GB2299723A - Suppressing video signals - Google Patents

Abstract

A system is shown for generating chroma-key related suppression signals. Luminance components and/or chrominance components of a video signal are suppressed. A generator 51 receives the chrominance information and generates a suppression signal from said information. A second generator 52, 53 receives luminance information and generates a suppression signal from said luminance information. Combiners 56, 58 combine the first and second primary suppression signals to generate composite chrominance and luminance suppression signals. A suppressor subtracts these signals from their respective video component.

Description

SUPPRESSING VIDEO SIGNALS The present invention relates to modifying suppression signals arranged to suppress luminance and/or chrominance components of a video signal.

Procedures for combining two or more video signals, representing different images, have been used for some time and now represent well established features of real time mixing desks and post-production facilities.

Using these procedures it is possible to produce composite images in which components are taken from a plurality of sources and combined in such a way as to provide a more artistic output image or, alternatively, to create an illusion to the effect that a video sequence has been shot with a particular foreground placed in front of a particular background. In this way, it is possible to place a presenter in front of any particular background without that presenter needing to be transported to the particular geographical location. This approach is also used in news broadcasts, where a seated presenter may report news items with video action also being displayed.

A mixer for combining digital video signals, allowing four real time signals to be combined, is disclosed in United States Patent No. 5,305,107.

This disclosure shows video signals that may be combined with drop shadows, borders and fills to produce a first stage composite video signal and a composite key signal. The key signal is used to switch between video sources and it is necessary to ensure that key transitions occur smoothly within each frame, so as to blend foreground object images into the background image, thereby avoiding the creation of undesirable jagged edges and artifacts.

Key signals may be generated by considering the colour of the foreground image or the luminance of said image. Colour keying is particularly attractive, given that foreground objects may be shot against a particular predefined colour, such as a saturated blue, thereby allowing all other colour combinations to be retained within the foreground image. In a composite video environment, such as environments embodying the PAL or NTSC recommendations, image data is conveyed as luminance signals, plus colour difference or chrominance signals and when deriving key signals from the chrominance information, the technique is often referred to as chroma keying. Thus, a key signal may be generated which selects a foreground image signal as an output when said signal is not blue, while selecting another video source when said foreground signal is blue.

In cinematographic films a similar result may be achieved by matting different areas of two source films to black and then exposing a third film with both of said sources. Thus, it will be appreciated that within the cinematographic film environment, it is not possible to generate a key switching signal unless the filmed images are sampled and processed within a video environment. However, it is possible to matte or suppress images within the films and thereby effect a combining process.

This matting process, in which image data is suppressed rather than being keyed, may be emulated within a video environment such that, when, for example, a blue background is detected, video signals may be suppressed to black, thereby allowing them to be additively combined with other video sources.

A system in which colour video signals represented as red, green and blue components are selectively suppressed, in response to detecting blue screens, is disclosed in United States Patent No. 3,595,987. In this patent, video signals are suppressed if a blue value is larger than green or red values detected for the same position. Thus, in areas where the input signal consists solely of blue, the blue signal is suppressed to black, thereby allowing it to be combined with another source. Chroma-matting images shot against a blue background is readily performed when processing RGB signals, because the blue background is easily identified as an area having relatively high blue values with relatively low red and green values.The technique is so successful that in some situations video source material recorded as luminance plus chrominance information may be converted into RGB, blue matted and then re-converted to luminance plus chrominance values after being combined within an RGB environment. However, such conversions lead to degradation and additional complexity.

A more direct approach may be adopted when a video signal is made up of components defining luminance (Y) and colour difference signals used to quadrature modulate a sub-carrier, for broadcasting or recording signals in composite form, such that the extent of modulation may be defined in terms of the amplitude and phase of the sub-carrier, representing saturation and hue respectively. Thus, when matting with reference to chrominance signals, it is possible to identify any hue or any range of hue signals as angles defined by the U and V vectors.

In theory, the processes of keying between video signals or suppressing and then adding video signals appear equivalent. However, in practice, it is often found that for particular examples of source material, a different result may be obtained by using the different processes. In sophisticated high-end professional equipment, it is therefore desirable to provide both techniques, so that a selection may be made by an operator, determined by the particular type of source material available. Ideally, the source material will have been recorded in real time in a way that facilitates the combining process.

However, in many situations, particularly in post-production, an editor is forced to use source material which is far from ideal, therefore it is desirable to provide the editor with a wide range of keying and matting facilities.

A common problem that occurs when keying or suppressing source material is that it is often not possible to derive suitable separation control signals on a single pass. For example, although many adjustments may be made to the equipment for deriving suppression or keying signals, it is often not possible to identify a configuration where all of the unwanted image portions are removed while all of the required image portions are retained.

In a digital environment, video matting consists of suppressing the Y, U and V samples to black in particular regions of the image, so that the image data may be added to another video signal. A suppression signal is generated in response to the detection of a particular range in hues, defined in terms of their U and V values. In order to produce smooth edges, it is necessary to consider a range of hues and the particular range selected may be defined by a wedge angle phi. Thus, it is possible to ensure that the whole of a screen image, usually blue, is suppressed from a foreground signal by increasing the acceptance angle. However, a problem with increasing the acceptance angle is that the number of non-suppressed colours within the foreground image will be reduced, therefore it is possible that required foreground colours will also be suppressed, effectively creating holes in the foreground image. Thus, it is desirable to maximise the suppression angle to ensure that the whole of the blue screen is suppressed while at the same time it is desirable to minimise the acceptance angle to ensure that all of the required colours in the foreground image are retained. Furthermore, the acceptance angle should not be made too small, so as to maintain smooth linear blending between the foreground signals and the background signals.

A known solution to mitigate this problem is to perform the chroma matting technique using a plurality of chroma matting devices connected in tandem. Thus, a first keying process may remove, say, 95% of the unwanted background, with the final 5% being removed by a more sensitive second pass. However, it would be desirable to achieve more efficient suppression signal generation on each pass, thereby improving the efficiency of the chroma matte or allowing it to effectively matte relatively lower grade source material.

According to a first aspect of the present invention, there is provided a method of generating a suppression signal arranged to suppress a luminance component and/or chrominance components of a video signal, comprising steps of generating a first primary suppression signal from said chrominance components; generating a second primary suppression signal from said luminance component; and combining said first primary suppression signal with said second primary suppression to generate a composite suppression signal.

In a preferred embodiment, the chrominance components define colour hue in terms of hue angle and the first primary suppression signal may be generated in response to identifying colours within a predetermined suppression angle. Preferably, hue angle vectors are rotated onto orthogonal axes prior to being processed in combination with said hue angle to produce said first primary suppression signal.

Preferably, the second primary suppression signal is derived in response to high values of said luminance component. In an alternative embodiment, the second primary suppress ion signal is derived from low values of the luminance component.

According to a second aspect of the present invention, there is provided apparatus for generating a suppression signal arranged to suppress a luminance component and/or chrominance components of a video signal, comprising means for generating a first primary suppression signal from said chrominance components; means for generating a second primary suppression signal from said luminance components; and combining means for combining said first primary suppression signal with said secondary primary suppression signal to generate a composite suppression signal.

In a preferred embodiment the apparatus includes a masking circuit arranged to generate a third primary suppression signal. Preferably, the masking circuit includes pixel period counting means and line period counting means.

The invention will now be described by way of example only, with reference to the accompanying drawings in which: Figure 1 shows a video mixing desk, including circuitry for combining video signals by suppressing background screen colours; Figure 2 shows a process for combining video signals by suppressing background screen colours including a suppression signal generator and a suppression signal subtractor; Figure 3 shows the identification of a particular suppression hue; Figure 4 shows the expansion of a selected hue into a suppression angle representing a range of colour values;; Figure 5 details a generator for generating suppression signals and suppression signal subtractors, of the type shown in Figure 2, embodying the present invention and including a generator for generating primary suppression signals from chrominance information, generators for generating primary suppression signals from luminance information, mask signal generators for generating masking primary suppression signals and video subtracting circuits for suppressing chrominance and luminance information; Figure 6 details the generator for generating a primary suppression signal from the chrominance information shown in Figure 5; Figure 7 details a generator for generating a primary suppression signal from luminance information, of the type shown in Figure 5; Figure 8 illustrates the operation of the primary suppression signal generator shown in Figure 7;; Figure 9 details a mask signal generator for generating a masking primary suppression signal of the type shown in Figure 5; Figure 10 illustrates the operation of the mask signal generator shown in Figure 9; Figure 11 details a suppression signal mixing circuit, of the type shown in Figure 5, including a suppression signal adding circuit and a suppression non-additive mixing circuit; Figure 12 details the suppression signal adding circuit shown in Figure 11; Figure 13 details the suppression signal non-additive mixing circuit shown in Figure 11; Figure 14 details a video subtraction circuit of the type shown in Figure 5; and Figure 15 illustrates the operation of the video subtraction circuit shown in Figure 14.

A video mixing desk 15 is shown in Figure 1, arranged to receive video signals from video cameras, such as video camera 16, video tape recorders, such as VTR 17 and other video sources, such as laser disc 18.

The mixing desk 15 is similar to that manufactured by the present assignee and sold under the trade mark "ALPHA 500". However, in addition to facilities found on said equipment, the mixing desk 15 is also capable of combining video images by suppressing portions of partial video signals such that said images may be additively combined.

Input video and output video clips may be viewed on monitors such as monitor 19 and monitor 20. Output signals from the mixer 15 may be supplied to broadcast mixing devices 21, for immediate broadcast, or to recording devices such as VTR 22.

An example of image combination is also shown in Figure 1. The first video clip VIDEO 1 consists of an actress shot in real time against a saturated blue background, illustrated in Figure 1 as image 23. In addition, the mixer 15 also receives a recorded background video signal, VIDEO 2, illustrated in Figure 1 as image 24, consisting of an outdoor scene including trees etc.

These two images are combined in real time to produce a composite signal, VIDEO OUT, illustrated as image 25, in which the actress in the studio appears to be positioned within the outdoor scene. Thus, in this way, the outdoor scene may be recorded during optimum weather conditions etc. and made available, via a video tape recorder, for use when required. Similarly, the actress may be shot in real time or, alternatively, images may be recorded onto video tape at an appropriate time, whereafter, as a post-production process, the two sources of video material may be combined to produce the final output.

As disclosed in United States Patent No. 5,305,107, several layers of video source material may be combined, particularly when the video image data is recorded in digital format. Furthermore, in addition to combining major storyline constituents, the system may also be used to add special effects and graphics etc.

The mixer 15 includes circuitry for combining the foreground video signal (VIDEO 1) with the background video signal, VIDEO 2. The foreground video signal includes a foreground image, the actress in image 23, which is to be retained, and the background screen colour which must be suppressed to black, so that background image data may be added in these areas. Similarly, the background video signal consists of an obscured area, that is an area obscured by the foreground image, and a background image which is to be retained in the final composite. The obscured area is defined by data generated by processing the foreground video signal and, depending on preferred implementations, the obscured area may be suppressed to black, like the background screen in the foreground video, or, alternatively, the obscured area may effectively be cut out by generating a background key signal.

The combining operation is shown schematically in Figure 2. The foreground video signal VIDEO 1 (represented by image 23) is supplied to a suppression signal subtractor 31 configured to subtract a suppression signal generated by a suppression signal generator 32 from said foreground video signal. All of the video signals processed in the equipment are defined in terms of three components, representing luminance Y, along with two colour difference signals U and V. Thus, suppression signals are generated by the suppression signal generator 32 for each of the Y, U and V components, which are subtracted from their respective video components for Y, U and V.

The suppression signal generator 32 also generates a key signal that is supplied to a keyer 33, arranged to receive the background video signal VIDEO 2 (represented by image 24). The subtractor 31 subtracts the suppression signal from the foreground video to produce a first intermediate video signal VIDEO 3, represented by image 34, in which the foreground image is retained but the background screen colour is suppressed to black.

Similarly, the keyer 33 effectively cuts out the obscured area from the background video to produce a second intermediate video signal VIDEO 4, represented by image 35. Thus, the VIDEO OUT signal is produced by adder 36 by adding the partially suppressed intermediate signals VIDEO 3 and VIDEO 4.

Traditionally, the background screen colour is a saturated blue to facilitate the suppression of said signal when the video information is derived in RGB format. The generation of the suppression signal in YUV colour space presents further complications, given that none of the three colour components alone are maximised, to the exclusion of the others, when a saturated blue image is detected. Thus, the saturated blue colour screen, or any other selected screen colour, must be defined within YUV colour space.

As is well known in the art, the U and V signals quadrature modulate a chrominance sub-carrier such that the instantaneous amplitude of said carrier represents saturation and the instantaneous phase angle represents hue. In theory, a colour screen of any particular hue may be selected and suppression signal generation then occurs in response to identifying hue values which fall within a particular region.

The procedure for generating suppression signals in response to hue detection is illustrated in Figure 3. The U and V colour difference signals are combined to generate a chrominance signal by quadrature modulation. A carrier sinusoid is quadrature modulated, resulting in adjustments being made to the amplitude and phase of said carrier. Thus, the instantaneous value of the carrier may be represented by a phasor defined with respect to mutually perpendicular U and V axes, illustrated as 41 and 42 respectively in Figure 3. The instantaneous colour hue may then be defined by a suppression angle theta measured anti-clockwise from the U axis 41.Thus, a particular hue is defined by considering both its U and V components but a suppression signal is defined solely in terms of its amplitude, therefore it is necessary to generate the suppression signal with reference to both the U and V components.

The system is not restricted to suppressing only saturated blue backgrounds and any particular hue may be selected. The suppression hue is identified by defining the suppression angle theta, illustrated in Figure 3, and in order to generate an actual suppression signal, an instantaneous hue value is rotated towards both the U and V axes by an angle determined by the hue angle theta.

The identification of one particular angle would result in a very hard edge, given that only pixels having the specific predetermined hue would actually be suppressed. It is therefore necessary to expand the suppression hue into a suppression wedge, defined in terms of a wedge angle phi.

Referring to Figure 4, the suppression hue has been rotated by suppression angle theta onto the hue axis. The suppression hue is then expanded symmetrically about the U axis by positive and negative angles phi. Thus, the suppression hue (theta) defines a central suppression colour, while the wedge angle (phi) defines the level of deviation away from said suppression colour that will still result in a suppression signal being generated.

An example of a suppression signal generation circuit 32 embodying the present invention is shown in Figure 5. The U and V chrominance components of the foreground video signal (VIDEO 1) are supplied to a primary suppress signal generator 51 arranged to generate a primary suppression signal in response to the U and V signals, in combination with operator defined values for the suppression hue theta and the wedge angle (phi), as illustrated in Figure 3 and Figure 4 respectively.

In addition, the suppression signal generator 32 is also arranged to receive the luminance component of the foreground video signal (VIDEO 1) that is supplied to primary suppress signal generators 52 and 53. Thus, the suppress signal generator 32 is arranged to generate primary suppression signals (arranged to suppress all components of the appropriate video signal) by processing signals derived from the chrominance information of the video signal to be suppressed. In addition, the suppression signal generator is also arranged to generate suppression signals derived from the luminance information, whereafter the chrominance derived suppression signal may be combined with the luminance derived suppression signal to produce a modified suppression signal derived with reference to both the chrominance and luminance information of the original foreground video signal.

Generator 52 is arranged to generate a primary suppression signal, based on the luminance information, in regions where the luminance signal is very low. Thus, in areas of the original image (VIDEO 1) where the luminance value is below a predetermined threshold, a primary suppression signal will be generated which may be used to suppress luminance and/or chrominance information in the relevant portion of the picture. This primary suppression signal derived from low intensity luminance values may be identified as a "lolite".

Similarly, primary suppression signal generating circuit 53 is arranged to generate a primary suppression signal where the luminance information is relatively high. Thus, in portions of the foreground video signal (VIDEO 1) in which the luminance values exceed a predetermined threshold, a primary suppression signal will be generated which may be used to suppress luminance and/or chrominance information. Similarly, such a primary suppression signal may be referred to as a "hilite".

Modified suppression signals suppress portions of the video image which are not required in the final output image. In some situations it is possible to identify relatively large regions of a video field which are not required in the output image. In the generator shown in Figure 5, operator controlled mask signal generators 54 and 55 are provided which mask out regions of the video image without requiring sophisticated suppression techniques based upon the input image data. In the present embodiment, the suppression signal generator 32 is provided with a first mask generator 54 and a second mask generator 55, each arranged to produce primary suppression signals by defining rectangular masks around areas of interest.

Circuits 51, 52, 53, 54 and 55 are all arranged to generate primary suppress ion signals which may be used to suppress chrominance information and/or luminance information from the foreground video signal. A first suppress signal mixing circuit 56 is arranged to generate a primary suppression signal for the chrominance information, consisting of the U colour difference signal and the V colour difference signal. The output from the first suppress signal mixing circuit 56 may be considered as a composite suppression signal which is supplied to a U and V suppression signal subtractor 57 within the suppression signal subtractor circuit 31. Similarly, a second suppress signal combiner 58 is arranged to generate a composite luminance suppression signal which is supplied to a luminance signal subtractor 59, again forming part of circuit 31.Each suppression signal combining circuit 56 and 58 is arranged to receive all five of the primary suppression signals, generated from chrominance information, via primary suppression signal generator 51, generated from luminance information, from the hilite and lolite primary suppression signal generators 52 and 53 and from user defined mask data, received from the mask primary suppression signal generators 54 and 55.

In each of the combiners 56 and 58 any combination of the primary suppression signals may be used in order to generate a composite suppression signal. Thus, after a plurality of primary suppression signals have been generated, via circuits 51 to 55, a composite chrominance suppression signal is generated by processing a first set of the primary suppression signals.

Independently, a composite luminance suppression signal, for suppressing luminance information, is generated from a second set of the primary suppression signals. The chrominance signal and the luminance signal are suppressed by their respective chrominance suppression signal and luminance suppression signal. However, in response to operator control, the selection of a set of primary suppression signals for the chrominance suppression is made totally independently of a selection of a second set of primary suppression signals for the suppression of the luminance information.

It should also be noted that the way in which the modified suppression signals are generated departs significantly from the way in which the signals were generated previously. It is known to generate suppression signals based on chrominance information and in the present embodiment a suppression signal of this type is supplied as a primary suppression signal to both of the combiners. However, in addition, primary suppression signals are also generated from the luminance components of the input video signal such that, within a combiner such as combiner 56 or combiner 58, a primary suppression signal derived from the chrominance components may be combined with a primary suppression signal derived from the luminance component, to produce a modified suppression signal for suppressing the luminance and/or the chrominance information.

The circuit 51 for generating a primary suppression signal from the chrominance information is detailed in Figure 6.

Hue values are rotated by the suppression angle theta onto the U axis and by an angle of 90 degrees minus theta onto the V axis, to produce rotated components U prime and V prime. A rotation processor 51 multiplies the U component by the cosine of the suppression angle theta and adds this to the V component multiplied by the sine of the suppression angle theta to produce the value U prime. Similarly, a V rotation processor 52 multiplies the V component by the cosine of the suppression angle theta and subtracts from this the U component multiplied by the sine of the suppression angle theta.

The values of U prime and V prime are supplied to a suppression signal generation processor 53, configured to take account of the wedge angle phi.

Thus, the modulus of the value V prime is divided by the tangent of the wedge angle phi, which is in turn subtracted from the value U prime in order to generate a chrominance based primary suppression signal.

A similar processor to processor 53, receiving the components U prime and V prime, may be used to generate a primary signal for the background video key signal, ie. the background image hole cutter.

It has been found that a more effective blending may be obtained if the background screen colour is allowed, to some extent, to be combined with the foreground image, to a greater extent than that to which the foreground image is allowed to blend with the background image. Thus, although it is known for blending to occur at the interface of the two images, it has now been found that a greater proportion of the background image should be allowed to be blended with the foreground image, rather than a straight 50-50 split.

This may be achieved by having a different wedge angle phi for the circuitry arranged to generate the background key signal K compared with the wedge angle phi used to generate the foreground suppression signal. Thus, the wedge angle phi when used in circuit 53 is made smaller, by an adjustable factor, compared to the wedge angle phi used when generating the key signal used in keyer 33, to cut the obscured area hole in the background image 24.

Circuit 53, arranged to generate a primary suppression signal from relatively high luminance values ("hilites") is detailed in Figure 7. Circuit 52, arranged to generate a primary suppression signal from relatively low luminance values ("lolites") is of a substantially similar configuration and the difference in operation is achieved by supplying different control values to the circuit.

Referring to Figure 7, the suppression circuit 53 includes an adder 71, a multiplier 72 and a limiting device 73. The adder 71 is arranged to lift the signal and receives a lift control value in addition to receiving the input luminance value. An example of an input luminance value is shown in Figure 8A. The graphical representation of the luminance signal shown in Figure 8A represents luminance values for a particular scan line of a particular field. Each pixel of a video image may have a luminance value which ranges between a lower bound 81 and an upper bound 82. In the example shown, the luminance values start off as being relatively high over a portion 83, whereafter they reduce and are relatively low over a portion 84, whereafter they increase again and are relatively high over a portion 85.

Thus, the highlight circuit 53 is configured to produce a primary suppression signal derived from the relatively high luminance values present over portions 83 and 85. Similarly, the lolite primary suppress signal generator 52 is configured to generate a primary suppression signal from relatively low luminance values, such as those present over portion 84.

In order for circuit 53 to operate as a hilite primary suppress signal generator, a negative value is supplied to the lift input of the adder 71, thereby effectively applying negative lift to the luminance signal as illustrated in Figure 8B. Thus, the central portion 84 has been reduced to a level which lies below the lower bound and the high luminance values over portions 83 and 85 have been reduced to relatively low luminance values, illustrated by portions 86 and 87.

The output from the summation circuit 71 is supplied to the multiplier 72 which also receives a gain value from a central control processor. The multiplier 72 is configured to effect real-time multiplications on the digital samples of the luminance input signal and is therefore a relatively high speed device implemented as a hardware multiplier. A positive gain value supplied to the multiplier 72, thereby expanding the lifted values shown in Figure 8 to provide much larger values in both positive and negative sense. These values are supplied to the limiting circuit 73, which in turns limits values to within the lower bound 81 and the upper bounded 82. The result of the signal shown in Figure 8B being applied to the multiplier 72 and the limiting circuit 73 is shown in Figure 8C.The portions 86 and 87 are increased significantly in value by the gain circuit 72 such that they will be greater than the upper bound value 82. The limiting circuit 73 then limits these values to the maximum permitted value, defined by the upper bound 82, to provide a well defined primary suppression signal. Thus, the suppression signal derived from the luminance hilite will have a full suppress value at portions 88 and 89 with zero suppress signals over portion 90.

Circuit 52 will also receive the luminance signal shown in Figure 8A.

However, on this occasion, positive lift will be supplied to the addition circuit, followed by negative gain supplied to the multiplication circuit 72, such that the resulting primary suppression signal will effectively be the inverse of the signal shown in Figure 8C, with positive suppression being supplied over the central low luminance region.

The mask signal generators 54 and 55 are substantially similar and the particular position of the masks is defined in response to user controlled parameters. Circuit 54 is detailed in Figure 9 and includes a first counter 91 arranged to count pixels and a second counter 92 arranged to count lines. A typical mask produced by the mask signal generator is shown in Figure 10.

The mask signal generator is arranged to generate a primary suppression signal which suppresses video information contained within a boundary region. Thus, video information present within the central rectangle is retained and the size of this retained rectangle may be specified in terms of a horizontal width H and a vertical height V. Thus, the position of a left edge 101 is specified by an operator, along with the position of a right edge 102, with the distance between these two edges being equal to the window width H.

The position of the left and right edges is defined in terms of a pixel count, therefore left and right edge definition positions are considered with reference to the count of pixels produced by counter 91. Similarly, a top edge 103 is defined along with a bottom edge 104. On this occasion, the position of the top and bottom edges are defined in terms of line positions and are considered with reference to the count produced by counter 92 arranged to receive the line clock.

A mask signal generator includes a first comparator 93, a second comparator 94, a third comparator 95 and a fourth comparator 96. It also includes a first adding circuit 97, a second adding circuit 98 and a third adding circuit 99. Each of the comparators is arranged to compare a count value against an edge definition, such that high values are generated in the boundary area, to produce a positive suppression signal, with low values being generated in the central window area. Gates 97, 98 and 99 are arranged to produce high values, representing a primary suppression signal, when either of their inputs are placed into a high condition. Alternatively, if both inputs are in a low condition a low output is generated representing the central mounded and non-suppressed region.

The output from the pixel clock counter is supplied to the left edge comparator 93 and to the right edge comparator 94. Comparator 93 also receives a value indicating the position of the left edge and a true (ie. high) output is produced if the clock count value is lower than the left edge value.

Under these circumstances a true value is supplied to OR gate 97, which in turn supplies a true value to OR gate 99 resulting in a positive suppression signal being generated by the circuit 54.

Right edge comparator 94 also receives the output from the pixel clock counter 91 along with a definition of the position of the right edge. A true output is produced when the pixel clock count is larger than the right edge position, again resulting in a positive suppression signal being generated. Top edge comparator 95 receives an output from the line clock counter 92 and produces a true output to gate 98 when the top edge value is larger than the line clock count. Similarly, the output from the line clock counter 92 is also supplied to the bottom edge comparator 96 which in turn produces a true output when the line clock count is larger than the bottom edge value.

It can be seen that a primary suppression signal is not generated for the central region 105 of the mask shown in Figure 10, which may be defined for conditions in which the pixel clock count is larger than the left edge definition but smaller than the right edge definition, while the line clock count is larger than the top edge definition but smaller than the bottom edge definition.

The suppression signal generator 32 is detailed in Figure 5 and includes a primary suppress signal combiner 56 for suppressing the chrominance values U and V and a separate primary suppress signal combiner 58 for combining primary suppress signals to generate a modified luminance suppression signal. A plurality of standards exist for transmitting digitised component video signals, often with variations occurring concerning the relative sample rates between the chrominance samples and the luminance samples. In particular, in some situations all the components are sampled at the same rate, for example 13.5 megahertz or, alternatively, in order to make full use of the available bandwidth, the chrominance components may be sampled at a lower sampling rate. It will be appreciated that when, for example, chrominance originating signals are used to generate luminance suppress ion signals, it may be necessary to provide over-sampling circuits, so that the sample rates are unified. It will be assumed that circuits of this type are included where necessary and that the combiner 56, arranged to generate the chrominance suppression signal and the combiner 58 for the luminance suppression signal are operating at equivalent sampling rates. Consequently, the circuitry contained within combiner 56 is substantially similar to the circuitry contained within combiner 58.

Combiner 58, arranged to receive primary suppress signals and to generate a modified suppression signal for the luminance signal, is detailed in Figure 11 although, as previously stated, circuit 56 is substantially similar.

The combiner 58 is arranged to receive primary suppression signals from both of the mask signal generating circuits 54 and 55. Both of these masking signals are supplied to each of the four mask select circuits 111, 112, 113 and 114. Each circuit includes invertors for inverting both of the mask signals, whereafter a selector may be positioned into one of five positions, in response to operator commands. Thus, the output from each mask select circuit 111 to 114 may comprise of the non-inverted mask 1 signal, an inverted mask 1 signal, a non-inverted mask 2 signal, an inverted mask 2 signal or a null signal, ie. effectively applying a constant level to subsequent circuitry.

The output from each mask select circuit 111 to 114 is supplied to respective masking circuits 115, 116, 117 and 118. The hilite primary suppress signal, generated via circuit 53, is scaled via a multiplier 119 and then masked by masking circuit 115, in response to a mask definition generated by the mask select circuit 111. Similarly, the lolite primary suppress signal generated by circuit 52 is scaled by a multiplier 120, before being masked in masking circuit 116 arranged to receive a masking signal from masking select circuit 112. The chroma primary suppress signal, generated by chroma suppression generator 51 is supplied directly to a similar masking circuit 117, arranged to receive a mask signal from the mask select circuit 113.

A non-additive mixing circuit 121 combines the outputs from the masking circuit 117 and the masking circuit 116. The output from the nonadditive mixing circuit 112 is itself supplied to a second non-additive mixing circuit 122 which, at its second input, receives an output from the masking circuit 115. Thus, the non-additive mixing circuit 112 is arranged to combine the combination of signals derived from the chroma primary suppress signal and the lolite primary suppress signal, with the signals derived from the hilite primary suppress signal. The output from non-additive mixing circuit 112 is supplied to a further masking circuit 118 which in itself is arranged to receive a masking signal from mask select circuit 114.

The masking circuits 115, 116, 117 and 118 are substantially similar and masking circuit 117 is detailed in Figure 12. The primary suppress signal derived from the chrominance information is supplied directly to a switch 121 directly in addition to being supplied to another terminal of switch 121 via an invertor 122. Thus, switch 121 is arranged, in response to operator commands, to select the non-inverted or the inverted chrominance- derived primary suppress signal. The output from switch 121 is supplied to an AND gate 122, such that a true output from AND gate 121 is produced when both of its inputs are true. Thus, in this way, the primary suppress signal derived from the chrominance information may be logically ANDed with the masking information which, as previously stated, may be derived from non-inverted or inverted mask signals.

A second switch 124 is provided, arranged to receive the non-inverted output directly from AND gate 123 in addition to an inverted version of said output, derived via an invertor 125. Thus, the output from a masking circuit may also be selected to its non-inverting or inverting polarity.

The non-additive mixing circuits 121 and 122 are substantially similar and mixing circuit 121. is detailed in Figure 13. Both inputs, derived from masking circuit 117 and masking circuit 116, are supplied to a comparator 131 and a switch 132, wherein said switch 132 is controlled in response to the output from said comparator 131. Thus, the input signals are combined in a non-additive fashion by being selected, one in preference to the other, in response to operations of the comparator 131.

The output from switch 132 is inverted by an invertor 133 and a switch 134 is provided so as to select either the non-inverted or inverted output from switch 132. In this example, the output from switch 134 provides an input to the second non-additive mixing circuit 122. The output from masking circuit 188 provides the modified suppress signal for the luminance information, which is supplied to the Y suppression circuit 59.

The Y suppression circuit 59 is detailed in Figure 14. The circuit shown in Figure 14 is arranged to suppress the luminance foreground signal, in response to a luminance suppression signal generated by the combiner 58.

Suppression circuit 57 includes two circuits of the type shown in Figure 14, arranged to suppress respective colour difference signals U and V in response to a chrominance suppression signal received from combiner 56.

The luminance suppression circuit 59 includes a comparator 141, a first gate 142 arranged to receive the suppression signal from the combiner 58, a second gate 143 arranged to receive the luminance input signal, a multiplier 144 arranged to receive a scaling factor and a subtractor 145, arranged to perform the subtraction of a modified suppression signal.

The combiners 56 and 58 each receive a plurality of primary suppression signals and, within each combiner, an independent selection of primary suppression signals are made in order to produce respective composite suppress signals, a first being arranged to suppress the U and V colour difference signals and a second being arranged to suppress the luminance signal. Referring to Figure 14, the composite suppress signal received from combiner 58 is supplied to comparator 141 and to gate 142.

An example of a composite suppress signal is shown in Figure 15A, having a less than ideal suppression characteristic. A suppression signal has been generated from a foreground video signal in which the central portion of the signal 151 represents the position of the foreground image, where suppression is not required, while the edge regions 152 represent the background screen where suppression is required. In the example shown in Figure 15A, the suppression signal dips below an ideal value over periods 152 therefore it is possible that some of the background screen signals will not be suppressed sufficiently.In the embodiment, the masking circuits may have been used to mask out large regions of the screen, however, due to the irregular shape of most real foreground images, some areas of the screen will still exist which need to be suppressed using more sophisticated techniques. However, as seen from Figure 1 SA, this may result in unsatisfactory suppression taking place.

A composite suppress signal, illustrated in Figure 15A, is supplied to the comparator 141 which compares the suppress signal against a references signal, adjustable in response to operator commands. The relative amplitude of the reference signal is indicated in Figure 15A by dotted line 153. Thus, the comparator compares the values of the composite suppress signal over regions 151 and 152, against the reference signal identified at level 153. The result of the comparison is shown in Figure 15B. When the composite suppress signal is larger than the reference signal the output from the comparator is set low. Similarly, when the suppress signal is lower than the reference signal the output from the comparator is set high. Thus, the comparator output is low over periods 152 and high over periods 151.

The output from comparator 141, illustrated in Figure 15B, is supplied to AND gate 142, which also receives the composite suppress signal illustrated in Figure 1 5A. AND gate 142 produces an output derived from the composite suppress signal, such that the composite suppress signal is allowed to pass through the gate when the output from the comparator 141 is high.

Thus, over periods 152 the output from comparator 142 is low therefore the output from AND gate 142 is also low. However, over period 151 where the output from comparator 141 is high, the composite suppress signal is allowed to pass, therefore the output from AND gate 142, represented in Figure 15C, is substantially similar to the representation of a composite suppress signal shown in Figure l5A.

The composite suppress signal is supplied to AND gate 142 via a multiplier 44, arranged to receive an operator controlled scaling factor as its other input. Thus, the composite suppress signal gated through AND gate 142 may be scaled, in response to an operator controlled scaling factor, before being used to suppress the modified luminance signal gated through AND gate 143. Scaling of this type is particularly important when a suppression signal is being used to suppress a video signal of the first type (for example luminance) derived substantially from video information of a different type, such as chrominance. Thus, the suppression signal is being used to suppress the luminance value in Figure 14, the actual suppression signal itself may have been derived substantially from chrominance information.Similarly, when suppressing chrominance, it is possible that scaling will be required if the suppression signal is derived substantially from the luminance information. Furthermore, given that the original colour difference signals U and V represented mutually displaced vectors, it is necessary to multiply a modified suppression signal for the U signal by the cosine of the suppression angle theta, while, similarly, the modified suppression signal for the V colour difference signal is multiplied by a representation of the sine of said angle theta.

The luminance signal (or when considering circuit 57, a colour different signal U or V) is supplied to AND gate 143, which is also arranged to receive, at its second input, the output from comparator 141. AND gate 143 performs a substantially similar function to AND gate 142. Thus, while the output from comparator 141 is low, the luminance signal will effectively be blocked by AND gate 143, thereby effectively providing a suppression function. Thus, if a constant luminance value, for example, is supplied to GATE 143, the output from gate 143 will substantially follow the output from comparator 141 and reflect the signal shown in Figure 1 SB. If the luminance value is fluctuating, these fluctuations will be reflected in the output from AND gate 143 over periods during which a high or true signal is generated by the comparator 141.

Thus, over periods 152, where suppression is required, the luminance value is suppressed by means of AND gate 143 and it is not necessary to rely upon the less than ideal suppression signal. However, over the period 151 where suppression is required, in order to effect smooth blending, the luminance signal is gated through AND gate 143 to the positive input of a summation circuit 145 while, similarly, the suppress signal, shown in Figure l5C, is also gated to the negative input of circuit 145. Thus, the suppression signal shown in Figure 15C is subtracted from the luminance signal to produce the suppressed luminance signal.If a constant luminance value is supplied to the gate 143, suppression of this signal will result in an output suppress luminance signal similar to that shown in Figure 1 SD. Thus, over periods 142 suppression is performed by the combination of the comparator 141 and AND gate 143. Within the central region 151 a suppression is performed by the subtraction process effected by circuit 145.

The modified suppression circuit generated by AND gate 142 is supplied to the subtracting input of adding circuit 145, thus it is this circuit 145 which actually performs suppression of the video information, by subtracting the modified suppress signal from the modified luminance.

The overall system shown in Figure 5 may be summarised as follows.

Source signals are provided in the form of the chrominance signals, the luminance signals and mask signals from which primary suppression signals are derived. Individual combiners are provided for receiving these primary suppression signals such that different selections of primary suppression signals may be combined in order to provide a chrominance suppression signal and a completely independently derived luminance suppression signal.

The output from the combiners may be referred to as composite suppression signals, given that they are derived by combining primary suppression signals.

Each composite suppression signal is in turn processed by comparators gates within the suppression circuit to produce a modified suppression signal before actual suppression takes place.

Although the system shown in Figure 5 provides a high level of sophistication for suppressing chrominance and luminance signals, it is also possible to see the system shown in Figure 5 as a building block forming a module in a larger suppression environment. Thus, for example, all the constituents of Figure 5 may be repeated such that the generation of suppression signals may be effected in two or more stages.

Claims (22)

1. A method of generating a suppression signal arranged to suppress a luminance component and/or chrominance components of a video signal, comprising steps of: generating a first primary suppression signal from said chrominance components; generating a second primary suppression signal from said luminance component; and combining said first primary suppression signal with said second primary suppression to generate a composite suppression signal.

2. A method according to claim 1, wherein said chrominance components define colour hue in terms of a hue angle.

3. A method according to claim 2, wherein said first primary suppression signal is generated in response to identifying colours within a predetermined suppression angle.

4. A method according to claim 3, wherein hue angle vectors are rotated onto orthogonal axes prior to being processed in combination with said hue angle to produce said first primary suppression signal.

5. A method according to any of claims 1 to 4, wherein said second primary suppression signal is derived in response to high values of said luminance component.

6. A method according to any of claims 1 to 4, wherein said second primary suppression signal is derived from low values of said luminance component.

7. A method according to claim 5 or claim 6, wherein said second primary suppression signal is generated by applying lift and gain to a luminance signal.

8. A method according to any of claims 1 to 7, wherein a third primary suppress ion signal is derived from a masking circuit.

9. A method according to claim 8, wherein said masking signal is produced by counting pixel periods and by counting line periods.

10. A method according to any of claims 1 to 9, wherein said first primary suppression signal is combined with said second primary suppression signal by a process of non-additive mixing.

11. Apparatus for generating a suppression signal arranged to suppress a luminance component and/or chrominance components of a video signal, comprising means for generating a first primary suppression signal from said chrominance components; means for generating a second primary suppression signal from said luminance components; and combining means for combining said first primary suppression signal with said secondary primary suppression signal to generate a composite suppression signal.

12. Apparatus according to claim 11, wherein said means for generating a first primary suppression signal is configured to process chrominance components defining colour hue in terms of a hue angle.

13. Apparatus according to claim 12, wherein said means for generating a first primary suppression signal is arranged to identify colours within a predetermined suppression angle.

14. Apparatus according to claim 13, wherein said means for generating a first primary suppression signal include means for rotating hue angle vectors onto orthogonal axes, prior to processing said rotated vectors in combination with said hue angle to produce said first primary suppression signal.

15. Apparatus according to any of claims 11 to 14, wherein said means for generating said second primary suppression signal includes means for generating said signal in response to high values of said luminance component.

16. Apparatus according to any of claims 11 to 14, wherein said means for generating said second primary suppression signal includes means for deriving said suppression signal from low values of said luminance component.

17. Apparatus according to claim 15 or claim 16, wherein said means for generating said second primary suppression signal includes means for applying lift and gain to a luminance signal.

18. Apparatus according to any of claims 11 to 17, including a masking circuit arranged to generate a third primary suppression signal.

19. Apparatus according to claim 18, wherein said masking circuit includes pixel period counting means and line period counting means arranged to generate said masking signal.

20. Apparatus according to any of claims 11 to 19, wherein said combining means is arranged to combine said first primary suppression signal with said second primary suppression signal using non-additive mixing means.

21. A method of generating a suppression signal arranged to suppress a luminance component and/or chrominance components of a video signal, substantially as herein described with reference to the accompanying figures.

22. Apparatus for generating a suppression signal arranged to suppress a luminance component and/or chrominance components of a video signal, substantially as herein described with reference to Figures 5, 6, 7, 9, 11. 12, 13 and 14.