GB2404106A

GB2404106A - Generating a test image for use in assessing display crosstalk.

Info

Publication number: GB2404106A
Application number: GB0316605A
Authority: GB
Inventors: Graham Roger Jones
Original assignee: Sharp Corp
Current assignee: Sharp Corp
Priority date: 2003-07-16
Filing date: 2003-07-16
Publication date: 2005-01-19
Also published as: GB0316605D0

Abstract

A technique is provided for generating a test image for use in assessing crosstalk in a multiple view display, such as a three dimensional autostereoscopic display. A test pattern is displayed by a display device of the display. A first intensity A% is displayed in a first view at a first display region and, for example, comprises 100% intensity. A second intensity B% is displayed in a second view at the first region and may, for example, be 0% intensity. One or more third intensities are displayed in the second view at a second display region of the display device. The or each third intensity is of the form (B + (A Ò X)/100)%, where X is a crosstalk value. A viewer may then compare the appearances of the first and second regions in the second view so as to assess the crosstalk generated in the display.

Description

2404 1 06

GENERATING A TEST IMAGE FOR USE IN ASSESSING DISPLAY

CROSSTALK

The present invention relates to generating a test image for use in assessing crosstalk in a multiple view display. Such techniques may be used with directional displays, such as two view three dimensional (3D) autostereoscopic displays, and dual view displays for showing different two dimensional (2D) images to one or more viewers depending on the viewing angle relative to the display. Such techniques may also be used with non directional displays, such as stereoscopic displays using anaglyph glasses, active shutter glasses and polarization sensitive glasses. For example, such techniques may be used to allow users to assess the quality of a display or for assisting in quality control on a production line.

Multi-view directional displays are well known and allow different views or images to be seen by a viewer in different viewing regions or zones. Thus, the image which is visible to an eye of a viewer depends on the angle between the display and the eye. In the case of 3D autostereoscopic displays, two or more views are visible in corresponding viewing zones such that a left eye view can be made visible in a left eye viewing zone and a right eye view can be made visible in a right eye viewing zone.

Provided the viewer is placed with the left and right eyes in the left and right viewing zones, a three dimensional effect can be seen.

In normal vision, the two human eyes perceive views of the world from different perspectives due to their separate location within the head. These two perspectives are then used by the brain to assess the distance to 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 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. Stereoscopic displays typically display both of the images over a wide viewing area. However, each of the views is encoded, for instance by colour, polarization state or time of display, so that a filter system of glasses worn by the observer can separate the views and will only let each eye see the view that is intended for it.

Autostereoscopic displays require no viewing aids to be worn by the observer but the two views are only visible from 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 region". If the observer is situated such that one of their eyes is in one viewing region and the other eye is in the viewing region for the other image of the pair, then a correct set of views will be seen and a three-dimensional image will be perceived.

For flat panel autostereoscopic displays, the formation of the viewing regions is typically due to a combination of the pixel structure of the display unit and an optical element, generically termed a parallax optic. An example of such an optic is a parallax barrier. This element is a screen with vertical transmissive slits separated by opaque regions. This screen can be set in front of a spatial light modulator (SLM) with a twodimensional array of pixel apertures as shown in Figure 1 of the accompanying drawings.

The display comprises a transmissive spatial light modulator in the form of a liquid crystal device (LCD) comprising an active matrix thin film transistor (TFT) substrate 1, a counter substrate 2, a pixel (picture element) plane 3 formed by a liquid crystal layer, polarisers 4 and viewing angle enhancement films 5. The SLM is illuminated by a backlight (not shown) with illumination in the direction indicated by an arrow 6. The display is of the front parallax barrier type and comprises a parallax barrier having a substrate 7, an aperture array 8 and an anti- reflection (AR) coating 9.

The SLM is arranged such that columns of pixels are provided extending vertically for normal viewing with the columns having a horizontal pitch p. The parallax barrier provides an array 8 of apertures or slits with the slits being parallel to each other and extending parallel to the pixel columns. The slits have a width 2w and a horizontal pitch b and are spaced from the pixel plane 3 by a separation s.

The display has an intended viewing distance rO with left and right viewing windows 10 and 11 at the widest parts of the viewing regions defining a window plane 12. The viewing windows 10 and 11 have a pitch e which is generally made substantially equal to the typical or average human eye separation. The centre of each primary viewing window lo, 11 subtends a half angle a to the display normal.

The pitch b of the slits in the parallax barrier is chosen to be close to an integer multiple of the pixel pitch p of the SLM so that groups of columns of pixels are associated with a specific slit of the parallax barrier. Figure 1 shows an SLM in which two pixel columns are associated with each slit of the parallax barrier.

Figure 2 of the accompanying drawings shows the angular zones of light created from an SLM and parallax barrier where the parallax barrier has a pitch b of an exact integer multiple of the pixel column pitch p. In this case, the angular zones coming from lS different locations across the display panel surface intermix and a pure zone of view for image l or image 2 does not exist. In order to address this, the pitch b of the parallax optic is reduced slightly so that the angular zones converge at the window plane 12 in front of the display. This change in the parallax optic pitch is termed "viewpoint correction" and is shown in Figure 3 of the accompanying drawings. The viewing regions created in this way are roughly kite shaped.

For a colour display, each pixel is generally provided with a filter associated with one of the three primary colours. By controlling groups of three pixels each with a different colour filter, substantially all 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, due 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 used with such an SLM such that three pixel columns are associated with each slit (or lenslet), only one colour will be visible in each viewing region. This may be avoided using, for example, the techniques disclosed in EP 0 752 610.

In an ideal display of this type, the left eye image would be visible only in the left viewing zone and would be completely invisible, in particular, in the right viewing zone, and vice versa. However, in practical displays, "crosstalk" occurs between views such that some of the light from the left eye image is propagated to the right viewing zone and vice versa. Crosstalk may be measured as a ratio of the left eye image seen by the right eye and vice versa. For crosstalk not to be apparent or result in uncomfortable viewing conditions, it must be at a very low level. For higher levels of crosstalk, it is possible to provide crosstalk compensation such that display contrast is sacrificed for providing acceptable crosstalk performance. A known technique for providing such crosstalk correction is disclosed in EP 0 953 962 and requires that the crosstalk level be measured in order to provide the appropriate correction or compensation.

British patent application No. 0216608.0 discloses an arrangement for use with transflective displays to allow a viewer to assess the amount of 2D reflected light compared with 3D transmitted light. Reflected light only contributes to crosstalk during the transmissive mode of operation of such a display so that this arrangement allows a viewer to determine the level of crosstalk caused by ambient reflected light. The resulting assessment may be used to compensate for crosstalk as described above.

However, this arrangement does not allow the assessment of inherent display crosstalk from other causes, such as a misaligned parallax optic. Further, this arrangement does not allow assessment of crosstalk for purely transmissive displays.

Various techniques are known for measuring crosstalk in such displays but all such known techniques rely on measurements made with the appropriate instruments and are therefore relatively expensive, time-consuming and inappropriate for consumer use. For example, Mirishama, Nose, Taniguchi, Inoguchi, Matsumura, "An Eyeglass- Free-Rear- Cross-Lenticular 3D Display", 1998 Sid International Symposium, Digest of Technical Papers, vol. 29, p923-6, May 1988 discloses a technique for measuring crosstalk in an autostereoscopic display based on measuring an intensity profile along the horizontal centre line of the viewing plane of the display with the left eye image set black (minimum intensity) and the right eye image set while (maximum intensity). This technique requires the use of optical detectors. s

Similarly, Hanazota, Okui, Yumama, "Subjective evaluation of crosstalk disturbance in stereoscopic displays", Conference Record of the 20th International Display Research Conference, p288-91, 25-28 September 2000 compares subjective quality assessments of 3D displays based on measured crosstalk values. In this case, the crosstalk is measured using a photometer. Kuo-Chung Huant, Chao-Hsu Tsai, Kuen Lee, Wen-Jean Hsuch, "Measurement of contrast ratios for 3D display", Proc. SPE - Int. Soc. Opt.

Eng. (USA) vol. 4080, p78-86, 26-27 July 2000 discloses measurements of crosstalk on different types of stereoscopic displays with the crosstalk being measured using a luminance meter. Tan, Stanley S.L, Woods, Andrew J. , "Characterising sources of ghosting in time-sequential stereoscopic video displays", Proceedings of SPIE - The International Society for Optical Engineering, vol. 4660, 2002, p66-77 discloses crosstalk assessment for a stereoscopic display using liquid crystal shutter glasses and a cathode ray tube (CRT) display. Total crosstalk of the system is deduced by optical measuring equipment with measurements being made for red, green and blue images.

According to a first aspect of the invention, there is provided a method of generating a test image for use in assessing crosstalk in a multiple view display, comprising the steps of: (a) selecting a first intensity having a value of A% of maximum intensity for display in a first view at a first display region of the display; (b) selecting a second intensity having a value of B% of maximum intensity for display in a second view at the first region; and (c) selecting at least one third intensity having a value of (B+(A.X)/100) % of maximum intensity, where X is a crosstalk value in per cent, for display in the second view at a second display region of the display.

The first intensity may have a value A of 100% of maximum intensity. In the case of a pixellated display, those pixels (picture elements) which are displaying the first intensity are set to their maximally transmissive, reflective or emissive states, for example corresponding to "white".

The second intensity may have a value of 0% of maximum intensity. For example, in the case of a pixellated display, those pixels required to provide the minimum intensity may be set to their minimally or zero reflective, transmissive or emissive states corresponding to "black".

The step (c) may comprise varying the at least one third intensity by varying the crosstalk value X. The step (c) may comprise selecting a plurality of different third intensities having different crosstalk values X for display in respective sub-regions of the second region.

The method may comprise displaying adjacent each of the sub-regions a text indication of the crosstalk value X corresponding to the third intensity of the respective sub- region.

The method may comprise adjusting the colour of the at least one third intensity to match the colour in the second view at the first region.

The method may comprise performing the steps (a) to (c) for at least one of a plurality of colours. The method may comprise performing the steps (a) to (c) for each of the colours. The display may be an ROB display and the colours may comprise red, green and blue.

The method may comprise selecting a fourth intensity, which is less than or equal to the third intensity, for display in a first view at the second region. The fourth intensity may be substantially equal to the second intensity. As an alternative, the fourth intensity may be substantially equal to the third intensity. The first and second regions may at least partially overlap each other and the method may comprise switching the display between a multiple view mode, in which the first and second intensities are displayed, and a single view mode in which the third and fourth intensities are displayed. As an alternative, the first region may provide multiple views and the second region may provide a single view. As a further alternative, the first and second regions may be provided by different display devices.

The first region may comprise a viewer position indicating region.

The method may comprise the further step of comparing the intensities displayed in the second view at the first and second regions. The comparing step may comprise visually comparing the appearances of the second view at the first and second regions.

The display may be a directional display. The intensities may be assessed in a viewing region of the second view.

According to a second aspect of the invention, there is provided a method of generating a test image for use in assessing crosstalk in a multiple view display, comprising the steps of: (a) selecting a first colour, which has N components having values of A'%, À .., AN%, of maximum intensity respectively, for display in a first view at a first display region of the display, where N is an integer greater than 1; (b) selecting a second colour, which has N components having values of Be%, ...., BN% of maximum intensity respectively, for display in a second view at the first region; and (c) selecting at least one third colour, which has N components having values of (B+(A.X)/100)%, , (BN+(AN.X)/100)% of maximum intensity respectively, where X is a crosstalk value in percent, for display in the second view at a second display region of the display.

The method may comprise displaying or printing the test image.

The method may comprise storing or transmitting data representing the test image.

N may be equal to 3.

The value of Al may be 100% of maximum intensity.

The values of A2 and A3 may be 0% of maximum intensity. s

The value of B', B2 and B3 may be substantially equal to each other.

The values of B2 and B3 may be substantially equal to half maximum intensity.

The step (c) may comprise selecting a plurality of different third colours having different values of (B'+(A.X)/100)% of maximum intensity for display in respective sub-regions of the second region.

The method may comprise the further step of comparing the colours displayed in the second view at the first and second regions. The comparing step may comprise visually comparing the appearances of the second view at the first and second regions.

The display may be a directional display. The colours may be assessed in a viewing region of the second view.

The display may be a two view display.

The display may be a three-dimensional autostereoscopic display.

According to a third aspect of the invention; there is provided a test image generator for generating a test image for use in assessing crosstalk in a multiple view display, comprising: means for generating first image data, representing a first intensity having a value of A% of maximum intensity, for display in a first view at a first display region; means for generating second image data, representing a second intensity having a value of B% of maximum intensity, for display in a second view at the first region; and means for generating third image data, representing at least one third intensity having a value (B + (A - X)/100) % of maximum intensity where X is a crosstalk value in per cent, for display in the second view at a second display region.

According to a fourth aspect of the invention; there is provided a test image generator for generating a test image for use in assessing crosstalk in a multiple view display, comprising: means for generating first image data, representing a first colour, which has N components having values of Al%, ...., AN% of maximum intensity respectively, for display in a first view at a first display region where N is an integer greater than 1; means for generating second image data, representing a second colour, which has N components having values of Be%, ...., BN% of maximum intensity respectively, for display in a second view at the first region; and means for generating third image data, representing at least one third colour, which has N components having values of (B+(A'.X)/100)%, , (BN+(AN.X)/100)%, of maximum intensity respectively, where X is a crosstalk value in percent, for display in the second view at a second display region.

According to a fifth aspect of the invention, there is provided a method of generating a test image for use in assessing crosstalk in a 3D display, comprising applying crosstalk correction to at least one stereoscopic pair of images and causing the display to display the or each corrected pair of images.

The amount of crosstalk correction may be selectable.

The method may comprise applying different amounts of crosstalk correction to a stereoscopic pair of images to create a plurality of corrected image pairs and causing the display to display the corrected image pairs.

The method may comprise selecting a value of crosstalk correction corresponding to a most natural 3D image appearance.

The crosstalk correction may be applied by: adding a grey level to a first image to form a first sum; adding said grey level to a second image to form a second sum; subtracting from the first sum an amount equal to a given fraction of said second image; and subtracting from said second sum an amount equal to the given fraction of said first image, wherein these steps comprise the calculation of a partial result which is used to determine crosstalk corrected picture elements for both the first and second images.

Each picture element may comprise M colour components having an intensity value, and the method may further comprise, for an intensity level Ix of each of the picture elements of said first image, determining a crosstalk corrected picture intensity level Iox according to: K(l -I -I -1) {JX (Im + 1) or an equivalent form thereof and for an intensity level Iy of each of the picture elements of said second image, determining a crosstalk corrected picture intensity level Ioy according to: K(l -I -I - 1) (Im + 1) or an equivalent form thereof, where K is the scalar crosstalk correction; and Im is the scalar maximum value of each of colour component.

The partial search result used to determine crosstalk corrected picture elements may be: K(lm-lX-Iy-1) According to a sixth aspect of the invention, there is provided an apparatus for generating a test image for use in assessing crosstalk in a 3D display, comprising means for applying crosstalk correction to at least one stereoscopic pair of images and means for causing the display to display the or each corrected pair of images.

The method may comprise determining from the comparison a display crosstalk value.

The method may comprise determining from the or each displayed corrected pair of images a display crosstalk value.

The method may comprise recording the display crosstalk value.

According to a seventh aspect of the invention, there is provided a display whose crosstalk is corrected in accordance with the display crosstalk value.

An image generated by the methods and apparatuses defined hereinbefore may be supplied by one entity to another entity. Accordingly, a further aspect of the invention provides use of such an image, comprising displaying the image. Such use may comprise assessing display crosstalk from the displayed image. Such use may comprise recording the display crosstalk.

A further aspect of the invention provides a display whose crosstalk is corrected in accordance with the display crosstalk recorded by such use.

All such uses and such a display are within the scope of the invention, irrespective of whether the same entity or different entities are involved in different aspects of the invention.

It is thus possible to provide a technique which allows a viewer or user of a display to assess the display crosstalk using only a test image and without requiring any additional equipment, although measuring equipment may be used if necessary or desirable. In embodiments where the user can determine the display crosstalk level, crosstalk correction can be applied so as to compensate for the prevailing level of display crosstalk. In cases where crosstalk varies over time, the user can periodically "calibrate" the display and crosstalk compensation so as to allow better image quality to be achieved. No optical measurement systems are required and no mechanical systems for moving a detector are required. This technique may therefore be implemented with very low cost, is quick and easy to set up, and allows crosstalk to be identified very quickly. Such a technique allows the advantages to be achieved when used for performing quality control tests or display calibration, for example on a production line.

The invention will be further described, by way of example, with reference to the accompanying drawings, in which; Figure 1 is a diagrammatic horizontal cross-sectional view of a known type of multiple view display; Figure 2 is a diagrammatic plan view illustrating angular viewing regions created by a non-viewpoint corrected display of the type shown in Figure 1; Figure 3 is a diagrammatic plan view illustrating viewpoint correction in the display of Figure 1; Figure 4 is a block schematic diagram of an apparatus for performing methods constituting embodiments of the invention; Figure 5 is a diagram illustrating a viewing arrangement for assisting in performing a method during display manufacture; Figures 6 to 14 are diagrams illustrating patterns displayed on a display device of two view directional displays when performing methods constituting embodiments of the invention; and Figure 15 is a schematic block diagram illustrating crosstalk correction.

Like reference numerals refer to like parts throughout the drawings.

Figure 4 illustrates an arrangement which may be used to generate test images and to assess display crosstalk. This arrangement is used with a display 20 of the multiple view type, for example as illustrated in Figure 1, for displaying stereoscopically related or unrelated views.

The arrangement comprises a controller 21 which generates a test image in the appropriate format and supplies this to the display 20. The controller 21 may be hardwired or may be of the programmed type, such as a computer, in which case storage 22 is provided containing software for generating the test image. In order to provide on-screen instructions of how to assess crosstalk, storage 23 may be provided for the appropriate instructions.

The arrangement may have one or more user-operable controls and two such controls are illustrated at 24 and 25. These controls may be of a dedicated type specifically for use in assessing crosstalk of the display 20 or may be conventional inputting devices, such as a keyboard and a mouse. The control 24 is operable by a user to adjust the crosstalk intensity or reference intensity of the test image as described hereinafter. The control 25 is operable by the user to adjust the colour of the crosstalk or reference intensities as described hereinafter.

In order to help a user to view the display correctly when assessing crosstalk, an arrangement illustrated diagrammatically at 26 and shown in more detail in Figure 5 may be used. For example, such an arrangement may be used on a production line during manufacture of the display 20 so as to ensure that a user is correctly positioned in order to make a rapid and reliable assessment of display crosstalk. The arrangement comprises an aperture in a mask 27 which is spaced by the display viewing distance from the display 20 and at a position such that the eye 28 of an examiner is in a viewing region of the display. This arrangement is particularly useful for directional displays, to such as autostereoscopic displays, which form viewing regions from each of which a single view may be observed. In the arrangement illustrated in Figure 5, the aperture is located so as to be substantially in the middle of the right eye viewing window at the viewing distance from the display 20.

Figure 6 illustrates diagrammatically a test image which is generated by the arrangement shown in Figure 4 and displayed by a display device forming part of the display 20, which by way of example is a two view 3D autostereoscopic display. The display device may be of the light-emitting type, such as a fluorescent or light emitting diode device, or of the light attenuating type, such as a liquid crystal device operating in a transmissive or reflective mode. In such displays, the display device is associated with a parallax optic, such as a parallax barrier, to form left and right viewing zones in which left-eye and right-eye images are visible to a viewer. Alternate columns of pixels display strips of left and right eye images which are interlaced on the device. However, the techniques disclosed herein are not limited to directional displays and may, for example, be used to assess crosstalk in stereoscopic displays.

The test image is supplied to the display device during a "testing" phase to allow a viewer to assess the crosstalk generated in the display. As shown in Figure 1, different patterns are supplied in "regions 1 and 2", which operate in the 3D mode during the test phase.

With the pattern shown in Figure 6, the pixels in the region 1 for the left-eye view are set to 100% intensity whereas the pixels for the righteye view in the region 1 are set to 0% intensity. In the case of light-emissive display devices, the pixels for the left and right eye images are set to give maximum light output and zero light output respectively. In the case of transmissive or reflective devices, the left and right eye pixels are set to give minimum and maximum attenuation, respectively. Thus, maximum intensity corresponds to "white" whereas minimum intensity corresponds to "black".

Because of the effect of display crosstalk, when viewed from the righteye viewing zone, the right-eye pixels do not appear black but instead have the appearance of a non zero intensity, for example Y% which provides a measure of the crosstalk which the right-eye of the viewer sees from the left-eye image.

In the region 2, the pixels which display the left-eye image are set for 0% intensity whereas the pixels for displaying the right-eye image are set for X% intensity, where X is between 100 and 0. The right-eye image in the region 2 therefore shows X% intensity regardless of any crosstalk generated by the display. Thus, when the viewer views the display from the right viewing zone, the viewer can compare the intensities of the right-eye images in the regions 1 and 2 and this comparison of the appearance provides information on the crosstalk generated within the display.

For example, in order to provide a quality control check during manufacture of the display, the "reference intensity" X% may be set to match the maximum acceptable crosstalk intensity which would appear in the right-eye view of the region 2 if the display generated a level of crosstalk represented by X% intensity. If the viewer judges that theintensity Y% in the region 1 is larger than the reference intensity X% in the region 2, the crosstalk is higher than the maximum acceptable level and the display may be rejected or allocated to remedial action.

In an alternative test mode, the viewer may be able to control the value of the reference intensity X% while viewing the regions 1 and 2 from the right viewing zone. In particular, the viewer may adjust the reference intensity until the intensities appear the same in the regions 1 and 2. When this condition is achieved, the reference intensity X% gives a nominal indication of the crosstalk level of the display. It is thus possible to perform a test which returns a sufficiently accurate measure of the actual display crosstalk, for example to allow crosstalk correction to be applied in accordance with the technique disclosed in EP 0 953 962.

Figure 6 illustrates a specific example of a more general test pattern. In particular, in Figure 6, the right eye pixels of the region l are set to 0%, the left eye pixels in the region 1 are set to 100% intensity and the left eye pixels in the region 2 are set to 0% intensity. In practice, the pixels in the region 1 and the left eye pixels in the region 2 can be set to more arbitrary values, although the pixels of the left image in the region l cannot be set to 0% intensity because they must produce some light in order to generate the crosstalk. If the left eye image in the region 2 has an intensity other than 0%, this will contaminate the reference intensity generated in the right eye image. This contamination may be sufficiently low to be ignorable or can be compensated.

In the most general case, the reference intensity in the left image in the region 1 is set to an arbitrary non-zero value A% of maximum intensity, the right image in the region l is set to an arbitrary value of B% of maximum intensity, and the right image of the region 2 is set to a value of (B+(A.X)/100)% of maximum intensity, where X is a crosstalk value in per cent. The value X may be varied as described hereinafter to provide continuous ranges of reference intensities or sub- regions of different reference intensities. The intensity value in the left image of the region 2 is arbitrary but, in many embodiments, is preferably 0%.

By way of example, the left image in the region 1 may be set to 20%, the right image in the region 1 may be set to 25% and the left image in the region 2 may be set to 0%. In this case, the right image in the region 2 is set to 28% for a crosstalk value X of 15%.

When viewed from the right eye viewing region, the region 1 will appear brighter than the region 2 when the crosstalk exceeds 15%.

This type of test pattern has potential advantages in that the use of low intensity levels, which may be non-linear, can be avoided. However, this pattern has the disadvantage that it is necessary to distinguish between smaller differences in intensity for given differences in crosstalk resulting in reduced sensitivity compared, for example, with the previous example described with reference to Figure 6.

In another example of a test pattern of this type, the left and right views in the region 1 are set to the same non-zero intensity, such as 50%. The right view in the region 2 is set to 55% so that, when the display exhibits 10% crosstalk, the regions 1 and 2 when viewed from the right eye viewing window will be seen as having the same intensity.

Figure 7 illustrates a test pattern for a modified test mode in which the part of the test pattern in the region 1 is the same as that illustrated in Figure 6 but the part of the test pattern in the region 2 is different in order to provide a calibrated scale of reference intensities. In the region 2, different sub-regions of the pixels for the right-eye image are controlled to provide intensities representing different crosstalk levels. The levels shown in the region 2 of Figure 2 represent crosstalk levels of 1%, 5%, 10% and 15% by way of example but other levels or other numbers of levels may be used in order to achieve a desired measurement range and/or resolution. Figure 7 also illustrates a third region where text or characters indicating the crosstalk levels are displayed against the sub-regions of the region 2 whose intensities match these levels. The viewer can therefore assess which of the sub-regions most closely matches the intensity of the right-eye image pixels of the region 1 and can determine the crosstalk value from the region 3. Crosstalk correction may then be applied as described hereinbefore.

The regions l and 2 are shown as being vertically adjacent each other in Figures 6 and 7 but any other juxtaposition may be used. Also, the test pattern shown in Figure 6 may be replicated with different reference intensities in the different patterns, corresponding to a larger scale version of the type of test pattern shown in Figure 7.

The crosstalk produced by a display may be coloured in that the mechanism or mechanisms producing the crosstalk may change the colour of the light as compared with the light produced by the display in the view for which it is intended. For example, crosstalk may occur when light is scattered from the left-eye image into the right-eye image and other mechanisms producing crosstalk may involve diffraction of light through the parallax optic or transmission of light through nominally opaque regions of a parallax barrier. The amount of such scatter, diffraction and transmission is dependent on the wavelength of light so that, if the left- eye image is white, the crosstalk generated by this in the right-eye view may be coloured. If the crosstalk is a different colour from the reference intensity, it may be difficult to match the crosstalk and reference intensities visually.

In order to reduce or overcome this difficulty, the colour of the reference intensity may be changed so as to match the crosstalk intensity and make relative intensity assessments easier. In the example shown in Figure 4, the viewer is provided with the manually operable control 25 so as to be able to adjust the reference intensity colour until it is judged to match the crosstalk colour.

Another technique for dealing with this difficulty is to use a coloured reference intensity. For example, in the case of ROB displays, the pixels displaying the left image in the region 1 may be set to maximum intensity for one of the colour components and zero intensity for the other two components. The pixels displaying the right image in the region 2 may, at least initially, be set for X% of the one colour and zero intensity for the other two. The colour of the right image in the region 2 may then be adjusted manually to improve the colour match with the crosstalk colour. This may be done in turn for each of the components. By reducing the spectrum of light in the reference intensity, colour deviation caused by crosstalk is less problematic. Also, measuring crosstalk for each of the colour components individually allows crosstalk correction to be applied to each colour component so as to improve such correction.

Other effects may occur to distort the intensities produced by the display device. For example, the grey scale produced by the display device may be non-linear so that some form of calibration may be necessary or desirable. For a display of known crosstalk, the reference intensity may be set to match the crosstalk intensity. The reference intensity then indicates the known crosstalk level. Alternatively, the intensities produced by the display device may be calibrated against measurements made by an optical detector.

Figure 8 illustrates a test pattern which differs from that shown in Figure 6 in that the regions 1 and 2 are disposed side-by-side and the reference intensity X% is displayed by the pixels in the region 2 in the left and right-eye images. For the left-eye, the reference intensity is nominally produced by the left-eye image. However, there may be some contribution from the right-eye image because of display crosstalk. The reference intensity would then be dependent on the crosstalk level but compensation for this may be provided by modifying the reference intensity.

Figure 9 illustrates another test image which is suitable for use with displays which can be switched between multiple view (3D autostereoscopic) and single view (2D) modes of operation. For example, in the single view mode, the left and right-eye images are the same. Also, in this arrangement, the part of the display device displaying the test image is not divided into the regions 1 and 2 but, instead, these two images are displayed alternately by the same region of the display device.

In the 3D mode, the pattern displayed by the display device is essentially the same as that illustrated in the region 1 in the previously described embodiments, with the pixels displaying the left-eye image being set to 100% intensity and the pixels displaying the right eye image being set to 0% intensity. In the 2D mode, all pixels are controlled to display a graded or stepped greyscale image, for example varying in intensity from 0% to about 15%. In both modes, a scale corresponding to the adjacent reference intensities is displayed.

In use, the display cycles between the 2D and 3D modes and the viewer assesses which part of the graded or stepped greyscale is closest in intensity to that which he sees with the display in the 3D mode in the right viewing zone. The corresponding reference intensity can then be read from the scale.

Where the display is of the type which can have different display device regions operating simultaneously in the 2D and 3D modes, the reference intensities may be displayed in the 2D region of the display and the pattern for generating crosstalk may be displayed in the 3D region of the display. As a further alternative, a separate display may be used to provide the reference intensity and may be disposed adjacent the autostereoscopic display to facilitate assessing the relative intensity levels by the viewer.

Arrangements in which switching or comparison between 2D and 3D modes of operation is performed, the displays will require calibration because the 2D mode is normally naturally brighter than the 3D mode. In particular, in the 2D mode a parallax generating device is disabled or removed so that the light attenuation resulting from the use of such a device is also removed.

Figure 10 illustrates another method of assessing crosstalk in which the viewer may use both eyes to view the display autostereoscopically, i.e. with the left and right eyes in the left and right viewing zones, respectively. In this case, the lower region 2 operates in the 2D mode or may be embodied by a separate 2D "reference" display.

The left and right views in the upper region 1 are set to the same nonzero intensity and the lower region provides the reference intensity. For example, if the left and right upper intensities are set to 50%, then in a display exhibiting 10% crosstalk, there will be a 5% intensity added to the right view from the left view and vice versa. The lower region is set to 55% intensity and will match the intensity of the upper region. Higher or lower levels of crosstalk will then result in the upper region being brighter or darker, respectively, than the lower region.

As described hereinbefore, because the upper and lower regions are operating in different modes, calibration may be necessary in order to compensate for the absence of a parallax generating device, such as a parallax barrier, which reduces the light intensity in the 3D mode. Also, this test pattern cannot be used to assess all sources of crosstalk.

For example, when crosstalk arises from scattered light, light from the left eye image may be scattered into the right eye image so that the left eye image is reduced in brightness. However, the left eye image is increased in brightness by scattered light from the right eye image. Thus, crosstalk is increased but the brightness of the upper region remains the same. Thus, crosstalk caused by scatter cannot be detected.

Figure 11 illustrates a crosstalk assessment method which may be used with a multiple view display of the 3D autostereoscopic type having a viewer position indicator, for example of the type disclosed in EP 0 860 728, the contents of which are incorporated herein by reference. Such indicators are arranged to provide a clear indication to a viewer to guide them so as to be correctly positioned with respect to the display for correct autostereoscopic viewing. When positioned correctly, both eyes of the viewer see the indicator as nominally black, but the black level is contaminated by crosstalk produced by the display. When the viewer is further away laterally from the correct viewing position, the indicator is seen as white or as a specific colour.

In order to assess crosstalk level, a region of the display device above the viewer position indicator (VPI) displays a greyscale of reference intensities in the pixels displaying both the left and right-eye images. The viewer views the VPI so as to position himself or herself in the correct viewing position relative to the display and then visually matches the intensity in the VPI region of the display with the nearest grey level so as to determine the display crosstalk.

As described hereinbefore, the actual reference intensities may vary from the nominal intensities because of the display crosstalk. However, this can be at least partly corrected by calibrating the reference intensity levels. Also, if the VPI displays a particular colour instead of white, the reference intensities may be displayed in the same colour so as to ease assessment of intensity matching.

In the embodiments described hereinbefore, crosstalk assessment is done by matching intensities. However, such assessments can be made by matching colours and test images for use in such colour matching methods are illustrated in Figures 12 and 13.

In Figure 12, a first outer or border region 110 of the display provides a grey image by illuminating the red, green and blue pixels for the left and right-eye images to 50% intensity. In this respect, although a grey reference image is illustrated, other colours may be chosen for the comparison. In an inner or central region 111, the pixels for displaying the left-eye image display a red image with the red pixel intensity being 100% and the green and blue pixel intensities being 0%. In the right-eye image, the red pixels are set to 35% intensity whereas the green and blue pixels are set to 50% intensity.

In order to assess crosstalk, the pattern is viewed from the right viewing zone. If the display crosstalk is 15%, the inner region 111 appears grey and matches the colour and intensity of the outer region 110. If the crosstalk is less than 15%, the inner region 111 appears more cyan than the outer region 110. If the display crosstalk is greater than 15%, the inner region l l l appears more red than the outside region 110.

This method may be repeated with the red and green components in the lefteye image of the inner region 111 set in turn to 100% and the other components set to 0% intensity. In each case, in the right-eye image of the inner region 111, the component whose crosstalk is being assessed is set to a lower level of intensity than the other two components by an amount corresponding to an acceptable level of display crosstalk.

The colouring of the inner region 111 compared with the outer region 110 seen by the viewer then allows the display crosstalk to be assessed and, in particular, allows a judgement to be made as to whether the display crosstalk is acceptable.

Figure 13 illustrates pattern having a common grey outer region 110 as illustrated in Figure 12 but with a plurality of (in this case three) inner regions 11 la, 11 lb and 11 to in which different intensities of the red component are displayed in the right-eye image in the different inner regions. For example, the red component has a 40% intensity in the region Lila, a 30% intensity in the region limb, and a 20% intensity in the region talc. The viewer then views the display from the right viewing zone and assesses which of the inner regions Lila, lllb and 111c most closely matches the colour and intensity of the outer region 110 so as to determine the display crosstalk.

Figure 14 illustrates another test pattern which may be used to assess the display crosstalk. In this case, the same stereoscopic image is displayed at several regions of the display device but with the different images having different levels of crosstalk correction, for example provided in accordance with the technique disclosed in EP O 953 962 and described hereinafter. The viewer views the autostereoscopic display in the normal manner with left and right-eyes in the left and right viewing zones. Each image is most comfortable to the viewer for a particular level of crosstalk and the viewer therefore judges which image is the most comfortable to view. The display crosstalk can then be deduced from this.

The method of providing crosstalk correction is based on adding a base level of grey to every pixel of both the left and right images in each region so as to raise the background grey level. The amount of grey is preferably equal to or greater than the amount of crosstalk correction required. A percentage of the left image corresponding with the amount of crosstalk to be corrected is then subtracted from the right image and vice versa. This results in a low intensity negative image in the background grey level.

When the corrected images are displayed, the crosstalk fills in the negative images so that a uniform background grey level is restored. Thus, image contrast is sacrificed so as to improve crosstalk and hence improve the perceived 3D image quality.

The amount of crosstalk correction required is chosen for the regions as 5%, 15% and 20% in the example shown in Figure 14.

The method is described in more detail hereinafter for a pixel in the left image and its corresponding pixel in the right image, where: 1- is - the incoming RGB colour vector for the left pixel; Ir- is - the incoming RGB colour vector for the right pixel; lb- is - the colour vector with raised background grey level; lo- is - the output colour vector with crosstalk correction; C- is - the scalar crosstalk correction in the range [0.255]; and Im = 255 - is - the scalar maximum value of each RGB component.

All of the individual values are integers in the range [0,255] assuming 8 bits per colour component in each 24 bit full colour pixel. For the upper region in Figure 14 having 5% of crosstalk correction, the value of C is given by 255 x 5/100 = 13. Similarly, C = 38 for 15% correction and C = 51 for 20% correction.

First, a background grey level is added to the left image pixel: I = I *(Im C:+c (1) The corresponding right image pixel crosstalk correction is subtracted from the new value of the left image pixel: Io =Ib-Ir *- (2) The value lo is then output as the new left image pixel colour value.

This method must be applied to every pixel in the left image to correct for right image crosstalk and to every pixel in the right image to correct for left image crosstalk.

The method may be implemented in software. Alternatively, for a hardware implementation, the method may be performed using only integer arithmetic. This significantly reduces the complexity of a hardware implementation by removing the IS need for a floating-point arithmetic unit.

For binary computing devices, the use of numbers that are a power of two has significant advantages. For this reason the input pixel colour values are raised by one from the range [0,255] to the range [1,256]. The above method can then be re-written as below where K is the scalar cross talk correction in the range [1,256].

From equations (1) and (2): ( ) ( (t + i)-K) ( )( K) (3) multiplying out the above gives: (Io +1)(Im +1)=(I, +1)(Im +1)+K(Im +1)-K(I, +1)-K(Ir +1) (4) rearranging this gives: I =I + K(Im( Ii)r) (5) This is computed efficiently using a bit shift operation instead of division since the value (Im +1) = 256 and is accounted for with a bit-wise right shift by 8 bits.

I = I + K(I -I -I -1) >> 8 (6) The output crosstalk corrected value for the right pixel is computed similarly: lo = Ir + K(Im-I,-Ir-1) >> 8 (7) This computation is performed in the controller 21 and is illustrated in Figure 15, where Is the partial result K(Im - I, - Ir -1) >> 8 is computed at 32 and is added to the left and right pixel values 1 and Ir at 33 and 34, respectively.

If the value of K is restricted to be a factor two, where K = 2n, then the computation is more simply implemented since the multiplication by K can be incorporated into the right shift, resulting in a computation requiring only addition, subtraction and right shift operations, ie: Io = I, + (Im-I,-I r-1) >> (8-n) (8) Use of expression (6) or (8) has the advantage that all the arithmetic is integer with the largest resulting number requiring only an 18 bit signed value, thus greatly reducing the implementation complexity of the hardware.

The crosstalk correction methods described hereinbefore assume that the display used has a perceived linear response to the input values. This is not normally the case and it is usually compensated for by using gamma correction in the video display driving circuit, for instance as disclosed in Glassner, "Principles of Digital Image Syntheses", Morgan Kaufman, 1995, Chapter 3, pp 97-100. With the above method, the gamma correction can be applied to every pixel after the crosstalk correction has been computed. Alternatively, it can be applied to the correction factor K alone before crosstalk correction is computed.

Also, the above methods assume a 24 bit or similar full colour pixel value. This is not always the case and, in some systems, colour indexing is used to save memory. This is where there is a limited range of colour values and the actual value stored in video memory is an index into a look-up table, which holds the full 24 bit RGB values. Often, with this approach, only 8 bits are stored per pixel, resulting in 256 possible colours on the display at any one time. For colour indexing systems, the crosstalk correction should be performed after the colour index has been decoded into its 24 bit RGB display driving values.

It is thus possible to provide a technique which allows a viewer to make an assessment of display crosstalk in a multiple view directional display requiring only the display of a test image. Thus, no measuring equipment or automated measuring technique is required and the result of the assessment may be used, for example, to provide correction of crosstalk or to provide a quality control check during display manufacture.

By repeating the test during use of a display, any changes in crosstalk level which occur during the life of the display may be corrected so as to optimise crosstalk correction and thus optimise the performance of the display during the working life of the display.

Claims

CLAIMS: 1. A method of generating a test image for use in assessing

crosstalk in a multiple view display, comprising the steps of: (a) selecting a first intensity having a value of A% of maximum intensity for display in a first view at a first display region of the display; (b) selecting a second intensity having a value of B% of maximum intensity for display in a second view at the first region; and (c) selecting at least one third intensity, having a value of (B + (A À X)/100)% of maximum intensity, where X is a crosstalk value in per cent, for display in the second view at a second display region of the display.
2. A method as claimed in claim 1, comprising displaying or printing the test Image.
3. A method as claimed in claim 1 or 2, comprising storing or transmitting data representing the test image.
4. A method as claimed in any one of the preceding claims, in which the first intensity has a value A of 100% of maximum intensity.
5. A method as claimed in any one of the preceding claims, in which the second intensity has a value B 0% of maximum intensity.
6. A method as claimed in any one of the preceding claims, in which the step (c) comprises varying the at least one third intensity by varying the crosstalk value X.
7. A method as claimed in any one of claims 1 to 5, in which the step (c) comprises selecting a plurality of different third intensities having different crosstalk values X for display in respective sub-regions of the second region.
8. A method as claimed in claim 7 when dependent on claim 2, displaying or printing adjacent each of the sub-regions a text indication of the crosstalk value X corresponding to the third intensity of the respective sub-region.
9. A method as claimed in any one of the preceding claims, comprising adjusting the colour of the at least one third intensity to match the colour in the second view at the first region.
10. A method as claimed in any one of claims 1 to 8, comprising performing the steps (a) to (c) for at least one of a plurality of colours. i5
11. A method as claimed in claim 10, comprising performing the steps (a) to (c) for each of the colours.
12. A method as claimed in claim 10 or 11, in which the display is an ROB display and the colours comprises red, green and blue.
13. A method as claimed in any one of the preceding claims, comprising selecting a fourth intensity, which is less than or equal to the third intensity, for display in a first view at the second region.
14. A method as claimed in claim 13, in which the fourth intensity is substantially equal to the second intensity.
15. A method as claimed in claim 13, in which the fourth intensity is substantially equal to the third intensity.
16. A method as claimed in claim 15, in which the first and second regions at least partially overlap each other, comprising switching the display between a multiple view mode, in which the first and second intensities are displayed, and a single view mode, in which the third and fourth intensities are displayed.
17. A method as claimed in claim 15, in which the first region provides multiple views and the second region provides a single view.
18. A method as claimed in claim 15, in which the first and second regions are provided by different display devices.
19. A method as claimed in any one of claims 15 to 18, in which the first region comprises a viewer position indicating region.
20. A method as claimed in any one of the preceding claims, comprising the further step of comparing the intensities displayed in the second view at the first and second 1 5 regions.
21. A method as claimed in claim 20, in which the comparing step comprises visually comparing the appearances of the second view at the first and second regions.
22. A method as claimed in any one of the preceding claims, in which the display is a directional display.
23. A method as claimed in claim 22 when dependent on claim 20 or 21, in which the intensities are assessed in a viewing region of the second view.
24. A method of generating a test image for use in assessing crosstalk in a multiple view display, comprising the steps of: (a) selecting a first colour, which has N components having values of AI%, 30.., AN% of maximum intensity respectively, for display in a first view at a first display region of the display, where N is an integer greater than 1; (b) selecting a second colour, which has N components having values of BI%, À. ., BN% of maximum intensity, respectively, for display in a second view at the first region; and (c) selecting at least one third colour, which has N components having values of (BI+(AI.X)/100)%, ...., (BN+(AN.X)/100)% of maximum intensity respectively, where X is a crosstalk value in percent, for display in the second view at a second display region of the display.
25. A method as claimed in claim 24, comprising displaying or printing the test Image.
26. A method as claimed in claim 24 or 25, comprising storing or transmitting data representing the test image.
27. A method as claimed in any one of claims 24 to 26, in which N is equal to 3.
28. A method as claimed in claim 27, in which the first component has a value A} of 100% of maximum intensity.
29. A method as claimed in claim 27 or 28, in which the value of each of the components A2 and A3 is 0% of minimum intensity.
30. A method as claimed in any one of claims 27 to 29, in which the values of the components B., B2 and B3 are substantially equal to each other.
31. A method as claimed in any one of claims 27 to 30, in which the value of each of the components B2 and B3 is substantially equal to half maximum intensity.
32. A method as claimed in any one of claims 27 to 30, in which the step (c) comprises selecting a plurality of different third colours having different first component values of (BI+(AI.X)/100)% of maximum intensity for display in respective sub-regions of the second region.
33. A method as claimed in any one of claims 24 to 32, comprising the further step of comparing the colours displayed in the second view at the first and second regions.
34. A method as claimed in claim 33, in which the comparing step comprises visually comparing the appearances of the second view at the first and second regions.
35. A method as claimed in any one of claims 24 to 34, in which the display is a directional display.
36. A method as claimed in claim 35 when dependent on claim 33 or 34, in which the colours are assessed in a viewing region of the second view.
37. A method as claimed in any one of the preceding claims, in which the display is a two view display.
38. A method as claimed in any one of the preceding claims, in which the display is a three-dimensional autostereoscopic display.
39. A test image generator for generating a test image for use in assessing crosstalk in a multiple view display, comprising: means for generating first image data, representing a first intensity having a value of A% of maximum intensity, for display in a first view at a first display region; means for generating second image data, representing a second intensity having a value of B% of maximum intensity, for display in a second view at the first region; and means for generating third image data, representing at least one third intensity having a value (B + (A X)/ 100)% of maximum intensity where X is a crosstalk value in per cent, for display in the second view at a second display region.
40. A test image generator for generating a test image for use in assessing crosstalk in a multiple view display, comprising: means for generating first image data, representing a first colour, which has N components having values of A1%, ...., AN% of maximum intensity respectively, for display in a first view at a first display region where N is an integer greater than 1; means for generating second image data, representing a second colour, which has N components having values of B1%, À., BN% of maximum intensity respectively, for display in a second view at the first region; and means for generating third image data, representing at least one third colour, which has N components having values of (BI+(AnX)/100)%, ...., (BN+(AN-X)/100)% of maximum intensity respectively, where X is a crosstalk value in percent, for display in the second view at a second display region.
41. A method of generating a test image for use in assessing crosstalk in a 3D display, comprising applying crosstalk correction to at least one stereoscopic pair of images and causing the display to display the or each corrected pair of images.
42. A method as claimed in claim 41, in which the amount of crosstalk correction is

selectable.
43. A method as claimed in claim 41 or 42, comprising applying different amounts of crosstalk correction to a stereoscopic pair of images to create a plurality of corrected image pairs and causing the display to display the corrected image pairs.
44. A method as claimed in any one of claims 41 to 43, comprising selecting a value of crosstalk correction corresponding to a most natural 3D image appearance.
45. A method as claimed in any one of claims 41 to 44, in which the crosstalk correction is applied by: adding a grey level to a first image to form a first sum; adding said grey level to a second image to form a second sum; subtracting from the first sum an amount equal to a given fraction of said second image; and subtracting from the second sum an amount equal to the given fraction of said first image, wherein these steps comprise the calculation of a partial result which is used to determine crosstalk corrected picture elements for both the first and second images.
46. A method as claimed in claim 45, in which each picture element comprises M colour components having an intensity value, and the method further comprises, for an intensity level Ix of each of the picture elements of said first image, determining a crosstalk corrected picture intensity level Iox according to: IN = I + K(lm-[x-Iy-1) (1,,, + I) or an equivalent form thereof and for an intensity level Iy of each of the picture elements of said second image, determining a crosstalk corrected picture intensity level Icy according to: K(lm-IX-Iy-1) (I m + 1) or an equivalent form thereof, where K is the scalar crosstalk correction; and Im is the scalar maximum value of each colour component.
47. A method as claimed in claim 46, in which the partial result used to determine crosstalk corrected picture element is: K(lm-Ix Iy I)
48. An apparatus for generating a test image for use in assessing crosstalk in a 3D display, comprising means for applying crosstalk correction to at least one stereoscopic pair of images and means for causing the display to display the or each corrected pair of images.
49. A method as claimed in any one of claims 20, 21, 23, 33, 34 and 36, comprising determining from the comparison a display crosstalk value.
50. A method as claimed in any one of claims 41 to 47, comprising determining from the or each displayed corrected pair of images a display crosstalk value.
51. A method as claimed in claim 49 or 50, comprising recoding the display crosstalk value.
52. A display whose crosstalk is corrected in accordance with the display crosstalk value determined by a method as claimed in any one of claims 49 to 51.
53. Use of an image generated by a method or apparatus as claimed in any one of claims 1 to 51, comprising displaying the image.
54. Use as claimed in claim 53, comprising assessing display crosstalk from the displayed image.
55. Use as claimed in claim 54, comprising recording the display crosstalk.
56. A display whose crosstalk is corrected in accordance with the display crosstalk recorded by use as claimed in claim 55.