation formula. As a variant, Sara K F (
XO Jo+2) * FcX,,, sYg12)
It is also possible to use the curve Fukuma.

This operation can be performed by a combination of multipliers and adders.

Output F(X, Y) is a 12-bit ODA converter 16
is applied to become an analog signal. This analog signal is amplified by an amplifier 17 and applied to an auxiliary deflection yoke 18, thereby performing auxiliary deflection of the electron beam according to F(X.Y).

Although not shown in the figure for simplification, since the direction of color shift is two-dimensional (horizontal/vertical), the weighted accumulation means 12, interpolation means 14, DA converter 16, amplifier 17 shown in the figure
One more set of 0 parts is used for each color, totaling two sets. The main deflection yoke 19 is driven by a separately known deflection circuit. 20 is CR! It is.

The well-known crosshatch signal generation circuit wt21 generates a crosshatch signal consisting of 16×16 grid points based on the input of XHYO. The lattice points of this circuit #is x a may be used. In that case, a crosshatch signal may be generated based on each of the upper three bits input to the XgeMCI.

Furthermore, crosshatch signals corresponding to 5×5=25 representative grid points corresponding to the representative grid points of the present invention may be generated. The pattern of this representative grid point crosshatch signal is shown in FIG.

In any case, its output 22 is delayed by a delay resistor 25, at least the signal representing the vertical line, so that CR? 2
0 cathode electrode.

Next, the 25 source correction functions 'IJ(XO
,YL,) will be explained.

The IO+YgO value is 0,
The values are ±4 and ±6.

Corresponding to this, 1. The jO value will be written as 0, ±4, ±6. It is simplified and expressed as X, Y instead of Xg*Mg.

The 25 source correction functions fIJ (Xy * Yo) are A2
(x) + B4 (X ,)') + A4
(x)e 86(X.y)J6(x*7) by the following formula■~
By defining and using κ written at the beginning of each of ■, the and j? at the end of each of the following formulas ■ to ■ 6 was selected. Inside, here A s (x) − B @ (x)
The subscript m corresponds to an m-th degree polynomial, and the signs 2 and 2 mean the same order on the left and right sides of the equation.

'o, o(X+Y) =' ・・・・・・・・・
■f±4,,(X.Y) =A2(±X)fo,±4(
X, Y) = A2 (±Y) ・・・・・・・・・・・・■ f Sat. , soil, (X, Y): B4 (soil X, ±Y)...
・・・・・・・・・■f±6,. (X.Y) = A4
(±X) f0+±6 (x, y) = A4 (±Y)・
・・・・・・・・・・・・■ f±6,±4(X,Y),. = 86 (±X, ±Y) f soil 4, ±6 (X, Y) = B6 (±X, ±Y)...
・・・・・・■ f Sat. , soil, (X.Y):B, (±X, soil Y)...
・・・・・・・・・■ The weighting coefficient kiJ of each function mentioned above in formula ■ is shown on the screen (
It is set so that the color shift at the representative grid point corresponding to Xo+Yo)=(itj) is minimized.

For example, the coefficients of Ik4 and -4 are adjusted and determined by focusing only on the representative grid point at the upper right corner of the screen.

The adjustment order is performed in the numerical order shown in FIG. The numbers shown in the figure correspond to the numbers in formulas (1) to (2).

Adjust from the smallest number to the largest number.

The degree of independence between the above functions will be explained next.

For example, fo,6(XIY) is obtained from the formula ■, so
This function is expressed as Yσ=O I M,=S49YO=-6oi![i
It is always zero in I state.

Therefore, in formula ■. Even if the k,6O weighting coefficient is varied, the corresponding correction amount does not change at the positions on these dotted lines.

That is, (1・j)=(0.0)・(±4*O)+(
O-±4), (±4.±4)w(-6*O)*(6sO
)e(0+-6) The correction amounts for a total of 12 representative grid points are not changed. This means that for the correction of color shift, the function f. ，
6 is f,,. , f±4, otfos±4sf±4t±4
, r-6, Olf6, 0, fo, -4.

In general, any function (n-4 to 8) in formula (■) is (n
-1) It has an orthogonal relationship with the following equation, and it also has an orthogonal relationship with a function other than itself in the equation (2).

Therefore, once correction adjustment for color shift is performed by focusing on each of the 25 representative lattice points in the order of the numbers in FIG. 5, there is no need to perform readjustment again.

In FIG. 5, grid points with the same number have an orthogonal relationship with each other, so the order of adjustment is arbitrary. However, care must be taken because the function "+3" corresponding to grid points with small numbers also affects grid points with large numbers.

Next, a second example will be described.

Here, we will analyze the color shift generation process in the optical system of a projection display and the CRT auxiliary deflection, and clarify the inherent irregularities.

Now, the process of color shift generation will be explained.

Figure 7 is a diagram showing the pattern of trapezoidal distortion generated in the projection optical system, and the magnitude of the trapezoidal distortion is . It is approximately proportional to the concentration angle ω and the tangent α of the angle of view α.

FIG. 8 is a diagram showing an analysis of trapezoidal distortion occurring in the projection system, in which the origin (oeos') is set at the center of the exit pupil of the green projection lens and x. 7. Place the g-axis in the direction of the arrow in the figure.

The y-axis is ima on the paper. 24 is screen D1
The coordinates of the green m image on it are (x g・!a*z(1)
shall be. It is assumed that the inner axis of the lens is fil[ on the screen 24.

In order to find the distortion of the red image, instead of arranging the red lens at an angle, an inclined screen 25 is assumed for the red color. The coordinate points located in the same direction on this virtual screen 25 when viewed from the lens are assumed to be (xR, yR, zR).

R.G. When congruent figures are displayed on each CR'J'' of B, the color shift between OR and G on the screen is given by the difference between the green and red coordinates.

x1 and X considering that Kg is a constant. Seek a relationship with. From the same figure, ZG ZB! , 1--α・1ω Xa )'o...
......O The above formula represents the size of color shift O measured on the screen.

10,000, the required amount of color shift correction should be calculated by converting it onto a green screen. This is because, under the practical condition that the geometric distortion of the projection lens is small, the coordinates on the crystal screen and the coordinates on the CRT are proportional. This is because there is a relationship between

That is, the corresponding coordinates on the green screen (x'6*7'g *zG
) to find the required correction amount (XG-x'o) From the formula [phase] and the $8 diagram, x/oy I. 1−−α′・1ωE0? ■-X'g x0 7G-)"G 7a = one α' ] from -α=-ZQ ・・・・・・・・・・・・・・・・・・OZ0 The above equation represents the required amount of color shift correction.

In the normal projection system, the value of (-α・1ω) is within about 1, i. e. Within about 10%.

Next, a description will be given of nonlinearity that occurs accompanying the auxiliary electromagnetic deflection of an electron beam for color shift correction.

D. Written by GJink'Tslevision Engine
eering Hand-book” pages 6-1 to 6-
According to the electromagnetic deflection analysis described on page 6, the sine of the deflection angle is proportional to the deflection tL flow. Therefore, if the unit system is chosen appropriately, the two can be equidistant. That is, Yuθ8=I8 Yuθ=Work! ! Here, tLtI, is horizontal, vertical deflection current θ, ri, θ are horizontal, vertical, total deflection angle is 10,000, and the deflection angle (XIY) on the CRT fluorescent aperture (planar surface) is given by the appropriate unit system. , is equal to the tangent of the deflection angle, that is, txlIsI θI.

・・・・・・・・・・・・・・・・・・・・・・・・＠
Therefore, the deflection amount increments ΔX, ΔY corresponding to the auxiliary deflection current increments ΔI, ΔI are obtained by the following equations as a result of differential calculation.

``Kufu a (101) ・・・・・・・・・・・・・・・@ Particularly, seagull 1 = , 1 -dn” 11, -da
'A must in combination with Bow 72〇! ! Auxiliary deflection current Δl
1Δk, p, the following equation is obtained.

Δengine, ・・・・・・・・・・・・・・・［phase］? Koni
l( w■ Zy If the above equation is ΔI=kx2 0 ΔI y =k x G7 6 , then in equation (■), the purpose of color shift correction can be achieved with only the term x2sxx. However, In reality, the formula O includes the non-reciprocal factor. Therefore, the formula ■
~0 and bB. It is necessary to include higher order terms.

The coordinate system of Th, X, Y is the above-mentioned bj5cR? The surface of the firefly is divided into about 16 horizontal and vertical synchronous sections, and in FIG. 1, X is proportional to -I8 and Y is proportional to blood #a.

From the relationship, formula 9#iX. It can be expressed as a function of Y,
Furthermore, it can be expanded to a polynomial by Taylor.

Is9 diagram is a diagram showing the main part of the 20th embodiment, and is 31
is the formula Qo function value as a function of (IOeYg) and 1tS
This is a memory element that stores the values at the relay points of XId in advance.

The output of this is added and accumulated by the addition accumulation means 12GC& of FIG.

Compared to the first embodiment, the second embodiment has the advantage that the time required for adjustment is further reduced.

For practical applications, the value of g (XIJI YO ) may be determined in advance through experiments instead of calculating the value of the formula Oo.

Next, some modified examples will be explained.

As the 1st modification, *z(x)~ in formulas ■~■
B@(X+7) may be given the following formula by performing 44+6 substitutions.

The required minimum number Nm of horizontal relay grid points is determined by the following equation.

・・・・・・・・・・・・・・・・・・O In this case, the 5th F! ! There is no difference in this trade except that substitutions of 508 and 406 occur in J.

In the second modification, the number of relay grid points is 1 dX1 6=2
Although the number of points was set at 56, it may be substituted with 12x12.

In this case, the 25 representative grid points are! ,=Ol±2,±4 Y,=OI±2,±4 shall correspond to

In particular, Trot is the horizontal retrace period! Yo is the horizontal scanning period. The meaning of this expression is the condition that two or more relay sections exist within the horizontal flyback period.

That is, this is a necessary condition for correcting color shift at the left and right edges of the screen.

The timing relationship between the relay grid points and the horizontal flyback period is shown in FIG.

Figure 10 shows representative grid points (long arrows) and horizontal retrace pulse 1.
The observation point in the figure corresponds to the point where the auxiliary deflection yoke 18 is reached after passing through the amplifier 17 in FIG. 1.

The long and short arrows represent a total of 12 relay lattice points.This figure helps to understand the meaning of the equation [phase].

i or f in formula ■. j (V.vO), the same value as the data corresponding to the right raco is used as the value of the correction data corresponding to the center of the horizontal retrace period (xrJ=±Nyo/2). That is, this is because correction of color shift within the retrace period is not necessary. In addition, since the horizontal retrace period is very short, the above-mentioned method is used in order to quickly respond to color shift correction.

Next, a fifth modification will be explained.

Although the case where the number of representative grid points is 5×5=25 has been described, nine points can be substituted for applications that do not require high definition.

In this case, A 2 (x) s B4 (x,
If you use only y), you will get 9 feet.

Furthermore, a modification of #I4 will be explained.

In Figure 1, [i[address signal Y is
J) wave 21 was obtained by converting with an AD converter, but a counter for counting the number of horizontal retrace pulses may be used for each vertical retrace.

Furthermore, a fifth modification will be explained.

In this 1, it has been assumed that the CR'l'' fluorescent surface and the screen surface are flat, but it can also be applied when these are curved surfaces.It can also be applied to cases and direct-view high-definition displays. be.

Among them, conventional technology uses a nonlinear element such as a diode to create a waveform that exists only in a specific diagonal corner of the image direction and is zero in other areas, and is used to create a local color image. There have been attempts to perform misalignment correction.

However, there is no physical reality that generates a distortion factor corresponding to a non-linear element such as a diode. Therefore, such means have the disadvantage that the adjustment accuracy is poor or the number of adjustment points is excessive.

〔Effect of the invention〕

As explained above ii2, according to the present invention, the orthogonal function expansion method is based on the physical reality. Highly accurate color shift correction can be achieved. Furthermore, since this method eliminates interference between representative grid points, there is no need for repeated adjustments. Therefore, its industrial value is high.

In addition, when applied to multi-scan projection displays, one representative grid point can always be maintained at approximately the center of the horizontal retrace period, making it possible to effectively reach every corner of the screen.
The color misregistration O correction can be performed by using the present invention, and a color misregistration correcting device with excellent functions can be provided, except for the problems of the prior art described above.

[Brief explanation of drawings]

ate is a block diagram showing an embodiment of the color shift correction device according to the present invention, FIG. 2 is a diagram showing 100 adjustment grid points of the prior art, and FIG. 5 is a diagram showing the pattern of a representative grid point crosshatch signal. Figure 4 is a signal waveform diagram for explaining the operation of the PLL section, Figure 5 is a diagram showing 25 adjustment grid points, Figure 6 is a diagram showing an example of representative grid points, and Figure 7 is a diagram showing an example of representative grid points. Is the Ha diagram, a diagram showing the trapezoidal distortion pattern that occurs in the projection optical system, a projection optical system? Figure 9 shows the analysis of the trapezoidal distortion that occurs. Figure 9 shows the main part of the second embodiment. Figure 10 shows representative grid points (long arrows).
3 is a diagram showing the timing relationship between the horizontal retrace pulse 1 and the horizontal retrace pulse 1. FIG. 1...Horizontal am pulse, 2...Vertical saw D wave, 5.6.25...Delay element, 4...
...Phase detector, 5...VC0,7...
...Counter, 8.9...Horizontal pulse, 10
. . . AD converter, 11 . . . Source correction function generation means, 12 . . . Weighted accumulation means, 13.
... Weighting coefficient generation means, 14 ... Interpolation means, 15 ... Latch circuit, 16 ... D
A converter, 17...Amplifier, 18...
...Auxiliary deflection yoke, 19...Main J direction w-1.
20...CRT, 21...Crosshatch signal generation circuit, 22...Crosshatch signal output, 31...Memory element.

Claims (1)

[Claims] 1. Means for generating a source correction function consisting of a plurality of polynomials including polynomials of at least fourth degree or higher regarding two-dimensional position coordinates on a surface, and the number of source correction functions is nine or more. The number of source correction functions is 25 or less, and the source correction functions are arranged to face the representative grid points on a nearly one-to-one basis, and the relatively high-order source correction functions are arranged at representative grid points corresponding to different homogeneous and relatively low-order source correction functions. The color shift is fixed without changing, and each of the source correction functions is configured to control the color shift in the entire screen excluding the fixed point based on a continuous polynomial function, and the weighting of the source correction function is means for accumulating data values at each relay grid point of the sum, and the number of relay grid points is equal to 4 of the number of representative grid points.
A color shift correction device for a CRT display, which interpolates a color shift correction amount for the remaining points based on the accumulated data, and includes three or more points in each horizontal retrace period. 2. In the color shift correction device according to claim 1, a delay means is provided in the PLL circuit for generating horizontal coordinates on the screen, and the delay means causes the timing of the output pulse of the PLL circuit to be set approximately at the center of the horizontal retrace pulse. A color shift correction device that precedes the color shift correction amplifier by an amount equivalent to the delay time of a color shift correction amplifier.