WO2001028253A1

WO2001028253A1 - Method and apparatus for convergence correction in a television set

Info

Publication number: WO2001028253A1
Application number: PCT/EP2000/009057
Authority: WO
Inventors: Jacques Chauvin; Albert Runtze
Original assignee: Thomson Licensing S.A.
Priority date: 1999-09-23
Filing date: 2000-09-16
Publication date: 2001-04-19
Also published as: DE19945623A1; AU7519400A

Abstract

The invention relates to a method for convergence correction in a television set, and to a television set, in particular a projection television set, having a monochromatic tube for each of the three primary colours, red, green and blue, whose images can be projected on a screen. Each tube has an associated correction device, which comprises a convergence circuit. The fundamental idea of the method according to the invention consists in the manufacturer already taking account of differences which result from convergence setting by means of an image acquisition system on the part of the manufacturer and such setting by means of optical sensors in the set. In this way, mechanical tolerances and also optical imaging defects can be taken into account equally with little outlay. As a result, it is possible for the user to obtain improved convergence setting.

Description

Method and apparatus for convergence correction in a television set

The invention relates to a method for automatic convergence correction in a television set, and to a television set having a convergence correction device.

Convergence correction devices are used for correction of parameters in the raster deflection, for example for correction of north/south and east/west distortion, pincushion distortion, nonlinearity in the deflection, and other geometric errors in the horizontal or vertical direction. One particular field of application is convergence correction in a projection television set, in which the pictures from three monochromatic tubes are projected onto a screen. The term television set in this context means any equipment with an electronic raster picture display. The equipment may be fed from a television broadcast signal or else, as a pure monitor, from an RGB signal, a composite video signal, or separately with a luminance signal and a colour sub-carrier from any desired video signal source.

The deflection parameters are corrected using correction values which are stored in a convergence circuit. The stored correction values are converted in a digital/analogue converter into an analogue control signal, and are fed to a driver circuit which comprises a preamplifier and an output amplifier. This driver circuit emits a current, which corresponds to the correction value, to a correction coil. The details of such a convergence correction circuit are disclosed, for example, in German Patent Application DE 42 14 317. The convergence circuit itself is not the subject matter of the present invention.

The convergence correction values also depend, inter alia, on the earth's magnetic field prevailing at the installation location of the television set. This means that the convergence correction must be recorrected after the set has been transported - for example from the manufacturer to the customer - m order to achieve an optimum setting. An entirely corresponding situation also occurs, of course, when the mechanical construction of the set changes as a result of parts being replaced for repair purposes, or as a result of external influences.

In known sets, a raster is displayed on the screen, using the three primary colours red, green and blue, m a special operating mode for convergence correction. The remote control, for example, can now be used to move the three primary colours of the raster such that they coincide, by correcting the convergence values successively for each raster crossing point. The convergence values are stored automatically m a memory provided for this purpose m the set. If, for example, there are eleven horizontal and fifteen vertical raster lines, a horizontal and a vertical convergence value can thus be stored for each of the three primary colours for 165 crossing points, that is to say a total of 990 values. This is time-consuming and tedious, not least because the optimum setting is often not found m a single attempt. Furthermore, this trimming can be carried out only by trained personnel.

Projection television sets are known m which the convergence correction is effected automatically, for example by pressing a button. A number of optical sensors are provided for this purpose m the known projection television set. The optical sensors are arranged outside the visible region of the screen and thus have no adverse effect on the picture impression on the screen. In order to illustrate this construction,

Figure 1 shows the view from the front of a known projection television set, with a screen 700. The visible region VA of the screen is provided with the reference symbol 800. The boundary of the screen 700 is represented by a broken line, m order to indicate an edge region OS which remains hidden within a casing C. However, pictures can nevertheless be displayed m the edge region if the television set is operated m an overscan mode. The edge of the imaging area, which is designated by the reference symbol OS in Figure 1, is depicted by a dotted line in the overscan mode. In the figure, eight photosensors I-VIII are arranged in the corners and in the centre of the screen edges outside the visible region 800, but inside the overscan region OS. These sensors therefore make it possible to measure an electronically generated test pattern in order to determine picture width and height, and also specific geometric defects, e.g. rotation, distortion, keystone distortion, pincushion distortion, etc., and to determine convergence corrections therefrom.

It is evident that for good convergence correction, firstly the positions of the optical sensors must be known exactly, and secondly the position of the overlaid raster on the screen can be determined exactly with regard to the sensors. In practice, it has now been shown that deviations which reduce the quality of the convergence correction can occur with regard to both of the aforementioned points.

Taking this as a departure point, the object of the invention is to specify a method for convergence correction which takes account of the aforementioned deviations and eliminates the disadvantageous consequences thereof. In this way, the convergence correction obtained is more accurate than that obtained by conventional methods.

This object is achieved by means of the method according to the invention according to Claim 1. A development of the invention also takes account of deviations which can occur during a changeover of the operating mode between an underscan mode and an overscan mode. This development is particularly advantageous if computer images are displayed by the projection television set.

In a modification of the method according to the invention, the convergence magnetic field defined in one direction is maintained while each of the others are determined. This makes it possible to move to the desired position of the marker more accurately. It has been shown to be advantageous if the initially determined convergence magnetic field is recorrected once again at the end, if the second convergence magnetic field is fixed. This allows increased convergence correction accuracy to be achieved.

The drawing illustrates a projection television set which is suitable for carrying out the method according to the invention. In the figures:

Figure 1 shows a known projection television set m a front view, Figure 2 shows a plan view of the screen of the television set from Figure 1 without convergence correction,

Figure 3 shows a convergence correction device schematically, Figure 4 shows the screen of the television set from Figure 2 with a convergence raster overlaid, Figure 5a shows a schematic illustration of a positional deviation of a sensor, Figure 5b shows an illustration of how raster lines continue non-lmearly m the non-visible screen region, Figure 5c shows an illustration of a parallax error, and Figure 6 shows a schematic representation of the successive approximation of a marker to a desired position. Figure 2 shows a plan view from the front of the screen 1 of a television set according to the invention, on which the pictures are projected from three monochromatic tubes 2, 3, 4. As is shown schematically m the figure, the tubes 2..4 are arranged geometrically differently. Imaging errors thus occur on the screen, which are different for the individual tubes. These picture errors are corrected by means of additional correction coils, which are mounted m front of the actual deflection coils, on the coil neck of the tubes. The correction for each individual tube 2..4 is carried out both m the horizontal direction and m the vertical direction, that is to say there are a total of six correction coils with the associated driver circuits m the television set, and these are each activated by one convergence circuit. A correction coil with the associated circuitry is referred to as a convergence channel, so that the television set has a total of six convergence channels. Figure 3 schematically shows the block diagram of a convergence channel which is designated overall by 5, of the kind known from the prior art. A convergence circuit 6, which is m the form of an integrated circuit, is connected by one output 7 to a driver circuit 8 which comprises a preamplifier 9 and an output amplifier 11. The output 12 of the output amplifier 11 is connected to a correction coil 13 which influences the electron beam m the associated tube. The correction coil 13 is connected m series with a measurement resistor 14. Each convergence circuit 6 together with the associated correction coil 13 and driver circuit 8 is referred to as a convergence channel.

The convergence correction is carried out using stored digital convergence correction values. The convergence correction values are stored m a memory 15, designated by M, m the convergence circuit 6, and are converted m a digital/analogue converter 16 into a corresponding analogue voltage. The voltage is amplified by an amplifier 17 that is integrated m the convergence circuit 6, and is emitted at the output 7 to the driver circuit 8, which produces the desired correction current m the correction coil 13.

The way m which the convergence correction values are processed m order to achieve the desired correction of the raster on the screen is not the subject matter of the present invention. Methods and devices relating to this are disclosed, for example, m German Patent Applications DE 197 35 681 and DE 197 04 775. Figure 4 shows the screen from Figure 2, on an enlarged scale. The visible region of the screen is surrounded by an edge 18, which is represented by a broad solid line. Eight optical sensors, which are designated by Roman numerals I to VIII, are arranged outside the edge 18. Furthermore, a convergence raster 19 is shown overlaid on the screen 1, and is formed by thirteen horizontal and fifteen vertical lines 21 and 22, respectively. The convergence raster 19 extends beyond the edge 18 in the overscan operating mode, so that the optical sensors I to VIII are located in the region of the raster 19. The sensors are connected to an evaluation circuit (not illustrated in the drawing) which emits a corresponding signal when light falls onto one of the sensors. This light is produced in particular by so- called markers, which can also be displayed in the non- visible region on the screen.

The way in which the position is determined is described in German Patent Application DE 197 00 204. According to this, the convergence magnetic field at which the position of the marker and the position of the sensor match is determined by moving two rectangular markers from a respective side towards the sensor that is used for measurement, match in this context meaning that the respective sensor responds. The precise correction value for the horizontal and vertical convergence is then calculated from the two measurements. For the sake of clarity, the present description refers to only one marker, which is indicated as a cross in the figures. For the sake of brevity, the term response of the sensor" in the text below also means any procedure which allows the position of a marker to be related to that of a sensor. For example, this term covers the situation where a marker is respectively approximated to a sensor from the left or right in order to use the measured electrical signals to tune the position at which the marker would be centred on the sensor in the horizontal direction. The same also applies correspondingly to determination of the position of the marker in the vertical direction.

The convergence correction for the entire screen can be calculated from the convergence correction values that have been determined in this way, but this is not the subject matter of the present invention. The quality of the convergence correction is limited, however, by a number of deviations during the determination of the position of the markers with regard to the sensors. These deviations are described below.

In each individual projection television set, mounting deviations of the optical sensors can occur, with the result that, by way of example, the connecting line between the sensors I, II and II is not a straight line, as assumed theoretically, but rather has a bend. The calculation of the convergence correction values, by contrast, is based on the assumption that the connecting line actually constitutes a straight line. Consequently, such a mounting deviation leads to erroneous calculation of the convergence correction. Figure 5a illustrates one example of such a positional deviation on a greatly enlarged scale, the desired position of the sensor III being represented by a broken line and the actual mounting position by a solid line. The incorrect position is illustrated by an arrow FI .

Figure 5b illustrates a different form of deviations based on the fact that the convergence which is set in the visible region of the screen by means of a camera does not necessarily lead to linear continuation of the overlaid convergence raster beyond the screen edge 18 into the overscan region. As can clearly be discerned in Figure 5b, the convergence raster lines can have a curved course in the overscan region. The setting of the convergence by means of the optical sensors, by contrast, is based on the assumption that the convergence raster extends linearly right into the overscan region, so that the convergence calibration set by means of a camera on the part of the manufacturer can be achieved. The deviation of the horizontal convergence raster lines 21 from the sensors shown is illustrated by the arrows F2 in Figure 5b.

Finally, Figure 5c illustrates a further deviation caused by the geometric arrangement of the various components of the projection television set. As can be seen from Figure 2, the beams of the individual tubes impinge at different angles on the sensors, which are not arranged in the screen plane SP but rather project above it with their light-sensitive surface 36. Figure 5c shows the sensor II, by way of example, on which the light beams of the red and blue tubes impinge from different angles. The light pencils are designated by the reference symbols R and B in Figure 5c. As is described in the introduction, according to the known method the desired position of the marker is determined in such a way that it is situated in the centre of the light of the light-sensitive region 36. This corresponds to position A in Figure 5c. In the screen plane S, on the other hand, this type of setting leads, however, to different positions A' and A' ' for the red and blue light beams. For optimum convergence setting, by contrast, it would be necessary for the different light beams to attain convergence in the screen plane SP, that is to say the convergence of the different colours must be obtained at point B. The spatial deviation between point B and point A' and A' ' is represented by arrows F3 in Figure 5c.

The method according to the invention functions as follows:

After the projection television set has been fully assembled by the manufacturer, a convergence raster extending right into the overscan region of the screen is overlaid for the individual primary colours. The convergence is set successively for the individual primary colours in the visible region by means of an electronic camera, with no consideration of how the convergence raster continues in the overscan regions, where the raster lines may, under certain circumstances, have a curved course.

The horizontal and vertical convergence values, at the location of a sensor, are designated by xO and yO, respectively, and stored in one area of the memory M. From this optimum convergence setting for the observer, markers are then overlaid in the primary colours and shifted until the sensors I-VIII each respond. In concrete terms, a marker is thereby shifted counter to the arrow directions FI, F2 and F3 in Figures 5a-5c, or convergence values (xl, yl) are determined which effect such shifting. The correction values (xl, yl ) are stored in another area of the memory M. These steps are performed by the manufacturer. By means of a convergence correction that is initiated later by the user, the markers are shifted from their respective initial position in such a way that the sensors respond one after the other, the corresponding convergence values (x2, y2 ) being determined. Such a convergence correction may become necessary for example in order to adapt the convergence setting to the magnetic field prevailing at the installation location. In order, from these values, once again to arrive at the optimum convergence settings for the observer, the difference - stored during manufacture - between the camera setting (xO, yO) and the setting by means of sensors (xl, yl) is taken into account by subtraction:

xO = x2 - xl (1) yO = y2 - yl (2)

In this way, it is possible for the user, despite the above-described deviations which inevitably result in the course of convergence setting using sensors, to achieve a convergence adjustment which is very close to that originally performed by means of a camera on the part of the manufacturer. The method described thus far can also be used for convergence adjustment when there are only computer images present, which cannot be displayed in the overscan mode because otherwise the information at the image edge would be lost. On the other hand, the automatic convergence adjustment on the part of the user can only be carried out in the overscan mode, because it is only in this mode that the overlaid markers can be imaged on the sensors I-VIII. It has been shown that different convergence settings are necessary for the two different operating modes, because, by way of example, the optimum convergence values in the underscan mode lead to curved raster lines of the convergence grid outside the visible region of the screen if the screen is operated in the overscan mode. These facts correspond to the illustration in Figure 5b. The method according to the invention makes it possible to determine the differences for the convergence settings entirely in accordance with the deviation errors and to store them as correction values (xlX yl' ) , in that first of all, in the underscan mode, the convergence is set by means of an optical image acquisition system and the corresponding convergence values (xO, yO) are stored. Subsequently, a changeover is made to the overscan mode and the convergence values xl, yl determined by means of the optical sensors I-VIII are stored. If the operating mode then changes, these differences, as described above, are taken into account by subtraction:

xO = x2 - xl' (1' ) yO = y2 - yl' (2' )

In this way, it is ensured that it is possible to obtain an optimum convergence setting by means of the optical sensors even in the underscan mode.

In order to obtain the best possible results, it is important for the markers to be moved as precisely as possible to the desired position. In this case, the difficulty arises that it is impossible to shift a marker purely horizontally in translation on the screen by passing current through a horizontal convergence coil. A corresponding situation applies to the vertical convergence coils and vertical translational shifting of a marker. This situation is illustrated in an axis system in Figure 6, in which the x and y axes indicate the horizontal and vertical directions, respectively. The direction in which a marker is shifted when current flows through a horizontal convergence coil is shown by the dotted line H in Figure 6. The direction in which a marker is shifted when current flows through its associated vertical convergence coil is shown by the dotted line V in Figure 6. It can clearly be seen that, in both cases, the actual movement of the marker on the screen 1 of the television set comprises a superimposed vertical and horizontal movement. The reasons for this may be, for example, imaging effects or non-homogeneous deflection fields. This phenomenon influences the accuracy with which the desired position of a marker can be determined, as is shown in Figure 6. At the start of the measurement procedure described in the introduction, a marker is located at the position designated by 29a at some distance from its desired position, which is defined by the intersections of the X and Y axes. The magnetic field of the horizontal conversion coils shifts the marker along the line 32 until the marker reaches the Y axis and assumes the position designated by 29b. The magnetic field of the horizontal convergence coils is then switched off, and the marker returns to its original position 29a. After this, a magnetic field is applied by means of the vertical convergence coils and shifts the marker along the line 33 to the position designated by 29c. Following this, the magnetic field from the vertical convergence coils is switched off once again. The convergence values corresponding to the convergence magnetic fields determined in this way are stored, for the relevant sensor and the relevant marker colour, in a memory. The convergence correction method according to the invention is based on the stored convergence values being suitable for shifting the marker from its initial actual position to its desired position at the intersection of the X and Y axes. On the assumption that the horizontal and vertical convergence coils shift the marker parallel to the horizontal axis and the vertical Y axis respectively, this aim would also be achieved. However, since the marker is actually shifted along the H and V axes, respectively, by the convergence coils, the final position of the marker, designated by 29d, deviates from the desired position, since the shift takes place along the lines 33 and 32' . The said deviation can have a disadvantageous effect on the accuracy of the convergence correction.

A modification of the method according to the invention allows the marker to approach more closely the desired position that is wanted. According to the modified method, the marker is shifted from the initial position 29a, by means of a magnetic field from the horizontal convergence coils, along the line 32 to the position 29b. This magnetic field is now maintained and, in addition to it, a further magnetic field is applied by means of the vertical convergence coils, which shifts the marker from the position 29b along the line 34 to the position 29e. As can clearly be seen in Figure 9, the position 29e is already closer to the desired position than the position 29d which is achieved using the previously described procedure. The marker can be moved even closer to the desired position by recorrecting the horizontal convergence field while maintaining the vertical convergence field. In this case, the marker is moved along the line 36 from the position 29e to the position 29f. It is obvious to a person skilled in the art that any desired approximation of the marker to the desired position is possible by iteration of these steps. However, as a rule, the accuracy achieved at the position 29f, which is reached in three steps, is sufficient for all practical requirements.

Claims

Patent Claims

1. Method for convergence correction m a projection television set having monochromatic light sources for a respective primary colour which are imaged on a screen having a visible and a non-visible region (VA, OS), a number of sensors (I-VIII) being arranged m the non- visible region (OS) , m which case, according to the method, a) a raster (19) having horizontal and vertical lines is overlaid m a primary colour; b) the convergence is set using the raster (19) by means of an image acquisition system and the associated convergence correction values are stored as a first convergence data record (xO, yO) ; c) the position of the raster lines m the non-visible region of the screen is determined with regard to the flxed-position sensors and the associated convergence correction values are stored as a second convergence data record (xl, yl); d) steps a) - d) are repeated for each primary colour.

2. Method according to Claim 1, characterized in that the overlaid raster is displayed m the underscan mode m order to set the convergence by means of the image acquisition system, and m that the overlaid raster is subsequently operated m the overscan mode m order to determine the second convergence data record (xl', yl') .

3. Method according to Claim 1 or 2, characterized m that, m a manner actuated by the user, the convergence correction values of the projection television set are determined by means of the sensors and are stored as a third data record (x2, y2 ) , and m that the difference between the second and third convergence data records is formed.

4. Method according to Claim 1, characterized in that a) the marker is initially shifted from an initial position in a first direction until the associated sensor responds; b) the magnetic field required for the shift in the first direction is maintained, and; c) the marker is shifted in a second direction until the sensor responds once again.

5. Method according to Claim 4, characterized in that the magnetic field required for the shift in the second direction is maintained, and in that the magnetic field applied for the shift in the first direction is recorrected until the sensor responds once again.