GB2532018A - Optical soft edge masks - Google Patents

Abstract

A method of applying an optical soft edge fall-off in brightness to a projector display channel 4, with known video display rate comprises: assessing the projector exit pupil for its shape and intensity profile; determining the multiple light paths from the projector exit pupil to locations of interest in the projected field; calculating the position, edge profile and magnitude of movement of an optical blend mask 112 to achieve the desired fall in brightness from consideration of the multiple light paths; and controlling the frequency or rate of movement of an optical blend mask such that unwanted visual artefacts are reduced. Optical blend masks are utilised in multi-channel projector systems to ensure that a large projected image composed of sub-images exhibits a uniform brightness. Preferably the mask moves continuously in two orthogonal directions such that the locus of the mask edge describes a circle or ellipse. The cross-section of the mask may be a complex shape such as an ellipse, oval or rugby ball shape.

Description

Optical Soft Edge Masks This invention relates to multi-channel projection display systems, and more especially optical blend masks used in multi-channel projection display systems as a means to combine sub-images into a larger single image.

Background of the invention

Optical blend masks have been used in many display systems to combine sub-images into a single, continuous larger image. Typically an optical blend mask is positioned between the exit pupil of the projector and the projection screen to provide a gradual decrease in brightness towards the edge of a display channel which is then matched with an opposite brightness profile from the adjacent projector channel. The peak intensity sum of the two images across the overlapping blend region is essentially the same as the unmasked images and consequently the display system exhibits a continuous image of uniform brightness. Varying the distance of the optical blend mask from the projector exit pupil will vary the width of the blend and this can be used to adjust blend profiles and performance to optimise the overall display.

Many examples of optical blend masks have been proposed and used successfully in many applications. Blend masks can be a hard edge, straight or contoured to match the shape of the blend or overlap region required, exhibit a sawtooth profile or may incorporate transparent and opaque regions on a blend mask. Use of a sawtooth profile was identified when overlapping images were used in a film projection display such as that used by the VITARAMA display in US2544116. This also described how the sawtooth profile could also be moved mechanically to further enhance the blend quality. This also disclosed methods and apparatus to move a straight edge mask perpendicular to the blend profile to produce the desired fall off in brightness. In this example the blend mask is moved by mechanical means linked to the carriage of the film and is synchronized to the film movement so was effectively fixed at an identical frame rate to the film. The movement could not be adjusted during use or set up as it was pre-determined by the mechanical assembly. Sawtooth optical blend masks are also described in W09525292 (Bridgwater) and W00141455 (Maximus et al.) Moving optical blend masks have also been disclosed in US2998747 in which sawtooth profile masks are arranged as vanes on a rotating assembly in order to achieve the desired fall off in brightness. US2610544 describes how masks with zones that transmit variable amounts of light can be used to achieve the desired fall off in brightness.

The above documents disclose the use of various blend masks to be used with film projectors however it is clear to one skilled in the art that all are suitable to a degree for use with the various types of video projectors available such as those incorporating LCD, LCOS and DLP spatial light modulators.

However in order to achieve high brightness from video projectors e.g. those using LCD, LCOS and DLP spatial light modulators, light must be collected from a large solid angle around the illumination source, usually a UHP type lamp that emits light from a short arc between electrodes. Parabolic or elliptical mirrors are typically placed behind the lamp to direct light towards the spatial light modulators. This alone would give a dimmer central region of the display since the lamp obstructs any light reflected from the central region of the mirror, which is normally cut away for insertion of the lamp. Light integration optics are therefore used between the lamp mirror and the spatial light modulators. This may be a pair of flyeye integrators each containing several lenslets with each lenslet on the first flyeye combining with the corresponding lenslet on the second flyeye to spread the light over the entire spatial light modulator or modulators. The desired effect is to increase uniformity of illumination across the entire display and hence produce an image that has a good uniform brightness on the screen.

However this has the result that the projector lens pupil is not illuminated uniformly, but rather consists of a pattern of bright spots corresponding to the flyeye integrator lenslets. These bright spots may also vary in colour hue across the pupil.

Integrating rods are also commonly used in video projectors. Integrating rods are typically shaped like rectangular pillars that may be slightly tapered. They can be either solid glass or hollow with mirrored internal faces. The rectangular cross section matches the aspect ratio of the spatial light modulator. Multiple reflections occur from the internal surfaces of the integrating rod, the light intensity profile is transformed into a rectangular one. The integrating rod effectively changes the luminous distribution of the light beam. The UHP type lamp light source is imaged onto the entrance surface of the rod and the resulting illumination at the exit surface is an overlap of the source images caused by the multiple reflections within the rod. The convergence angle of the beam from the reflector and the length of the rod determine the number of the reflected images. The overlap of the multiple virtual sources smoothes out the non-uniform intensity profile of the UHP type lamp light source when illuminating the image spatial modulator. However the resultant multiple virtual images result in multiple images of the illumination source within the projector lens pupil and these multiple images have to be considered when optical blend masks are used. A further possibility exists with different types of known illumination source to an arc lamp, for example LED sources, where multiple LED emitters are combined optically before being focused onto the entrance aperture of the light pipe. The multiple images of the LED emitters will form in the same way at the projector lens exit pupil as for a UHP source, but the images of the individual emitters are typically displaced from each other. If these sources are of different colours, typically red, green and blue, then it can result in a significant colour non-uniformity at the projector lens exit pupil.

When used with video projectors incorporating integration optics, prior art optical blend masks do not produce a blend region that falls smoothly from maximum brightness to zero, but tend instead to produce a region exhibiting a rippled structure. The fall-off in brightness in the blend region varies between projectors depending upon the components in the illumination optics for the particular projector. Consequently when two regions with rippled structures are overlapped to form a blended image, this almost invariably gives rise to significant brightness ripple across the blend region. This variation in brightness is unacceptable and is distracting in many applications. The blend may also exhibit colour differences across the overlap region, as a result of any colour non-uniformity at the projector lens exit pupil. The brightness ripple and colour differences across the blended overlap region represent unwanted visual artefacts.

It is an aim of the present invention to obviate or reduce the above mentioned drawbacks, providing optical blend masks that do not give rise to unacceptable artefacts or variations in brightness across the overlapped regions.

Accordingly the present invention provides a method of applying an optical soft edge fall-off in brightness to a projector display channel with known video display rate the method comprising, assessing the projector exit pupil for its shape and intensity profile, determining the multiple light paths from the projector exit pupil to locations of interest in the projected field, calculating the position, edge profile and magnitude of movement of an optical blend mask to achieve the desired fall in brightness from consideration of the multiple light paths and controlling the frequency or rate of movement of an optical blend mask such that unwanted visual artefacts are reduced.

The optical blend mask may be moved using known means such as a voice coil or other linear actuator, the rate or frequency and magnitude of movement being controlled by control means, the movement having been calculated in consideration of the optical properties of the projector. The frequency of the movement of the optical blend mask may be at a frequency that differs from the frame rate of the video being displayed. The magnitude of movement is such that the optical blend achieved gives the desired fall off in brightness. The movement may be symmetrical around a determined position or may be asymmetrical around a determined position.

In determining the rate or frequency and magnitude of movement consideration is given to the optical layout and configuration of the display, the video being displayed and also the optical properties of the projector in particular the illumination optics within the projector, the projector pupil and the projection lens.

The present invention also provides optical blend masks for use with one or more projectors for projecting an image onto a screen, the optical blend masks being in continuous motion relative to one or more projectors or one of the light paths of the one or more projectors, characterised in that the continuous motion is determined by consideration of the optical properties of the projector and the image displayed and in that the edge of the optical blend mask intersecting the light path continuously varies its position relative to the light path in a first direction relative to the light path whilst also continuously va.ry*ing in a second direction perpendicular to the first direction.

The optical blend mask is chosen in consideration of the ripple shown in the blend region when the optical blend mask is stationary and when the optical blend mask is stationary in a plurality of locations. As described the optical blend mask may be located between the projector and the projection screen. The mask is typically located closer to the projection lens with the optimum location being determined by the properties of the projection lens and other factors such as the illumination optics within the projector. If projectors are used which have the integration optics and consequent characteristics displayed by the resultant multiple pupils, ripples will be seen in the blend region with a stationary mask. The location of the ripples is determined by the location of the mask relative to the light path and the intensity distribution of the projector pupil. The optical blend mask may be moved in a direction perpendicular to the light path e.g. further into the light path or a location which lesser impedes the light path, and it can be seen that the ripples demonstrated also move. The optical blend mask may also be moved in a direction along the light path e.g. moving the optical blend mask further away or closer to the projection lens. Motion imparted to a blend mask will provide for a temporal averaging of the ripple and colour variation over the fall-off region and therefore reduces the artefacts or ripples in the blend region.

In a projector having complex integration optics motion of the blend mask edge in one direction may not he sufficient to remove ail the artefacts from the blend region. This is due to the projector exhibiting multiple exit pupils and therefore the light impinges on the blend mask from slightly different directions causing artefacts. Moving the blend mask in two directions, that is effectively into and out of the light path and also varying the distance from the projection lens reduces the artefacts further. Prior art solutions suggested using rotating vanes to achieve this but this does not produce continuous motion as the intersection of the light path with the vanes jumps at partial revolutions. Relatively small jumps at partial revolutions may result in artefacts that are not too distracting but larger jumps will result in unacceptable artefacts being seen.

The optical blend mask of the invention may be a hard edge blend mask that is moved in one direction, into and out of the light path, or moved in two directions simultaneously, that is continuously into and out of the light path whilst being continuously being moved in the orthogonal direction closer to and further from the projector. This may be achieved by known mechanical means and the frequency and magnitude of the movement can be determined or calculated to reduce artefacts and to minimize the introduction of any new artefacts that may be caused by the movement of the optical blend mask.

Alternative embodiments and methods of achieving an optical blend mask that has the required movement are also described below. An optical blend mask that has a desired cross section can achieve the required movement when rotated in the light path at the determined location. The various cross sections, profiles and shapes will be described.

The optical blend mask of the invention may be an extruded shaped mask and has a cross-section that is a continuous convex shape. The cross section may be an ellipse, oval, rugby ball shaped or a combination of contiguous convex sections forming a continuous convex profile. A continuous convex exterior profile is preferred as this presents an edge profile at all times during rotation which continuously intersects the light path of the projector. This profile effectively moves the location of the optical edge mask relative to the light path and projector continuously in two axes as the optical blend mask is rotated. The edge is effectively moved into and away from the light path whilst being simultaneously being moved closer and further away from the projector. The rate at which the mask is rotated is determined by calculation considering the system layout, the optical properties of the projector illumination system and the overlap or blend region. The rate of rotation may be continuous at the same rate or vary with time but is chosen to reduce the artefacts seen in the blend region and also chosen such that no new visible artefacts, such as flicker, are introduced. As discussed above small discontinuities in the shape of the cross section of the optical blend mask may be acceptable, that is the optical blend mask has a cross section that is net convex at al*. points but has short straight or concave sections.

The resulting blend fall-off or profile produced by the continuous motion of the optical blend masks homogenises the blend eliminating the ripples. The motion, size and location of the mask as described will vary with the optical properties of the projector or projectors. These properties include projector pupil size, projector pupil emissivity distribution, integration optics within the projector, projection lens and for a DLP projector whether the projector has an off-axis or on-axis configuration.

The edge of the optical blend mask is preferred to be always be in the light path of the projector without discontinuities. The optical blend mask of the invention achieves this as it has an exterior profile shaped for the purpose. Discontinuities can give rise to artefacts such as non-uniform blend roll-off and these have been seen with prior art moving blend masks such as those disclosed by Waller et al. Prior art has shown that motion can be used to produce roll-off at the edge of a film frame but this was mechanically linked to the film mechanism and the motion would therefore have been timed to the frame rate of the film. For the optical blend mask of the invention consideration is given to the amplitude and frequency of the motion such that the introduction of the motion does not itself introduce distracting artefacts such as flicker.

In a side-by-side configuration of projectors the profile of the optical blend mask could be chosen to facilitate blending which requires geometric correction. Optical blend masks that combine from adjacent or overlapping channels in a multi-projector display may have movement that is synchronised or not synchronised and motion in each is selected to reduce any resultant artefacts. Motion may also be random e.g. the motion is not constant in terms of amplitude and frequency but is continuously varying in one or both of these properties.

The optimum profile of the blend mask can be determined from analysis of the display system layout. Bespoke software solutions are typically used to determine display layout and optical blend mask profiles can be calculated from this software using certain design criteria. The design can also be adapted to compensate for properties such as projection pupil aberrations that may arise which could otherwise give rise to a distorted or incorrect blend profiles or geometry.

A method of determining the size, location and motion of the optical blend masks is to analyse the ripple exhibited by a stationary mask located in the position which achieves the desired roll-off in brightness throughout the blend or overlap region albeit with the ripples shown. Analysis of the ripples or brightness variation can be carried out through the data collected from a measurement device such as a digital camera from, measurements taken on the screen or measurements taken from the observer's location. Measurements could be taken automatically. Another method of determining the optical blend mask design includes the measurement of the projector lens pupil intensity and colour distribution, which may be performed using known means, such as optically re-imaging the pupil onto a screen surface and capturing this pupil image using a digital camera.

It is also noted that there are many ways to impart motion to the rotating optical blend mask each with the ability to control and vary the motion as desired or calculated. The obvious means is to use a motor, to which the mask is attached, to spin the mask imparting a motion to the mask. Consequently the speed of rotation is easily controlled and can also be easily adjusted to optimise performance.

The optical blend mask may be straight or may have a rotational twist of for example 90degrees, or 180degrees along its length. Various methods and materials exist for manufacturing the optical blend mask. These may include traditional methods such as casting and machining but can also include rapid prototyping methods such as the various 3D printing methods. Materials used can also be selected from a range including traditional materials such as metal and plastic but materials also include plastic foams which would benefit the invention as their reduced weight would allow smaller less powerful motors to be used and also add less weight to any system.

The edge contour of the optical blend mask where it intersects the light path may, due to the nature of the display layout, need to be a complex shape other than straight. The blend region may also vary in width. Characteristics of the projection pupil, such as variation with position or field of projection, may also result in a non-straight optical blend mask being required. The shape, length and cross-sectional size of the optical blend mask can be varied in order to achieve the desired blend. A complex shape for the optical blend mask may be required even if the blend region itself is straight.

The description and drawings only show two adjacent channels but it is obvious that the invention would also be suitable for multi-channel systems that could include overlaps on one or more sides of a projector channel. The invention could be used in multi-channel systems with overlaps on all sides or edges of a channel or channels.

List of Figures and Drawings Figure 1 shows a projector projecting an image onto a projection screen.

Figure 2 shows a projector rojecting an image onto a projection screen with * n optical soft edge mask.

Figure 3 shows the extent of the typical* blend region in a single channel of a multichannel projected display.

Figure 4 shows the typical blend region with anomalies.

Figure 5 shows two adjacent channels with a common overlapping blend region.

Figure 6 shows the desired optimum brightness across a two channel display with overlap.

Figure 7 shows the overlapping region in a two channel display, each channel exhibiting anomalies.

Figure shows the brightness across a two channel display with anomalies.

Figure 9 shows the plan view of a projector and screen with moving optical blend mask. Figure 10 shows projector and image on a screen with moving optical blend mask. Figure 11 shows the plan view of a projector and screen with a rotating blend mask. Figure 12 shows a rotating planar blend mask with a single vane.

Figure 13 shows a rotating planar blend mask with 2 crossing vanes.

Figure 14 shows an embodiment of the invention in which the blend mask moves in a first direction and in an orthogonal direction.

Figure 15 shows the locus described by the edge of the blend mask when moved equally in a first direction and an orthogonal direction.

Figure 16 shows the locus described by the edge of the blend mask when moved in a first direction and a second direction by unequal displacements Figure 17 shows another embodiment of the invention comprising a rotating mask with elliptical cross section.

Figure 18 shows the cross section through the rotating sk Figure 19 shows an embodiment of the rotating blend mask in plan view the comprising a rotation in cross section along its length.

Figure 20 shows an embodiment of the rotating blend mask which has an elliptical cross section and is tapered along its length.

Figures 21, 22, 23 and 24 show cross sections of alternative rotating blend masks.

Figure 25, 26 and 27 shows the interaction of the optical blend mask with light from a single exit point of a projector to demonstrate how the interaction varies as the mask rotates.

Figure 28 shows a cross section through a blend mask of the invention and the interaction with rays originating from more than one location as in known video projectors.

Figure 29 shows an example of a projection pupil of a typical commercially available video projector.

Figure 1 shows a display system 2 with a projector 4 projecting an image onto the screen 6. This display system has no overlap regions.

Figure 2 shows display system 10 with a projector 4 projecting an image onto screen 6. A known prior art blend mask 12 is located in the light path and produces a roll-off in brightness on the screen 6 indicated by the hatched region 16. 14 shows the width of the blend region required. Typically the brightness will fall off evenly towards the edge of the screen 6, The location of the blend mask determines the width and position of the blend region. The blend region obtained will vary in width depending on the distance that the optical blend mask is away from the projector pupil and also by the size of the projector pupil, which itself may vary through the projected field.

Figure 3 shows a display 18 with an image projected onto screen 6 with the desired blend region indicated by the hatched region 24. The desired width of the blend region is indicated by the arrow, 14. Ideally the brightness will roll-off from maximum brightness to zero at the left edge of the screen.

Figure 4 shows another display 28 having a screen 6 with a desired blend region indicated by the arrow 32. The blend region in this case shows a hatched region 34 but within this is another region 38, in this case exhibiting vertical lines of irregular intensity roll-off. The width of this region is shown by arrow 36 and is less than the width of the blend region. This type of artefact is typical of what is seen when prior art blend masks are used with video projectors having complex integration illumination optics. The artefacts shown here are not acceptable in a display system as they are distracting to the user.

Figure 5 shows a display system 40 with two adjacent channels the width of each shown by the arrows 46 and 48, There is shown an overlap region 44 indicated by the hatched area and the width of this is shown by the arrow, 42. In an optimised display system the left channel brightness would roll-off across the blend region and the right channel would roll-off in the opposite direction resulting in brightness across the blend region that is the same as the non-overlapped regions on either side. No artefacts are seen in the blend region.

Figure 6 shows graphically an intensity distribution 50, which indicates how the left and right channels' images in figure 5 are summed. 52 indicates the blend region where the two channels overlap, 54 is the desired brightness across the two overlapping channels, dotted line 58 shows the brightness roll-off for the left channel and 56 shows the brightness roll-off for the right channel and the summation of the two channels in the blend region gives the desired brightness 54 across the blend region.

Figure 7 shows a display system 60 with two adjacent channels the width of each shown by the arrows 66 and 68. There is shown an overlap or blend region indicated by the hatched area 64, the width of this region indicated by the arrow 62. In this example changes in brightness or artefacts are indicated by the vertical lines 70. These artefacts are typical and are seen when prior art optical blend masks are used with video projectors having complex integration illumination optics and are exaggerated in the blend region as artefacts are present from both the left channel and right channel and combine to create the artefacts seen. The artefacts may be variations in brightness or variations in colour or both.

Figure 8 shows graphically 74, how the left and right channels in figure 7 are summed. 76 indicates the blend region where the two channels overlap, 78 shows the achieved brightness level across the combined display. Dotted line 82 shows the brightness roll-off of the right hand channel in the blend region and dotted line 80 shows the brightness roll-off in the blend region for the left hand channel. 82 and 80 do not show a smooth roll-off for each channel and as a consequence the summed brightness shown by 78 within the blend region is also not smooth containing variations in brightness that are undesirable.

Figure 9 shows a display system 84 with a projector 4 projecting an image onto screen 6. The plan view shows a prior art optical blend mask 86 used to achieve the blend region indicated by the arrow 90. In this prior art example the blend mask is moved or oscillated in the direction shown by the arrow 88.

Figure 10 shows display system 94 with projector 4 projecting an image onto the screen 6. The blend region, 98, with a width indicated by the arrow 90, is achieved by the blend mask 86 being moved or oscillated in the direction 88 indicated. Known prior art discloses the mask moving in time or synchronous with the display or film frame rate. This type of moving blend mask does not achieve satisfactory results. If the display frame rate is 60Hz then flicker may be observed for example. Prior art methods typically moved the mask at the same frame rate as the film which results in artefacts being seen. Using other means to impart motion to the optical blend mask, voice coil for example, and a controller for controlling the movement, it is possible to impart motion at a frequency and magnitude that reduces visible artefacts.

Figure 11 shows an alternative prior art solution in which a rotating vane or blade type blend mask 102 is located in the light path of display system 100 between projector 4 and screen 6. The blend mask is rotated as indicated by the arrow 104.

Figure 12 shows prior art blend mask 102 that is rotated about an axis. This mask has a single vane or blade that interacts with the light path twice per revolution of the assembly Figure 13 shows prior art blend mask 106 having two vanes that interacts with the light path four times per revolution of the assembly. This is rotated about the axis shown.

Figure 14 shows display 108 comprising an optical blend mask 112 of the invention located between projector 4 and the screen 6. The optical blend mask 112 achieves the desired fall off in brightness in the region 110 at the screen with reduced artefacts as the mask is moved continuously in a first direction indicated by arrow 114 and continuously in a second orthogonal direction indicated by arrow 116. The edge of the optical blend mask is moved continuously into and out of the light path whilst simultaneously being moved closer and further away from the projector or exit pupil.

Figure 15 shows the optical mask 112 of the invention and the dotted line 118 describes the locus the edge of the optical blend mask describes when it is moved in a first direction indicated by arrow 122 whilst simultaneously being moved by an equal displacement in an orthogonal direction by arrow 120. Equal movement in the first and second directions means that the edge of the optical blend mask moves in a circle relative to the projector and light path.

Figure 16 shows the optical mask 112 of the invention, the dotted line 124 showing the locus of the edge of the optical blend mask, the locus being an ellipse, when the optical blend mask is moved in a first direction indicated by arrow 128 whilst being moved simultaneously in a second orthogonal direction indicated by arrow 126, the magnitude of the movement in each direction not being the same. In this example the movement into and out of the light path is greater than the movement towards and away from the projector or projector exit pupil.

Figure 17 shows an optical blend mask 110, of the invention. The optical blend mask has an eiliptical cross section and a length 132 with a consistent cross-sectional profile. The length of the blend mask is determined by analysis of the overall display system. Known means, such as a motor would be used to impart the rotational motion to the optical blend mask. Various types of motor would be suitable. Motors would ideally be chosen to facilitate the control required which could include constant or varying rotational speeds, and also achieving suitable rotational speeds such that other artefacts caused by the moving blend mask were not introduced into the display system. The motor and control means are not shown for clarity. The frequency of rotation would be chosen or determined by calculation such that the rotation itself does not introduce distracting artefacts into the visual display. The frequency of rotation may be less than the frame rate of the video being displayed, synchronized to the frame rate of the video being displayed, synchronized and equal to the frame rate of the video being displayed or greater than the frame rate of the video being displayed. The frequency of rotation would also take into account any temporal characteristics such as bit planes or combination of bit planes in a digital video projector that may give rise to distracting artefacts. For example bit plane characteristics of DLP projectors may introduce colour break-up artefacts and the frequency of rotation of the mask should* be chosen to reduce or eliminate these if they occur in the blend region. It will be shown that other optical blend masks may be used and the frequency of rotation may be varied and is dependent on the cross section of the optical blend mask. Optical blend masks that are essentially triangular, that is have three distinct corner lobes, do not need to be rotated at such high frequency as an optical blend mask with two distinct corner lobes such is the elliptical optical blend mask 132.

Figure 18 shows a cross-sectional view 134 of the optical blend mask 110 which is elliptical in cross-section.

Figure 19 shows an alternative profile of the blend mask that also has an elliptical cross-section. The figure shows a vertical view of the blend mask and the top of the mask is shown by 136. Along the length of the blend mask the profile is rotated through 90 degrees such that the bottom of the mask is shown by 138. The mask still has an edge that is continuous and consistent cross-section although rotated down the length. The rotation could be more or less than 90 degrees.

Figure 20 shows an alternative blend mask 120 of the invention. In this embodiment the cross-section is again elliptical but the biend mask is tapered aiong its length. The tapering can be consistent or vary along the length and is calculated to compensate for display system characteristics such as projection lens variation, off-axis projection or complex overlap or blend regions.

Figure 21 shows an alternative cross-section 144 of the blend mask of the invention. It is noted that as described the cross section shown in this figure and the following figures do not have any straight or concave edges or part straight or concave edges to the cross section. All edges are preferably convex as described.

Figure 22 shows an alternative cross-section 146 of the blend mask of the invention. Other cross-sections are available but 144 146, 148 and 150 have been shown as examples. The cross-section preferably always has the characteristic that the edge is continuously in the light path and this is achieved by having a continuous convex cross-section profile with no straight or concave portions.

Figure 23 shows the cross section 148 of an optical blend mask of the invention. The cross section is essentially triangular in profile but it is be noted that the edges as described above are convex with no straight or concave portions. As the number of distinct corners of the cross section increases the size of the optical blend mask also increases in order to achieve the same results. This is due to the displacement of the edge into and out of the light path being dependent on the number of distinct corners.

Figure 24 shows the cross section of an optical blend mask of the invention 150 having four distinct corners. Again the cross section edge profile has no straight or concave sections.

It is clear that other cross sections are suitable for use with increasing numbers of corners. The cross sections above have been shown as examples.

Figure 25 shows a cross-section 134 of the optical blend mask. A bundle of light rays is projected from a projector 156 and the optical blend mask 134 is at a location that intersects the bundle of rays. In this example a single projection point is used to demonstrate how the position at which the optical blend mask intersects the bundle of light rays varies continuously as the optical blend mask rotates. The edge of the optical blend mask intersects the bundle of light rays 152 at a position 154 that is at a distance 158 in a first direction from the origin of the bundle of light rays and is also at the position of the central ray of the light bundle. The following figures show how the location at which the edge of the optical blend mask intersects the bundle of light rays varies as the optical blend mask is rotated.

Figure 26 shows a cross-section of the optical blend mask 134 that has been rotated through approximately 45 degrees. The edge of the optical blend mask now intersects the bundle of light rays 152 at a point 160 that is now at a distance 164 from the origin of the bundle of rays and is also now not on the central ray but at a distance 162 away from the centra! ray. It can be seen that the point of intersection has moved closer to the projector whilst also moving out of the light path. The point of intersection is varying continuously in the two orthogonal directions as the optical blend mask rotates. This is further shown in Figure 27.

Figure 27 shows the cross section through the optical blend mask 134 now rotated through 90 degrees from the positon shown in figure 25. The edge of the optical blend mask 134 now intersects the bundle of light rays 152 originating from the projector 156, at a point 170. Point 170 is at a distance 174 from the origin of the bundle of light rays and at a distance 172 from the central ray of the bundle of light rays 152. The point of intersection has again moved in two directions.

Figure 28 shows the cross-section 134 through the optical blend mask. Video projectors with complex integrating illumination optics may exhibit multiple exit pupils or partial pupils which are shown in the figure by light paths 180, 182 and 184. The optical blend mask intersects all light paths from each of the pupils and as the optical blend mask is rotated the edge of the blend mask will be continuous intersecting all of the light paths without interruption.

Figure 29 shows a projection pupil 190, of a typical commercially available video projector. The circle, 192, shows the extent of the projection pupil. It can be seen that the pupil consists of a plurality of light patches, indicated by the rectangular shapes 194. The shape and distribution of the light patches, 194, will vary between projectors and will be dependent on the illumination system used within each projector. The illumination systems may be flyeye integrators or light pipes for example. Typically the light patches are not distinct individual shapes as shown in the figure but are also likely to overlap and exhibit colour variation along with brightness variation. This non-homogenous illumination exhibits itself as multiple projection pupils and gives rise to the multiple light paths shown in figure 28.

It is to be appreciated that*. the embodiments of the invention described above with reference to the accompanying drawings have been given by way of example only and that modifications may be effected. Thus, for example, modifications to the optical blend masks and alternative optical blend mask profiles or cross sections may be used other than those disclosed above along with various known means for rotating the optical blend masks. Alternative methods and materials to manufacture the optical blend masks may also be used. Another example is that the shape of the mask profile may allow for minor discontinuities in the intersection of the mask profile with the image light beam, perhaps by exhibiting small non-convex regions in the profile. Such sma*:l discontinuities may only have minor detrimental impact on the blend achieved, yet may, for example, assist in manufacturing options for the profile extrusion, such as use of standard extruded parts. Another modification to the method may include a simplification of determining the parameters of the mask where the projector pupil characteristics are assessed by simple observation of a projected blended edge and a trial mask having non-optimised parameters is tested, for example one having arbitrary estimated movement range or contour. The resultant performance of the trial mask is then re-assessed to attempt parameter changes to arrive at an improved trial mask and the modified method is iterated until satisfactory performance is obtained.