A Multiple-View Directional Display The present invention relates to a
multiple-view directional display, which displays two or more images such that each image is visible from a different direction. Thus, two observers who view the display from different directions will see different images to one another. Such a display may be used as, for example, an autostereoscopic display device or a dual view display device.
For many years conventional display devices have been designed to be viewed by multiple users simultaneously. The display properties of the display device are made such that viewers can see the same good image quality from different angles with respect to the display. This is effective in applications where many users require the same information from the display - such as, for example, displays of departure information at airports and railway stations. However, there are many applications where it would be desirable for individual users to be able to see different information from the same display. For example, in a motor car the driver may wish to view satellite navigation data while a passenger may wish to view a film. These conflicting needs could be satisfied by providing two separate display devices, but this would take up extra space and would increase the cost. Furthermore, if two separate displays were used in this example it would be possible for the driver to see the passenger's display if the driver moved his or her head, which would be distracting for the driver. As a further example, each player in a computer game for two or more players may wish to view the game from his or her own perspective. This is currently done by each player viewing the game on a separate display screen so that each player sees their own unique perspective on individual screens. However, providing a separate display screen for each player takes up a lot of space and is costly, and is not practical for portable games.
To solve these problems, multiple-view directional displays have been developed. One application of a multiple-view directional display is as a dual-view display', which can simultaneously display two or more different images, with each image being visible only in a specific direction - so an observer viewing the display device from one direction will see one image whereas an observer viewing the display device from another, different direction will see a different image. A display that can show different images to two or more users provides a considerable saving in space and cost compared with use of two or more separate displays.
Examples of possible applications of multiple-view directional display devices have been given above, but there are many other applications. For example, they may be used in aeroplanes where each passenger is provided with their own individual in-flight entertainment programmes. Currently each passenger is provided with an individual display device, typically in the back of the seat in the row in front. Using a multiple view directional display could provide considerable savings in cost, space and weight since it would be possible for one display to serve two or more passengers while still allowing each passenger to select their own choice of film.
A further advantage of a multiple-view directional display is the ability to preclude the users from seeing each other's views. This is desirable in applications requiring security such as banking or sales transactions, for example using an automatic teller machine (ATM), as well as in the above example of computer games.
A further application of a multiple view directional display is in producing a three- dimensional display. In normal vision, the two eyes of a human perceive views of the world from different perspectives, owing to their different location within the head.
These two perspectives are then used by the brain to assess the distance to the various objects in a scene. In order to build a display which will effectively display a three dimensional image, it is necessary to recreate this situation and supply a so-called "stereoscopic pair" of images, one image to each eye of the observer.
Three dimensional displays are classified into two types depending on the method used to supply the different views to the eyes. A stereoscopic display typically displays both images of a stereoscopic image pair over a wide viewing area. Each of the views is encoded, for instance by colour, polarisation state, or time of display. The user is required to wear a filter system of glasses that separate the views and let each eye see only the view that is intended for it.
An autostereoscopic display displays a right-eye view and a left-eye view in different directions, so that each view is visible only from respective defined regions of space.
The region of space in which an image is visible across the whole of the display active area is termed a "viewing window". If the observer is situated such that their left eye is in the viewing window for the left eye view of a stereoscopic pair and their right eye is in the viewing window for the right-eye image of the pair, then a correct view will be seen by each eye of the observer and a three-dimensional image will be perceived. An autostereoscopic display requires no viewing aids to be worn by the observer.
An autostereoscopic display is similar in principle to a dual-view display. However, the two images displayed on an autostereoscopic display are the left-eye and right-eye images of a stereoscopic image pair, and so are not independent from one another.
Furthermore, the two images are displayed so as to be visible to a single observer, with one image being visible to each eye of the observer.
For a flat panel autostereoscopic display, the formation of the viewing windows is typically due to a combination of the picture element (or "pixel") structure of the image display unit of the autostereoscopic display and an optical element, generically termed a parallax optic. An example of a parallax optic is a parallax barrier, which is a screen with transmissive regions, often in the form of slits, separated by opaque regions. This screen can be set in front of or behind a spatial light modulator (SLM) having a two- dimensional array of picture elements to produce an autostereoscopic display.
Figure 1 is a plan view of a conventional multiple view directional device, in this case an autostereoscopic display. The directional display 1 consists of a spatial light modulator (SLM) 4 that constitutes an image display device, and a parallax barrier 5.
The SLM of Figure 1 is in the form of a liquid crystal display (LCD) device having an active matrix thin film transistor (TFT) substrate 6, a counter-substrate 7, and a liquid crystal layer 8 disposed between the substrate and the counter substrate. The SLM is provided with addressing electrodes (not shown) which define a plurality of independentlyaddressable picture elements, and is also provided with alignment layers (not shown) for aligning the liquid crystal layer. Viewing angle enhancement films 9 and linear polarisers 10 are provided on the outer surface of each substrate 6, 7.
Illumination 11 is supplied from a backlight (not shown).
The parallax barrier 5 comprises a substrate 12 with a parallax barrier aperture array 13 formed on its surface adjacent the SLM 4. The aperture array comprises vertically extending (that is, extending into the plane of the paper in Figure 1) transparent apertures 15 separated by opaque portions 14. An anti-reflection (AR) coating 16 is formed on the opposite surface of the parallax barrier substrate 12 (which forms the output surface of the display 1).
The pixels of the SLM 4 are arranged in rows and columns with the columns extending into the plane of the paper in Figure 1. The pixel pitch (the distance from the centre of one pixel to the centre of an adjacent pixel) in the row or horizontal direction being p. The width of the verticallyextending transmissive slits 15 of the aperture array 13 is 2w and the horizontal pitch of the transmissive slits 15 is b. The plane of the barrier aperture array 13 is spaced from the plane of the liquid crystal layer 8 by a distance s.
In use, the display device I forms a left-eye image and a right-eye image, and an observer who positions their head such that their left and right eyes are coincident with the left-eye viewing window 2 and the right-eye viewing window 3 respectively will see a three-dimensional image. The left and right viewing windows 2,3 are formed in a window plane 17 at the desired viewing distance from the display. The window plane is spaced from the plane of the aperture array 13 by a distance r0. The windows 2,3 are contiguous in the window plane and have a pitch e corresponding to the average separation between the two eyes of a human. The half angle to the centre of each window 10, 11 from the normal axis to the display normal is c.
The pitch of the slits 15 in the parallax barrier 5 is chosen to be close to an integer multiple of the pixel pitch of the SLM 4 so that groups of columns of pixels are associated with a specific slit of the parallax barrier. Fig. 1 shows a display device in which two pixel columns of the SLM 4 are associated with each transmissive slit 15 of the parallax barrier.
Figure 2 shows the angular zones of light created from an SLM 4 and parallax barrier 5 where the parallax barrier has a pitch of an exact integer multiple of the pixel column pitch. In this case, the angular zones coming from different locations across the display panel surface intermix and a pure zone of view for image I or image 2 (where image 1' and image 2' denote the two images displayed by the SLM 4) does not exist. In order to address this, the pitch of the parallax barrier is preferably reduced slightly so that it is slightly less than an integer multiple of the pixel column pitch. As a result, the angular zones converge at a pre-defined plane (the "window plane") in front of the display.
This effect is illustrated in Figure 3 of the accompanying drawings, which shows the image zones created by an SLM 4 and a modified parallax barrier 5'. The viewing regions, when created in this way, are roughly kite-shaped in plan view.
Figure 4 is a plan view of another conventional multiple view directional display device 1'. This corresponds generally to the display device 1 of Figure 1, except that the parallax barrier 5 is placed behind the SLM 4, so that it is between the backlight and SLM 4. This device may have the advantages that the parallax barrier is less visible to an observer, and that the pixels of the display appear to be closer to the front of the device. Furthermore, although figures 1 and 4 each show a transmissive display device illuminated by a backlight, reflective devices that use ambient light (in bright conditions) are known. In the case of a transfiective device, the rear parallax barrier of Figure 4 will absorb none of the ambient lighting. This is an advantage if the display has a 2D mode that uses reflected light.
In the display devices of Figures 1 and 4, a parallax barrier is used as the parallax optic.
Other types of parallax optic are known. For example, lenticular lens arrays may be used to direct interlaced images in different directions, so as to form a stereoscopic image pair or to form two or more images, each seen in a different direction.
Holographic methods of image splitting are known, but in practice these methods suffer from viewing angle problems, pseudoscopic zones and a lack of easy control of the images.
Another type of parallax optic is a micropolariser display, which uses a polarised directional light source and patterned high precision micropolariser elements aligned with the pixels of the SLM. Such a display offers the potential for high window image quality, a compact device, and the ability to switch between a 2D display mode and a 3D display mode. The dominant requirement when using a micropolariser display as a parallax optic is the need to avoid parallax problems when the micropolariser elements are incorporated into the SLM.
Where a colour display is required, each pixel of the SLM 4 is generally given a filter associated with one of the three primary colours. By controlling groups of three pixels, each with a different colour filter, many visible colours may be produced. In an autostereoscopic display each of the stereoscopic image channels must contain sufficient of the colour filters for a balanced colour output. Many SLMs have the colour filters arranged in vertical columns, owing to ease of manufacture, so that all the pixels in a given column have the same colour filter associated with them. If a parallax optic is disposed on such an SLM with three pixel columns associated with each slit or lenslet of the parallax optic, then each viewing region will see pixels of one colour only.
Care must be taken with the colour filter layout to avoid this situation. Further details of suitable colour filter layouts are given in EP-A-0 752 610.
The function of the parallax optic in a directional display device such as those shown in Figures 1 and 4 is to restrict light transmitted through the pixels of the SLM 4 to certain output angles. This restriction defines the angle of view of each of the pixel columns behind a given element of the parallax optic (such as for example a transmissive slit).
The angular range of view of each pixel is determined by the pixel pitch p, the separation s between the plane of the pixels and the plane of the parallax optic, and the refractive index n of the material between the plane of the pixels and the plane of the parallax optic (which in the display of Figure 1 is the substrate 7). H Yamamoto et al. show, in "Optimum parameters and viewing areas of stereoscopic full-colour LED displays using parallax barrier", IEICE Trans. Electron., vol. E83-C, No. 10, p1632 (2000), that the angle of separation between images in an autostereoscopic display depends on the distance between the display pixels and the parallax barrier.
The half-angle a of Figure 1 or 4 is given by: sina = nsinlarctan(-&i) (1) 2s)) One problem with many existing multiple view directional displays is that the angular separation between the two images is too low. In principle, the angle 2a between viewing windows may be increased by increasing the pixel pitch p, decreasing the separation between the parallax optic and the pixels s, or by increasing the refractive index of the substrate n.
Co-pending UK patent application No. 0315171.9 describes a novel pixel structures for use with standard parallax barriers which provides a greater angular separation between the viewing windows of a multiple-view directional display. However, it would be desirable to be able to use a standard pixel structure in a multiple-view directional display.
Co-pending UK patent application Nos. 0306516.6 and 0315170.1 propose increasing the angle of separation between the viewing windows of a multiple-view directional display by increasing the effective pitch of the pixels.
JP-A-7 28 015 propose increasing the pixel pitch and therefore the angular separation between viewing windows of a multiple-view directional display by rotating the pixel configuration such that the colour sub pixels run horizontally rather than vertically.
This results in a threefold increase in pixel width and therefore roughly three times increase in viewing angle. This has the disadvantage that the pitch of the parallax barrier pitch must increase as the pixel pitch increases which, in turn, increases the visibility of the parallax barrier to an observer. The manufacture and driving of such a non- standard panel may not be cost effective. In addition there may be applications in which the increase in viewing angle needs to be greater than three times the standard configuration and in these cases simply rotating the pixels will not be sufficient. This is often the case with high resolution panels.
In general, however, the pixel pitch is typically defined by the required resolution specification of the display device and therefore cannot be changed.
It is not always practical or cost effective significantly to change the refractive index of the substrates, which are normally made of glass.
Other attempts at increasing the angular separation between the viewing windows of a multiple-view directional display device have attempted to reduce the separation between the parallax optic and the plane of the pixels of the SLM. However, this has been difficult as will be explained with respect to Figure 5, which is a schematic block view of the display device I of Figure I with an LCD as the SLM 4.
The LCD panel which forms the SLM 4 is made from two glass substrates. The substrate 6 carries TFT switching elements for addressing the pixels of the SLM, and is therefore known as a "TFT substrate". It will in general also carry other layers for, for example, aligning the liquid crystal layer 8 and allowing electrical switching of the liquid crystal layer. On the other substrate 7 (corresponding to the counter substrate of Figure I) colour filters 18 are formed, together with other layers for, for example, aligning the liquid crystal layer. The counter substrate 7 is therefore generally known as a "colour filter substrate" or CF substrate. The LCD panel is formed by placing the colour filter substrate opposite to the TFT substrate, and sandwiching the liquid crystal layer 8 between the two substrates. In previous directional displays the parallax optic has been adhered to the completed LCD panel as shown in figure 5. The distance between the LCD pixels and the parallax optic is determined primarily by the thickness of the CF substrate of the LCD. Reducing the thickness of the CF substrate will reduce the distance between the LCD pixels and the parallax optic, but will make the substrate correspondingly weaker. A realistic minimum for LC substrate thickness is about 0.5mm, but the pixel-to-parallax optic separation would still be too large for many applications if a parallax optic were adhered to a substrate of this thickness.
Japanese Patent No. 9-50 019 discloses a method for increasing the angular separation between the viewing windows of a multiple-view directional display device thereby to decrease viewing distance. This patent proposes reducing the thickness between the LC and barrier. This is done by constructing the stereoscopic LCD panel with the following order of components: LCD panel, parallax barrier, polariser. Previously the order had been: LCD panel, polariser, parallax barrier, as shown in Figure 1. This reduces the separation between the parallax barrier and the pixel plane by the thickness of the polariser, but this results in only a limited increase in the angular separation between the viewing windows of a multiple-view directional display device.
GB2405542 discloses a multiple view display, in which a parallax optic such as a parallax barrier is formed between the substrates of an image display device, such as a liquid crystal display. Figures 6 and 7 of the accompanying drawings illustrate two examples of the display disclosed in GB2405542.
The display device 52 comprises a first transparent substrate 6 and a second transparent substrate 7, with an image display layer 8 disposed between the first substrate 6 and the second substrate 7. An array of colour filters 18 is provided on the second substrate 7, and the second substrate will therefore be referred to as a colour filter substrate.
The first substrate 6 is provided with pixel electrodes (not shown) for defining an array of pixels in the image display layer 8, and is also provided with switching elements (not shown) such as thin film transistors (TFT5) for selectively addressing the pixel electrodes. The substrate 6 will be referred to as a TFT substrate'.
The colour filter substrate 7 comprises a base substrate 19 made of a light-transmissive material such as glass. A parallax barrier aperture array 13 is disposed on one principle surface of the base substrate 19 and 15 is formed by depositing opaque strips 14 on the surface of the base substrate, thereby defining transmissive slits between the opaque strips.
A spacer layer 20 formed of light-transmissive resin is provided over the parallax barrier aperture array 13. Finally, the colour filters 18 are disposed on the upper surface of the spacer layer 20.
A plurality of prisms 53 are provided on the external surface of the base substrate 19 of the colour filter substrate 7. The prisms 53 are shown as having a triangular cross- section. The prisms 53 work in combination with the parallax barrier 13 provided inside the display device. In use, the device is illuminated by a light provided behind the TFT substrate 6, so that the base substrate 19 of the colour filter substrate 7 forms the exit face of the display device. The prism structure varies the angle of separation between the left and right images induced by the parallax barrier.
Figure 6, the prisms are arranged so that they reduce the angle of separation between the viewing windows of different images.
The display 52' shown in Figure 7 corresponds generally to the display of Figure 6, except that the prisms 53 provided on the surface of the base substrate 19 are intended to increase the angle of separation between the two viewing windows.
H Yamamoto, S Muguruma, T Sato, K Ono, Y 1-layasaki, Nagai, Y Shimizu, N Nishida, Optimum parameters and viewing areas of stereoscopic full-colour LED displays using parallax barrier, IEICE Trans. Electron., vol. E83-C, No. 10, October 2000, p 1632 discloses the geometry involved in parallax barrier design and considers using parallax barriers for large screens such as those in stadiums.
Patent US5774262 discloses the use of a prism structure to create autostereoscopic 3D.
The system works by using collimated light. The prisms are aligned with the pixels.
The prisms may be used in conjunction with a phase mask which has random spatial fluctuations in phase shift. There is only one layer of prisms, so the image splitting angle will be small.
Figure 8 of the accompanying drawings illustrates diagrammatically a known type of dual view display, for example as disclosed in GB2405542. The display comprises a spatial light modulator, such as a liquid crystal display with pixels such as 60 and 61 disposed in front of or behind (as shown) a parallax barrier 63. The pixels 60, 61 are arranged as rows and columns and two images are displayed by the modulator such that alternate columns of pixels display vertical strips of the two views. This may be referred to as "spatial interlacing" or "spatial multiplexing". Each pair of columns of pixels 60, 61 is associated with a respective slit 64 of the barrier 65 so that the two images are visible in respective viewing regions.
As illustrated in Figure 8, such a display provides relatively low brightness, particularly when used in the reflective mode. For example, the slits 64 of the barrier 63 typically occupy about 25% of the barrier area so that, of the light incident on the display from a light source 65, approximately 25% passes through the barrier 63 whereas approximately 75% is blocked by the barrier 63.
In the case of a reflective display, the spatial light modulator contains a reflective structure such that light modulated by the pixels 60, 61 is reflected diffusely. Thus, reflected light from the pixels 60, 61 is incident not only on the slits 64 but also on the opaque regions of the barrier 63 so that, again, approximately 75% of the reflected light is lost and does not contribute to the brightness of the displayed images. The presence of the parallax barrier 63 may therefore result in a reduction in display brightness of the order of 95%.
According to first to twelfth aspects of the invention, there are provided apparatuses as defined in the appended claims 1, 8, 15, 27, 31, 35, 42, 49, 54, 58, 59 and 64, respectively.
Embodiments of the invention are defined in the other appended claims.
The invention will be further described, by way of example, with reference to the accompanying drawings, in which: Figures 1 to 8 are diagrams illustrating known types of multiple view displays; and Figures 9 to 24 are diagrams illustrating various embodiments of the invention.
Although the displays described hereinafter are of the dual view type, the same techniques may readily be applied to displays for displaying more than two views, as is readily apparent to a person of ordinary technical skill in this technological field. Such displays may be used to display views of independently selectable image content for viewing by different viewers so that the contents of the different views may be related or unrelated according to the desired application. At least some of the displays may also be used to provide a 3D display for one or more viewers where the views are stereoscopically related and provide binocular disparity.
Figure 9 illustrates a reflective or transfiective (transmissive/reflective) display of a type similar to that illustrated in Figure 8. The display comprises a reflective or transfiective spatial light modulator including pixels such as 60 and 61 and a parallax barrier 63.
The pixels 60, 61 contain reflective micro-structures such as 66 having specularly reflective surfaces. The structures 66 are disposed at the rear of the pixels so that incident light on the reflective surfaces passes through the light-modulating layer of the pixel before being reflected back through this layer. The reflective surfaces are disposed and oriented so that light passing through the barrier slits 64 and incident on the surfaces is reflected back towards and through the slits 64.
In a typical example of such a display where the slits 64 occupy approximately 25% of the parallax barrier area, approximately 75% of the ambient light incident of the display is blocked by the barrier 63. However, a high proportion of the light passing through the slits 64 and modulated by the pixels if reflected back through the slits 64 so that the image brightness provided by the display shown in Figure 9 is substantially greater than that provided by the display shown in Figure 8. For example, an increase in brightness of up to four times may be achieved by such a display.
Figure 10 shows a design that may be applied to a large dual view display such as that seen in a football stadium. These displays have big pixels (for example a large LED array 70), with LEDs which may be 1cm in diameter. This means the light directing structures such as baffles 71 are feasible to manufacture as shown in the diagram. The light directing structures 71 may be a similar size to the pixels (e.g. 1 cm tall), so that they could easily be made by injection moulding plastic (for example) , and easily attached to the LED display. The light directing elements shown are absorbing.
Similarly, similar structures may be produced on a microscopic scale inside an LCD panel.
Figures 11 and 12 show an arrangement for a rear projection screen that creates dual view. It is a time sequential arrangement. In time frame 1, S polarised light is used. S polarised light is focused by s polarisation sensitive lenses 75 to particular spots on a diffuser screen 76. Lenticular lenses 77 direct light from these spots to the left viewer.
In time frame 2, p polarised light is used. p polarised light is focused by p polarisation sensitive lenses 78 to different spots on the diffuser screen 76. The lenticular lenses direct light from these different spots to the right viewer.
Figures 13 and 14 show a different arrangement for a rear projection screen that creates dual view. It is a time sequential device. In time frame 1, S polarised light is used. S polarised light is focused by lenses 79 to particular spots on the diffuser screen 76. The lenticular lenses 77 direct light from these spots to the left viewer.
In time frame 2, p polarised light is used. p polarised light is focused by the lenses 79, but then a p sensitive prism film 80 redirects the lightto different spots on the diffuser screen 76. The lenticular lenses 77 direct light from these different spots to the right viewer.
Another version of the device may be used for a projector seen in a cinema (for example). One image is projected in a first polarisation and the second image in another polarisation. The viewer can choose which image to watch, by switching between glasses that allow just one image to pass.
Although three embodiments have been described as operating timesequentially, they may operate non-time-sequentially, for example such that both views, encoded in orthogonal polarisations, are projected simultaneously.
Figure 15 illustrates a dual view display which makes use of a collimated backlight 81 and internal prisms 82 arranged as first and second arrays on either side of a layer of liquid crystal display pixels 83 and between LCD substrates 84 and 85. The two layers of prisms 33 increase the angular separation at the display between the two viewing regions produced by the display. Thus, for example, a display of the type shown in Figure 15 may be used in the dashboard of a vehicle to permit a driver and a passenger to see different images.
Such an arrangement is capable of providing brighter image display than arrangements using parallax barriers. For example, an increase in brightness of up to four times may be achieved. Further, it is not necessary to make use of thin glass substrates or separators, for example of the order of 50 micrometres in thickness, so that the problems of handling such thin glass during manufacture are avoided.
Figure 16 illustrates a collimated backlight 81 which may, for example, be used as the backlight with the display shown in Figure 15. The backlight makes use of a backlight 22 of the type disclosed in British patent application No. 0510192.8.
The backlight 22 comprises a waveguide 26 that is illuminated by one or more light sources 27. Two light sources are shown, arranged against opposite side edge faces of the waveguide 26, but the backlight may have only one light source or may have more than two light sources. Preferably, the or each light source extends along substantially the entire length of the respective side edge face of the waveguide 26, and may be, for example, a fluorescent tube.
The light from the or each light source 27 enters the waveguide 26 of the backlight. The "exit face" of the waveguide has regions ("TIR regions") that are totally internally reflecting for light propagating within the waveguide, and has regions ("non-TIR regions") that are not totally internally reflective for light propagating within the waveguide. When light propagating within the waveguide is incident on a region of the exit face which is not totally internally reflecting, light is refracted out of the waveguide through the exit face.
Any light emitted from the rear face of the waveguide 26 would be wasted. The rear face of the waveguide 26 is therefore preferably totally internally reflecting for light propagating within the waveguide.
The regions of the exit face of the waveguide that are not totally internally reflecting are arranged to have a pitch p1 which is equal to the pitch of the pixels of the associated SLM or which is an integer multiple of the pitch p2 of the pixels of the SLM. That is, p1 = n p2 where n = 1, 2, 3... Where this embodiment is applied to a dual view display, the pitch p1 of the non-TIR regions of the exit face of the waveguide is preferably twice the pixel pitch p2, so that one non-TIR region is provided for every pair of a "left pixel" column and a "right pixel" column, where a "left (right) pixel" column is a column of pixels that displays the image intended for display to the left (right) observer. (In principle however, one non-TIR region could be provided for, for example, every set of m left pixel columns and m right pixel columns, where m = 1.) The centre of a non-TIR region is preferably laterally aligned with the non-display portion between a left pixel column and a right pixel column. Where this embodiment is applied to a conventional 2-D display, however, the pitch p1 of the non-TIR regions of the exit face of the waveguide is preferably equal to the pixel pitch p2, so that one non-TIR region is provided for every pixel column.
The regions of the exit face of the waveguide that are not totally internally reflecting (these regions will be referred to as "non-TIR regions") are obtained by disposing a plurality of protrusions 28 over the exit face of the waveguide 26. Each protrusion 28 has a refractive index that is not lower than the refractive index of the waveguide 26, so that total internal reflection does not occur at the boundary between the waveguide 26 and the protrusion 28. The refractive index of the protrusions 28 may be the same as the refractive index of the waveguide, or it may be greater than the refractive index of the waveguide 26.
Regions of the exit face of the waveguide 26 between neighbouring protrusions 28 are preferably coated with a material 29 having a refractive index lower than the refractive index of the waveguide 26. This ensures that light is totally internally reflected while the protrusions 28 are not provided. As an example, if the waveguide 26 has a refractive index of approximately 1.5, regions where the protrusion 28 are not present are preferably coated with a material 29 having a refractive index that is below 1.4.
The backlight shown in Figure 16 further comprises a substrate 86 on which are formed a parallax barrier 87 and an array of lenses 88. The lenses 88 comprise a lenticular array of cylindrical converging lenses which are substantially at the slits of the barrier 87 and are aligned with the non-TIR regions and protrusions 28.
Light from the portion 22 of the backlight has its angular distribution controlled so that substantially all of the light is incident on the lenses 88. The lenses 88 are arranged to improve the collimation of light from the backlight and, together with the barrier 87, effectively define a collimated parallel-striped light source for a spatial light modulator of a dual view display. Thus, such a backlight is suitable for use with the display shown in Figure 15.
The barrier 87 may be in the form of a switchable liquid crystal layer having a barrier mode as illustrated and a non-barrier mode, in which the opaque barrier portions between the slits are transmissive. Such an arrangement allows the backlight 81 to be switched between a highly collimated mode and a less collimated mode to provide viewing angle control. For example, when used in the display shown in Figure 15, this may allow the display to be switched between dual view and single view wider viewing angle modes.
In order to reduce scattered light, it may be beneficial to extend the light-blocking parts of the barrier 87 a little way up the "sides" of the lenses 88.
Figure 17 illustrates another dual view display using a liquid crystal device having a layer of pixels 83 and a collimated backlight 81, for example of the type shown in Figure 16. In this display, control of viewing angles and creation of viewing regions is provided by a mirror arrangement. In particular, light passing through each pixel 83 is incident on a curved mirror 89 (thus having "optical power"), which deflects the modulated light on to another curved mirror 90. The mirrors 89 and 90 are disposed internally of the LCD substrates and arranged to direct light to the respective viewing regions. The mirrors 89 and 90 are arranged to provide viewpoint correction, for example by choosing their positions with respect to the pixels 83, their orientations and/or their shapes appropriately. The mirrors 89 may also be disposed and arranged so as to improve the blocking of light so as to reduce crosstalk.
Figure 18 illustrates a known type of display in which the pixels 83 are of "asymmetric" shape and the parallax barrier has relatively narrow slits 64 of constant width. The lower part of Figure 18 illustrates how the image brightness varies with the lateral position of the head of a viewer. Because of the varying height of each pixel 83, the brightness is non-uniform with viewing angle and varies perceptibly and undesirably within each viewing region.
Figure 19 illustrates a display which reduce the dependence of brightness on viewing angle for pixels 83 of the same shape. The parallax barrier 63 comprises slits having regions 64a and 64b of different widths with the slit width variation corresponding to the pixel aperture variation. The barrier slits may be formed as continuous "vertical" slits with the slit width variation having the same pitch as the vertical pixel pitch. The graph in the lower part of Figure 19 illustrates the improvement in uniformity of brightness with lateral head position. Also, it is possible to increase the display brightness.
Figure 20 illustrates a pixel and barrier design which may be used to create a black viewing angle range, for example so as to prevent unwanted reflection from a car windscreen 91. The pixels 83 have a relatively small vertical aperture and each barrier slit 64 is of reduced length, for example created by horizontal light-blocking lines 92 having the same pitch as the vertical pixel pitch. The barrier 63 is aligned so as to block light in an angular region 93 and tends to prevent light from being reflected from the windscreen 91 to a driver so as to reduce or eliminate potentially distracting reflections.
Figure 20 illustrates part of a multiple view display disclosed in GB2405542. A barrier substrate 100 has formed thereon a lenticular lens array 101 of high refractive index. A black mask 102 is formed between the lenses of the array 101. This assembly is attached to a colour filter substrate 103 by means of a layer 104 of glue of low refractive index.
An example of a known technique for manufacturing the lens array 101 is "resist reflow". Blocks of resist are created where the lenses are required and the resist is melted such that it flows into a cylindrical lens shape. This has the disadvantage that resists which reflow tend to absorb too much light.
Clear resists exist but they are too thin to create lenses. Negative resists exist which are clear and thick but processing the resist into blocks by lithography involves hardening the resist. After hardening, negative resists do not reflow when heated.
A technique for creating lenses which is suitable for negative (and positive) resists is to print the resist on to the glass substrate 100 in blocks, for example by screen printing.
The resist is then reflowed by melting and then hardened for durability. This process is suitable for negative resists because no hardening is required to create the resist block.
Lenses may be difficult to make whereas prisms are easier to make although their image splitting qualities are not as good as lenses. In applications where the performance of prisms is acceptable, such prisms may be made by processing a clear negative resist by means of lithography. The processing is performed so as to create structures as close to lenses as possible or as necessary for the application by altering the lithography processing conditions. Figure 22 illustrates a typical structure resulting from such a technique. In this case, the resulting structures 105 are of prism shape.
Figure 23 illustrates the result of extending this technique to form the black mask as an array of light-absorbing prisms 106. The black mask prisms 106 may be formed by any suitable process and the light-absorbing prisms at least partially block undesirable secondary viewing windows or regions which might otherwise be formed. As shown in Figure 24, the black mask prisms 106 may be made sufficiently deep to space the barrier from the substrate 103, for example where the array 101 is a lenticular lens array.