JPS61274493A - Digital decoder - Google Patents

Abstract

PURPOSE:To prevent the quality of a picture from deteriorating on account of a Y/C separation and to obtain a still picture, quality improved after performing a movement adaptive type scanning line interpolation by using a movement adaptive type circuit as well as a picture adaptive type circuit in case of the Y/C separation. CONSTITUTION:Outputs of the picture adaptive type Y/C separation circuit 4 and a Y/C separation circuit 6 between frames seek an integrated Y/C separation data after weighing so that a larger movement makes a weight for the Y/C separation data by the picture adaptive type Y/C separation circuit larger and a smaller movement allows the weight for the Y/C separation circuit between frames to enlarge. Furthermore, being added to a IQ demodulator 16, C seeks I and Q and again I as well as Q together with Y are added to a IQ matrix circuit 17 to output RGB signals. Signals passing through a delay circuit 18 from a movement detection circuit 8 as well as RGB signals delayed by one field through a field delay circuit 20 are supplied a movement adaptive type scanning line interpolation circuit 19 to output the required R'G'B' signals from a terminal 21 after performing a scanning line interpolation to accommodate movements.

Description

Detailed Description of the Invention (Industrial Application Field) The present invention is a decoder that converts a color television signal into red, green, and blue RGB signals, and in particular outputs high-quality RGB signals. Regarding.

(Prior art) To convert a color television signal to an RGB signal,
Normally, a 2H comb filter (filter with comb-shaped characteristics) using a line memory is used to separate the luminance signal Y and color signal C (hereinafter abbreviated as "rY/C separation"), and the C signal is applied to an IQ demodulator to produce the I and Q signals. A method is used in which the required RGB signals are obtained by demodulating the signal and adding it to a matrix circuit together with the Y signal. When such a method is used, in the striped pattern near the color subcarrier fsc, the C signal has Y
The components leak in, resulting in cross color, and in portions where there is a sudden change in the vertical direction, the C component leaks into the Y signal, causing dot crawl, which impairs image quality. Several methods have been proposed to prevent this drawback, one of which is to detect motion by taking the difference between each corresponding pixel between two adjacent frames. performs Y/C separation and scanning line interpolation between frames,
In areas with a lot of movement, Y/ within each field.
Methods of C separation and scan line interpolation have been used.

(Problem to be Solved by the Invention) When motion is detected using the method described above, there may be a portion where there is movement between adjacent fields but there is no movement between frames, and no movement is detected. Since interpolation is performed between adjacent scan lines, a non-existent image or afterimage may be produced. In order to overcome this drawback, a method is used in which inter-frame differences are determined for adjacent fields, and when there is a significant difference in any of these, it is determined as motion. However, when this is done, there are parts where movement is detected even though there is no movement between adjacent scanning lines, and a new drawback arises in that appropriate interpolation cannot be selected.

Furthermore, even if Y/C separation and interpolation adapted to image movement are performed as described above, in each field plane, the C component leaks into the Y signal in parts where the image changes rapidly in the vertical direction, resulting in dots. Crawl occurs, parts with changes in the horizontal direction are emphasized, and thin lines in the vertical direction in particular become thicker or overlap, resulting in an unpleasant-looking image.

In the case of videos, even if some of the above-mentioned drawbacks are not noticeable, the drawbacks are noticeable when one screen of the video is frozen and viewed as a still image on a monitor.

It's coming. This tendency becomes even more obvious when it comes to hard copies, and the shortcomings that were not an issue when it came to video become a problem. For this reason, in devices that take hard copies, RGB from television signals is
Even as a decoder that converts signals, it is necessary to adopt a circuit system that does not degrade image quality as much as possible.

Therefore, an object of the present invention is to use a motion adaptive circuit and an image adaptive circuit for Y/C separation.
To provide a high-quality digital decoder capable of preventing deterioration in image quality due to /C separation and obtaining still images with improved image quality by performing motion-adaptive scanning line interpolation.

(Means for Solving the Problems) The digital decoder of the present invention is a decoder that converts a color television signal into red, green, and blue signals. An image adaptive luminance signal/chrominance signal separation circuit that performs separation, a motion adaptive luminance signal/chrominance signal separation circuit that separates luminance signals and chrominance signals according to image movement, and scanning line interpolation according to image movement. A motion adaptive scanning line interpolation circuit is provided.

(Function) The motion-adaptive Y/C separation circuit itself is a well-known circuit, for example, it was announced at the 1982 National Conference of the Television Society. The part without is the interframe Y
/C separation allows for ideal Y/C separation, and takes advantage of the property that moving parts cause lag between frames, so it is better to perform Y/C separation within the field to obtain better results. be. The circuit detects the movement of the image by taking the difference between frames, and weights the inter-frame Y/C separation and the intra-field Y/C separation based on this movement to obtain Y/C separation data.

However, just using a motion adaptive Y/C separation circuit,
It is impossible to avoid the drawbacks of dot crawling in parts where there is a sudden change in the vertical direction and vertical lines being emphasized in parts where there is a sudden change in the horizontal direction. Image adaptive Y/C
The separation circuit is intended to correct this defect, and selects an appropriate Y/C separation method according to the degree of change in the vertical and horizontal directions within the l-field plane, and selects a Y/C separation method that has fewer defects overall.
/C This is to obtain separated data.

FIG. 1 shows the configuration of an image adaptive Y/C separation circuit, in which a color television signal is supplied to a 2H comb filter 201 and a bandpass filter (hereinafter abbreviated as BPF) 202, and each Y/C separation signal is is created and supplied to the weighting circuit 203.

On the other hand, the input color television signal is also supplied to a circuit 204 that detects a change within one field, and this change detection signal is supplied to a weighting circuit 203, which outputs Y/C separated data of the 2H comb filter 201. The Y/C separated data of the BPF 202 is appropriately weighted and output as Y and C signals. A conversion coefficient is used to express the degree of this change. The conversion coefficient that represents the degree of conversion in the vertical direction is inversely proportional to the larger difference between the vertically adjacent scanning lines, and the conversion coefficient that represents the change in the horizontal direction is a number that is inversely proportional to the difference between the two adjacent pixels on the left and right on one scanning line. The number is proportional to the larger one of the differences between the vertical and horizontal conversion coefficients, and the Y/C separated data by the comb filter 201 with good resolution and the B P without dot crawl are determined by the larger of the vertical and horizontal conversion coefficients.
Y/C by adding weight to Y/C separated data by F202
It's about separation. As the weighting circuit 203, a known circuit used for weighting of inter-frame Y/C separation, intra-field Y/C separation, etc. can be used as is. In the above calculation, a coefficient (
Hereinafter, this will be referred to as the "change coefficient." ), the larger one of the change coefficients "K," representing changes in the vertical and horizontal directions and rKHJ is selected. The vertical change coefficient Kv is calculated by taking the difference in the Y component between the scanning line for which the change is to be calculated and the upper and lower scanning lines, and taking the larger absolute value as the amount of change, and calculating the change as shown in Figure 2. The number is inversely proportional to the amount. The horizontal portion occurs because the difference is considered significant when it exceeds an appropriate threshold level.

When there is a significant difference in the absolute value of the difference between the mean value of the level at the in-phase point of the color subcarrier from the center value on the scanning line and the center level, the lateral change coefficient KM is a number proportional to this difference. do. The part where the conversion amount is small is caused by assuming that there is a significant difference when the above difference exceeds an appropriate threshold level.

Next, the larger one of K and □ is used as a comprehensive conversion coefficient, and as described above, the Y/C separated data by the comb filter and the Y/C separated data by the BPF are weighted.

By determining Kv as described above, Kv becomes small in a portion with a large amount of conversion where dot crawl is likely to occur, so that the weight of the BPF 202 becomes high and dot crawl is less likely to occur. At the point where KV=0, the data separated by the B P F 202 is used, and at the point Kv=1, the data separated by the comb filter 201 is used.

By determining KH as described above, the following effects can be obtained. That is, when the signal on the scanning line does not change, or even if it changes, it changes in the same direction at the same rate or at a small rate, the value of KI4 will be 0 or a small value. In comparison, in the comb filter 201, K is always 1 and the vertical line is emphasized, but if image adaptive Y/C separation is performed, the comb filter 201 becomes The weight is reduced and the drawbacks of vertical line emphasis are alleviated. In other words, the hatched portion will be salvaged. In a portion where changes occur discontinuously, K becomes large, so the weight of the comb filter 201 becomes large and works to prevent resolution from dropping.

The problem in which motion is detected even though there is no motion between adjacent fields due to motion detection based on inter-frame differences between adjacent fields is dealt with as follows. First, motion determination is performed using the above-mentioned inter-frame difference for three adjacent fields. Determine the range in which any of these will be a movement. next,
For each RGB signal converted by the decoder, within the above range of motion, two adjacent
The difference between the scanning line of the book and the corresponding scanning line of the field delayed by one field is calculated, and if there is a significant difference and the sign is the same, it is determined that there is movement. In other words, the fact that there is a significant difference when taking the difference between two adjacent scanning lines in the current field and the corresponding scanning lines in the previous field means that there is no image in the current field and the previous field, or there is no image in the previous field. This means that the same sign means that there is a degree of redundancy in the vertical direction. Even if there is a significant difference, if they have different signs, they are treated as different images. When interpolating scanning lines, it is sufficient to select an appropriate interpolation method depending on the presence or absence of movement, but in this embodiment, bilinear interpolation is performed between adjacent scanning lines within the current field for the portion determined to be movement. Interpolation calculations are performed using the method, and for parts that are determined to be not moving, in the case of non-increment signals, the data of the previous field is interpolated, and in the case of incremental signals, interpolation calculations are performed between the current field and the adjacent field using the bilinear method. It is carried out.

By taking measures to improve image quality as described above, even within a single screen, it is possible to always perform the best Y/C separation and interpolation depending on the image movement and changes in the vertical and horizontal directions. Image quality can be improved to the utmost by performing appropriate scanning line interpolation that accurately detects motion.

(Example) Next, the present invention will be explained in detail using the drawings.

Figure 4 is a system diagram of the digital decoder according to the present invention.The human signal is an NTSC VH3 signal (a signal consisting of video, burst, and synchronization signals) used in broadcasting stations, but it is a baseband RGB signal. It's okay. It is necessary to correct a signal with a time axis deviation, such as an output signal of a VTR, by using a time base corrector, for example. Further, although the circuit shown in FIG. 4 does not include a circuit for suppressing noise (for example, a digital noise reducer), this may be inserted.

The VBS signal input from the input terminal l is converted into a digital signal by the A/D converter 2, and is sent to the phrase memory 3.
supplied to When a freeze command is issued, the current frame is frozen in the memory 3, and at the same time, the previous frame is frozen in the one frame delay 5. The signal read out from the memory 3 enters the image adaptive Y/C separation circuit 4, and the signal is inputted into the image adaptive Y/C separation circuit 4, and is converted into a Y/C separation circuit adapted to the vertical and horizontal changes in the image.
C separation is performed and Yc and Cc+a are output. Details of this method will be described later. The frame read from the freeze memory 3 and the frame read from the one-frame delay 5 are applied to an interframe Y/C separation circuit 6 to perform interframe Y/C separation and output Yp and CF. A known circuit can be used as this circuit. Image adaptive Y/C
Is the output signal of separation circuit 4 a filter? (BPF and L
PF) and outputs C5 and YL. Image adaptive Y/C
The output of the separation circuit 4 and the output of the interframe Y/C separation circuit 6 are determined by increasing the weight of the Y/C separated data by the image adaptive Y/C separation circuit when the movement is large, and increasing the weight of the Y/C separation data by the image adaptive Y/C separation circuit when the movement is small. Weighting is performed to increase the weight of Y/C separated data by the Y/C separating circuit to obtain comprehensive Y/C separated data. The weighting method is
First, the output of the filter 7 and the output of the 1-frame delay 5 are added to the motion detection circuit 8, the difference is taken to detect motion, and the output is added to the J generator 9 to generate J and (1-J). .

J is a coefficient that decreases as the movement increases.

Next, CL, YL, CF, YF and the above J and (1-
J) to the multipliers 10, 11, 12, and 13 respectively,
The outputs are added to adders 14 and 15, and comprehensive Y/C separated data Y and C are obtained using the following equation.

Y= (I J )Ct +J CFC= (I
J) CL +J CF Next, C is added to the IQ demodulator 16 to obtain I and Q, and I and Q are added together with Y to the IQ matrix circuit 17 to output RGB signals. The motion adaptive scanning line pre-compensation circuit 19 includes a motion detection circuit 8 to a delay circuit 1.
1 by the signal passed through 8 and the field delay circuit 20.
Field-delayed RGB signals are sent, scanning line interpolation adapted to the motion is performed, and the required R'G' is output from the terminal 21.
Outputs B' signal. The operation of this circuit will be described later.

Next, an outline of the image adaptive Y/C separation circuit 4 will be explained.

FIG. 5 is a system diagram of the image adaptive Y/C separation circuit, and FIGS. 6 and 7 are drawings showing the relationship between scanning lines and sampling points involved in determining KV and KH. The image input signal (Y+C) is input to I by line delays 31 and 32.
Delayed by H, the human output is delayed by each adjacent scanning line (n
+1), n (n-1). Adjacent scan lines are
Since the phase of the C signal is inverted, adder 33 adds (n+1) line and (n-1) line, and amplifier 3
If we take 1/4 of that by 5, the output is 1/2 (
Y 1C), and the n line is connected to 17 by the amplifier 34.
If we take 2, it becomes 1/2 (Y-C), so adder 3
If the difference is calculated by 6, the Y component becomes 0, and only C due to the 2H comb filter remains. This output is passed through the BPF 38 to remove noise components, and is added to the modulator 40 with a change coefficient K.
It is modulated by and added to the adder 44. On the other hand, if the n line is passed through the BPF 37 corresponding to the BPF 202 in FIG. 1 and the Y component is removed, a C signal by the BPF can be obtained. By adding this to the modulator 39, modulating it by (1-K), and adding it to the adder 44, a total C signal can be obtained. To destroy the total Y signal, the adder 45 can add n lines and C obtained by the above calculation.

To find the longitudinal change coefficient, use the same method as above, delay by IH by line delays 47 and 48, and adders 49 and 50 to calculate (n
By drawing the +1) line and the (TI-1) line, the C component cancels out and becomes 0, leaving only the Y component. These outputs are supplied to absolute value circuits 51 and 52 to take their absolute values, and a comparator 53 compares their magnitudes. A selector 54 takes the larger one and applies it to a K7 generator 55. Kv
Generator 55 produces an output having the characteristics shown in FIG.

Next, a method for determining the lateral change coefficient Kn will be explained with reference to FIG. 7 showing sampling points on the color carrier wave.

To find the horizontal change coefficient Kl, sample the nth scanning line at a frequency 4f, which is four times the color subcarrier frequency fic, and examine the data at the fourth position on the left and right from the median value. Sampling at 4f and c means sampling every 90° of the color subcarrier, and comparing the data at the fourth position on the left and right from the center value means that the same phase of the color subcarrier is sampled. This corresponds to checking the data at the position.The method is to apply 1/4f, c (approximately 70nS) to the signal delayed by IH from the input signal.
) Delay circuit 5 is a cascade connection of four delay circuit elements.
6 and 57 are connected, the sum of the sampling points (m+4) and (m-4) at both ends is taken by the adder 58, and 1/2 is taken by the amplifier 59, then it becomes (Y-C), and the center Since the value is also (Y-C), if the difference is taken by the adder 60, the difference in the Y component and the difference in the C component remain. By supplying this output to the absolute value circuit 61 and taking the absolute value, the amount of change in the horizontal direction can be determined. K) In the I generator 62,
An output having the characteristics shown in FIG. 3 is generated. If you examine data in this way, even if there is a significant difference between the median value and the left and right sampling points, horizontal changes can be detected in areas where the signals on the scan line change smoothly in the same direction. Since the amount is below the threshold, it is assumed that there is no change.

KV and Kn are compared in size by a comparator 43,
The larger one (say K) is selected by selector 42, and K and (K-1) are generated by K generator 41 and applied to modulators 39 and 40 as previously described.

Although the above circuit has been described for one embodiment, it goes without saying that other known circuits may be used.

Such an image adaptive Y/C separation circuit can adapt to changes in the vertical and horizontal directions of an image in one field, and take advantage of the advantages of Y/C separation using a comb filter and a BPF. Since Y/C separation can be performed, it is possible to reduce dot crawling or overlap of vertical fine lines in a portion where there is a sudden change in the vertical direction, and to greatly improve image quality. In addition to the present invention, Y due to image movement
If /C separation is used in combination, image quality can be further improved.

Next, the motion adaptive interpolation circuit 19 will be explained.

In FIG. 8, as described above, motion detection is performed by taking the difference between corresponding pixels between two adjacent frames. This diagram illustrates a motion image using a conventional method in which Y/C separation and scanning line interpolation are performed between frames in areas with little movement, and Y/C separation and scanning line interpolation within each field in areas with a lot of movement. It is. When detecting motion by taking the difference between the image P of the first field and the image P1-2 of the (I-2)th field one frame before, a portion A is detected as motion and a portion B is not detected. Sometimes. For the portion of Δ that has movement, in order to avoid the influence of movement, a new scanning line (indicated by a black circle) is interpolated using the bilinear method, that is, the average value of the adjacent scanning line (indicated by a white circle) is taken within the field. B without
In order to obtain accurate data, a new scan line (indicated by .DELTA.) in the I field is interpolated by taking the average value of the I field and the (I-1) field. At this time, in the B part, an image of the (1-1) field, which does not exist in the I field, is generated and is observed as an afterimage. In order to overcome this drawback, a method is known in which the inter-frame difference is calculated for each of two adjacent fields, and when there is a significant difference in either field, it is determined as motion. Presented at the national convention. However, when doing this, it is determined that there is movement even in the portion C that does not move between the I field and the (I-1) field that are actually interpolated, and the data of the previous field remains unchanged or between the fields. While it would be preferable to interpolate pi-linear data within a field, interpolating pi-linear data within a field introduces new drawbacks. As described above, the conventional motion detection and determination method in motion-adaptive scanning line interpolation is not optimal, and has the drawback that a high-quality image cannot be obtained from interpolated scanning lines.

In order to eliminate these drawbacks, the present invention performs interpolation calculations by separating and determining parts that actually have movement and parts that do not. The objective is to eliminate portions such as C that are determined to have no motion even though there is motion, and portions such as C that are determined to have motion even though there is no motion between adjacent fields.

To this end, in the present invention, first, motion determination is performed on three adjacent fields based on the above-mentioned interframe difference, and a range in which any one of these fields constitutes motion is determined. Next, for each RGB signal converted by the decoder, within the range of movement described above, the difference between two adjacent scanning lines of the current field and the corresponding scanning line of the field delayed by one field is calculated. , when the magnitude of these differences is significant and the signs of these differences are the same, it is determined that there is movement. In other words, if there is a significant difference between two adjacent scanning lines in the current field and the corresponding scanning lines in the previous field, it means that there is no image in the current field and the previous field, or an image that was not present in the previous field is missing. Therefore, having the same sign means that there is redundancy in the vertical direction. Even if there is a significant difference, if they have different signs, they will be treated as different images.

When performing scanning line interpolation, it is sufficient to select an appropriate interpolation method depending on the presence or absence of movement, but in this embodiment, bilinear interpolation is performed between adjacent scanning lines within the current field for the portion determined to be movement. Interpolation calculation is performed using the method, and for the part determined to be not motion, the data of the previous field is interpolated in the case of a non-increment signal, and the data of the previous field is interpolated in the case of an increment signal. Interpolation calculations are performed using the bilinear method.

With the above configuration, moving parts and non-moving parts are strictly separated on one screen, and the most appropriate interpolation calculation can be performed for each.Therefore, no afterimages occur, and the areas where inter-field interpolation should be performed are RG with the best image quality and is the most faithful to the original picture, without being in the field.
B signal can be obtained.

FIG. 9 shows a detailed configuration of the motion adaptive scanning line interpolation circuit 19. The circuit 19 includes an IQ matrix circuit 17.
and the R0 signal delayed by one field by field delay 20 are added. On the other hand, subtracters 71 and 72 separate the R signal and R signal.

The difference between the signal RH and the R9 signal delayed by the IH delay 70 and the R signal is taken, and the absolute value thereof is taken by the circuits 73 and 74, and the threshold level S generated by the threshold generator 77 is obtained. and in comparators 7.5 and 76,
An output that goes high when the difference is greater than S is applied to the AND gate 80. On the other hand, the signals representing the signs from the subtractors 71 and 72 are applied to the exclusive OR circuit 78 and then to the NOT circuit 79, so that when the signals are of the same sign, the signal becomes high level.

When this is added to the AND gate 80, the output of the AND gate becomes high level only when the output signals of the subtracters 71 and 72 are higher than the threshold level S and have the same sign. The above circuit operation is the same for the circuits that process green signal G and blue signal B, and the outputs from AND gates 80G and 80B of both circuits are applied to OR gate 87.

The current signal for motion detection is shown in Figure 10 in the ■ field and (I
-2) field, and the output signal of the field delay 15 is the (I-1) field and (I-3) field.
) field, and the output signal of the field delay 16 is the difference between the (I-2) field and (I-4) field, so ORing these fields expands the range that is determined to be movement. It has to be outside. With this measure,
In FIG. 10, it is possible to correct a portion such as B where it is determined that there is no movement even though there is movement between the I field and the (I-1) field. However, as it is, the ■ field that performs interpolation calculations and (I-
1>This results in a portion such as C where it is determined that there is movement even though there is no movement between the fields. In order to correct this, the present invention calculates the difference between two adjacent scanning lines of the current field and the corresponding scanning line of the previous field, and determines whether the differences have a significant difference and have the same sign. If there is a significant difference, it is determined that there is a movement, and if there is no significant difference, or even if there is a difference, the sign is different, then it is determined that there is no movement. That is, when the presence or absence of images in the current field is reversed with respect to the previous field and there is redundancy in the vertical direction, it is determined that there is movement. Figure 11 shows this explanation.
In Figure 11, which shows the case where a figure moves on the screen, when the figure indicated by the dotted line in the (■-1) field moves to the position indicated by the solid line in one field,
If the range of -X and YY is shown in an enlarged view, it becomes as shown in FIG. 11B. ■ Two adjacent scanning lines of the field (L
-1) to Cl-1) corresponding scan line L' of field
If you subtract , it is determined that there is movement because there is a significant difference in the area where there is no figure in one field as shown in Figure 11C, and the changes are in the same direction. Since there is no significant difference in the part where there is no figure, it is determined that there is no movement. If you do this, the 10th
The part C in the figure is determined to have no motion, and it is possible to eliminate the above-mentioned problem in which motion is determined even though there is no significant difference between adjacent fields. In a part where there is a sudden change in the vertical direction, that is, a part without redundancy, even if there is a significant difference between the adjacent scanning line and the scanning line of the previous field, the signs are different, so it is determined that there is no movement. .

The AND gate 88 includes the output of the OR gate 87 and the output of the OR gate 86.
Since the outputs of the two adjacent scanning lines of the current field and the previous field correspond to each other within the range of possible motion determined by the delay circuit 18, the output of When the difference between the scan line and the scan line is significant and has the same sign, that is, when there is movement, the level is high. On the other hand, the data of two adjacent scanning lines are added by an adder 81, averaged by a 1/2 amplifier 82, and applied to a switch 83 together with delayed field data Ro. Switch 83 is AND gate 88
The average value of two adjacent scanning lines (indicated by a black circle) is taken for a moving part (the part indicated by D in Figure 10), and the R is controlled by the output for a stationary part.
Output 1+ (indicated by the X mark) as is. The same goes for the G and B signals (and gate 80 in the drawing).
Compatible Antogame) 80G.

80B, a switch 83G corresponding to switch 83,
The average value of adjacent scanning lines and either Gfl or So are selected by the output of the AND gate 88 which shows only 83B.

As explained above, according to the present invention, it is not determined that there is no movement between the scan lines where interpolation calculations are performed, nor is it determined that there is movement when there is no movement, and the actual movement of the image is not detected. If you understand it correctly, you can do calculations. For this reason, it is possible to obtain extremely high quality image signals.
The effect of improving image quality is truly significant.

FIG. 12 is a system diagram of an apparatus that uses a digital decoder according to the present invention and creates a hard copy from a television signal using instant film that can be used as an opaque card. As mentioned earlier, still images require extremely high image quality, so in addition to using the digital decoder according to the present invention,
It has well-known means of improving image quality, and has made various other improvements as well. One is to digitize the circuit, thereby eliminating the use of analog circuits and the use of A/D converters or digital converters.
This prevents the image quality from deteriorating due to the /A converter. Another reason for digitization is that each part adopted a method that takes advantage of the advantages of digital circuits, such as the use of a table look-up method. Furthermore, each circuit is controlled by a combi coater, achieving automation and simplification of operation.

Next, the outline will be explained. The human signal is an NTSC-type VBS signal (a signal consisting of a video burst and a synchronization signal) used in a broadcasting station, but the baseband may be an RGB signal. It is necessary to correct a signal with a time axis deviation, such as an output signal of a VTR, by using a time base corrector, for example. Further, although the circuit shown in FIG. 12 does not include a circuit for controlling noise (for example, a digital noise reducer), this may be inserted. A VBS signal input from an input terminal 101 is applied to a digital decoder 103 according to the present invention, which outputs a frozen RGB signal. The human input signal can be observed on the monitor 102 and a screen can be selected.

The output of digital decoder 103 is applied to image enhancer 104. It is also possible to convert the RGB signal from the terminal 133 into a digital signal by the A/D converter 134, and to switch and apply the RGB signal stored in the frame memory 135 to the image generator. The image enhancer performs edge enhancement in order to compensate for the drop in aperture response when producing a color film and tighten the image, and a known circuit can be used. At this time, the degree of emphasis in the vertical direction (V, DETAIL) and the degree of emphasis in the horizontal direction ()1. DETAI can be adjusted.

The signal from the image enhancer 104 is then applied to a color corrector 105, and an operation box 129 performs gamma correction, contrast correction, and brightness correction for each color of the RGB signal and for the entire image. The correction pattern is RGB
It can be observed on the monitor 127, and at this time,
A cursor created by the cursor generator 125 can be used and can be projected at any position using the joystick on the operation box. Gamma correction is used to correct color distortion, etc. Gamma characteristics are calculated using a microprocessor based on data from the variable resistance force 1 on the operation box, stored in RAM, and table lookup method (table ``table'') is performed. This method is used to refer to the data, process it according to the purpose, and output it. Contrast correction involves inputting the input signal and data from the operation box into a multiplier to calculate and output characteristics. For brightness correction, input signals and data from the operation box are input to an adder to calculate and output characteristics.

The signal from color corrector 105 is then written to frame memory 106. Writing is at field frequency 6
It is an increment signal of 0H2, but the readout is 30Hz.
This is done as a non-increment signal.

The signal read from the frame memory 106 is subjected to gamma correction of the CRT 114 and film 120 that are part of the printer by a gamma correction circuit 107. The gamma of a CRT is usually 2.2, but to correct this, the test signal generated by the test signal generator 132 is
In addition to the CRT via the converter 110, a photomultiplier tube (P
The photoelectrically converted signal is read by the microprocessor (MT) 121, the deviation from the test signal and the data of T=1 is calculated by the microprocessor, and this correction data is read into the RAM and corrected by a table lookup method. The film gamma is corrected by writing the /-car data of the two types of films used in the RAM and using a table lookup method.

The gamma-corrected signal is then sent to the shading correction circuit 1.
Shading correction is performed using 08.

The shading correction is disclosed in Japanese Unexamined Patent Application Publication No. 1983, filed by the applicant.
This method is similar to the method of No. 51676 "Shading correction device". That is, the CRT screen is divided into small areas, and for each area, the average value of the output of the PMT 121 when a white light signal is applied to the CRT is stored in the ROM and corrected using a table lookup method. be.

The signal that has undergone shading correction is sent to the scanning line interpolation circuit 1.
09 doubles the scan line. Scanning line interpolation increases the number of scanning lines to obtain a fine-grained /%-tocoby, but in normal hard copy, doubling 525 lines to 1050 lines will provide a sufficient resolution. can get. R, G, and B are sequentially selected by a selector that performs interpolation, and the interpolation calculations are sequentially output. Regarding the interpolation method, please refer to Japanese Patent Application Laid-open No. 58-27 filed by the applicant.
As described in detail in No. 146 "Scanning line interpolation device in video image prepress system", the nearest neighbor method takes the same value as the adjacent scanning line, the bilinear method takes the average value with the adjacent scanning line, and the nearest neighbor method takes the average value of the adjacent scanning line. There are methods such as cubic convolution that take into account the influence of 16 pixels surrounding the interpolation point, but in this embodiment, cubic convolution was adopted to obtain good image quality. If you want to double the number of scanning lines, you just need to set the interpolation point at the midpoint on the pixel column, and the influence of pixels in other columns will be eliminated, so the calculation formula is simplified, and the pixel data of the upper and lower lines are ,
- If P2, P3, and P4, the pixel data of the interpolation point to be obtained is expressed by the following equation, so it is sufficient to convert this into a circuit.

P=(1,625(P2+P3) -0,125(P
, +P4) The signal that has completed scan line interpolation is fh = 15
．． 75kHz.

The RGB frames are sequentially read out at fv=15 Hz and the number of scanning lines=1050, converted into analog signals by the D/A converter 110, and supplied to the printer circuit.

Reading at fv = 15Hz means that the frequency characteristic of commercially available products such as amplifier ICs that can usually be done manually is f h = 15.75.
This is because it becomes worse at frequencies higher than Hz, and if this problem can be solved, fh = 30.15Kflz, fv = 3Qt
(z, number of scanning lines = 1050).

The video signal amplified by the amplifier 111 is displayed on the CRT 114. This image is reflected by the mirror 116,
The light passes through an objective lens 117 and an RGB filter 118, is reflected by a rotating mirror 119, and is imaged onto a film 120, thereby exposing the necessary number of frames for exposure. RGB
The signals are applied frame sequentially and RGB filter 118
Switch to perform overlapping exposure.

When performing gamma correction and shading correction, the angle of the rotating mirror 119 is changed so that the light beam from the CRT 114 enters the PMT 121.

Switching of RGB signals, switching of ROB filter 118,
Control of the number of exposure frames, control of the rotating mirror 119, control of the shutter, etc. are performed by the printer control 123. System control 128 is accessed through printer interface 122 and controlled by operation box 129.

(Effects of the Invention) According to the present invention, the most appropriate Y/C separation can be performed corresponding to moving parts and non-moving parts, and parts that change vertically and horizontally in one screen, and the movement can be detected correctly. Appropriate interpolation can be performed using the image processing method, and even small defects that are not noticeable in a moving image but noticeable in a still image can be completely corrected, and extremely high-quality images can be obtained.

The present invention is applicable not only to a hard copy production device that captures an image on instant film as in the above embodiment, but also to a R
All can be used in GB signal conversion circuits. Therefore, it can be used in a plate-making device for printing separation plates, etc., and can also be used in high-quality facsimiles. It is also possible to record the RGB signals as they are on a digital videotape or video disc and use them as needed. It goes without saying that the present invention can be used not only for NTSC television signals but also for PAL and SECAM television signals.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing the basic configuration of the image adaptive Y/C separation circuit used in the digital decoder of the present invention. FIGS. 2 and 3 are characteristic diagrams of the image adaptive Y/C separation circuit. Figure 5 is a block diagram showing the overall configuration of the decoder of the present invention, Figure 5 is a block diagram showing the configuration of an example of an image adaptive Y/C separation circuit, and Figures 6 and 7 similarly explain its operation. 8 is an explanatory diagram of a conventional motion detection method, FIG. 9 is a block diagram showing the configuration of an example of a motion adaptive scanning line interpolation circuit according to the present invention, and FIGS. 10 and 11. FIG. 12 is a block diagram showing the configuration of an example of a hard copy production apparatus having a decoder of the present invention. 1... Input terminal 2... A/D converter 3...
- Freeze memory 4... Image adaptive Y/C separation circuit 5... 1 frame delay 6... Inter-frame Y/C separation circuit 7... BPF Death... Motion detection circuit 9...
・K generation circuit 10, 11.12.13... Multiplier 14, 15... Adder 16... IQ demodulator 1
7...IQ matrix circuit 18...Delay circuit 19...Motion adaptive scanning line interpolation circuit 20...Field delay 21...Output terminal Fig. 1 Fig. 2 Fig. 3 Fig. 5 Sun 71 Procedural Amendment Written June 3, 1986 Michibe Uga, Commissioner of the Patent Office1, Indication of the case 1985 Patent Application No. 1155732, Name of the invention Digital Decoder3, Person making the amendment Case and Relationship between the patent applicant Ikegami Tsushinki Co., Ltd. 4 and the representative's specification, page 2, line 18, ``two'' is corrected to ``two''. 2. Correct "conversion amount" in lines 9 and 18 of page 8 to "amount of change," and correct rK=OJ to rKH=OJ in line 10 of the same page. 8. On page 16, lines 18-19, ``If you subtract 111 from In'', correct it to ``If you add In, (n-1) line, and (n+1) line.'' 4 Substitute r(K-1)J on page 17, line 16 of the same page with "(1-K)
” is corrected. 5. Correct "significant" to "significant" on page 21, line 8. 6. Correct "15" in line 16 of page 28 to "84" and correct "16" in line 17 of the same page to "86." F. Correct "significant difference" to "significant difference" in lines 4, 7, 10, and 14 of page 25 of the same page. Correct. (Correction area 9

Claims (1)

[Claims] 1. In a decoder that converts a color television signal into red, green, and blue signals, an image adaptive luminance signal/chrominance signal separation circuit that performs luminance signal/chrominance signal separation according to the amount of change in an image within a field screen; Motion-adaptive luminance signal and color signal separation that separates luminance and color signals according to image movement
A digital decoder comprising a color signal separation circuit and a motion adaptive scanning line interpolation circuit that performs scanning line interpolation according to the movement of an image.