Currently, there is a wide range of 3D display technologies, but not all of them are appropriate for mobile use. For example, wearing glasses to aid the 3D perception of a mobile device is highly inconvenient. The limitations of a mobile device, such as screen size, CPU power and battery life limit the choice of a suitable 3D display technology. Another important factor is backward compatibility – a mobile 3D display should have the ability to be switched back to “2D mode” when 3D content is not available.
Autostereoscopic displays are a class of displays which create 3D effect without requiring the observer to wear special glasses. Such displays are using additional optical elements aligned on the surface of the screen, to ensure that the observer sees different images with each eye. Typically, autostereoscopic displays present multiple views to the observer, each one seen from a particular viewing angle along the horizontal direction. However, the number of views comes at the expense of resolution and brightness loss – and both are limited on a small screen, battery driven mobile device. As mobile devices are normally watched by only one observer, two independent views are sufficient for satisfactory 3D perception. At the moment, there are only a few vendors with announced prototypes of 3D displays, targeted for mobile devices. All of them are two-view, TFT-based autostereoscopic displays.
The basic operational principle of an autostereoscopic display is to “cast” different images towards each eye of the observer. This is done by a special optical layer, additionally mounted on the screen surface which redirects the light passing through it. There are two common types of optical filters – lenticular sheet which works by refracting the light, and parallax barrier which works by blocking the light in certain directions. In both cases, the intensity of the light rays passing through the filter changes as a function of the angle, as if the light is directionally projected. These two technologies are shown in the figure below.
The main advantage of parallax barrier is the ability to be switched off, thus providing backwards compatibility with 2D content. The main disadvantage of the parallax barrier is that it blocks part of the light, resulting in a lowered brightness of the display. In order to compensate for that, one needs an extra bright backlight, which would decrease the battery life if used in a portable device.
TFT displays recreate the full color range by emitting light though red, green and blue colored components (sub-pixels). Sub-pixels are usually arranged in repetitive vertical stripes as seen in Figure 1. Since sub-pixels appear displaced in respect to the optical filter, their light is redirected towards different positions. One group will provide the image for the left eye, another – for the right. In order to be shown on a stereoscopic display, the images intended for each eye should be spatially multiplexed. This process is known as interleaving, or interzigging, and depends on the parameters on the optical filter used. Two topologies are most commonly used. One is interleaving on pixel level, where odd and even pixel columns belong to alternative views. The other is interleaving on a sub-pixel level – where sub-pixel columns belong to alternative views. In the second case, differently colored components of one pixel belong to different views. The next figure shows pixel-level (left) and sub-pixel-level (rigth) interleaving.
Multiview 3D displays work by simultaneously showing a set of images (“views”), each one seen from a particular viewing angle along the horizontal direction. For a large number of head positions, the eyes of an observer can perceive a scene at different angles, which enables 3D perception without wearing glasses. The advantage of multiview displays is that they provide some degree of freedom to observe a 3D scene from different angles, and the user is not “tied” to a particular “sweet spot”. The main disandtantage of multiview displays is that with incleasing the number of views, the resolution of each view decreases.
All modern multiview displays use TFT screens for image formation. The light generated by the TFT is separated into multiple directions by the means of special layer additionally mounted on the screen surface. Such layer is called “optical layer”, “lens plate” and “optical filter”.
However, TFT displays do not have full-colour pixels. They recreate the full colour range by emitting light though red, green and blue coloured components (sub-pixels), usually arranged in repetitive vertical stripes as illustrated in Fig 1a. The optical filter mounted on top of the screen also has repetitive structure, which redirects the light passing through it. There are two common types of optical filters – lenticular sheet which works by refracting the light, and parallax barrier which works by blocking the light in certain directions. In both cases, the intensity of the light rays passing through the filter changes as a function of the angle, as if the light is directionally projected. Since sub-pixels appear displaced in respect to the optical filter, their light is redirected towards different positions. The image, seen on the screen from a particular direction is said to form a view. As a result, differently coloured components of one pixel belong to different views. Respectively, the image formed by one view will be combination of colour components (sub-pixels) of various pixels across the TFT screen. When red, green and blue sub-pixels are visible from the same direction and appear close to each other, the triplet is perceived as one pixel. Such pixel is a building block of the view seen from that direction, and is sometimes referred to as “poxel”. For every poxel there is a certain angle, from which it is perceived with maximal brightness – that angle we call optimal observation angle for the poxel. The vector, which starts from the poxel, and follows the optimal observation angle, is the optimal observation vector for the poxel. As shown later, the optimal observation vectors for all poxels of the same view are designed to intersect in a tight spot in front of the multiview display. From this spot, the view will be perceived with its maximal brightness, and we denote that spot as being the optimal observation spot of the view. Outside of the optimal observation spot, there is a range of observation angles, from which a given view is still visible, even though with diminished brightness. We refer that range to as the visibility zone of a view.
For most multiview displays, visibility zones of the views are ordered in horizontal direction. Notable exception is the SynthaGram display produced by StereoGraphics, which has 9 views ordered in 3-by-3 grid. As the amount of the pixels provided by the underlying TFT is limited, there is a trade-off between the number of views casted by a 3D display and the resolution of each view. As stereoscopic depth cues are perceived mostly in horizontal direction, most multiview display designs do not allocate pixels for extra views in vertical direction.
When horizontally ordered, the visibility zones appear in a fan-shaped configuration. The repetitive structure of the optical filter creates several observation zones for any view, which follow the fan-shaped configuration as well. After the visibility zone of the last view, the first view becomes visible again. This creates one central set of visibility zones straight in front of the screen, and a number of identical sets to the side as shown in Fig. 3b. An example for two observation angles from which the same set of sub-pixels is visible is shown in Fig. 3c. The observation angles marked as “1” and “1R” are optimal observation angles of the view marked with “1” in Fig. 3b. However, angle “1” belongs to the central set of viewing zones, while angle “1R” belongs to the set in the right set.
Having discrete boundaries between the viewing zones creates two common artefacts, found in autostereoscopic displays – image flipping, caused by the noticeable transition between the viewing zones, and picket fence effect, a moire-like artefact caused by the gaps between sub-pixels being magnified by the lenticular sheet. To mitigate these effects, some vendors intentionally broaden the observation angle of the pixels, interspersing the viewing zones. It is also speculated, that blurring the boundaries between the viewing zones can increase the apparent number of views. In 1996, van Berkel proposed that the lenticular sheet could be placed at a slant over a standard LCD screen, as shown in Fig. 3a. This approach removes the picket fence effect, creates smooth transition between the views and at the same time balances the horizontal vs. vertical resolution of a view. The idea has been patented by Philips and utilized in 3D displays produced by the company. Another solution with similar effects is called “wavelength-selective filter array” as proposed by 4D-Vision GmbH. Essentially, the filter is a slanted parallax barrier which covers the display and defines particular light penetration direction of each sub-pixel. Depending on the observation angle and the distance to the observer, most of the sub-pixels are masked. The effect is illustrated in Fig. 4a. The sub-pixels which are visible form one view of the screen. The process of mapping sub-pixels of a multiview display to a certain view is called “interdigitation”, “view multiplexing”, or sometimes “interzigging”. The last term comes from the name of Interzigg TM - the interdigitizing tool developed by StereoGraphics for their monitors. The structure of the optical filter defines certain correspondence map between sub-pixels and views, which we call “interdigitation map”. The interdigitation map for the screen produced by 4D-Vision GmbH is shown in Fig. 4b.
Although effective, both approaches – slanted lenticular sheet and slanted parallax barrier – create a number of secondary problems, hopefully to be resolved by software means. The sub-pixels of a view are now placed on a non-rectangular grid, which creates the need of specially designed anti-aliasing filters. Interocular crosstalk between the interspersed images can hinder the ability of an observer to fuse the views of a 3D scene. Furthermore, as various slanted slices of rectangular shaped sub-pixels are coming into and disappearing from view, the images produced are accurate in average, but imbalanced in detail. Depending on the design parameters, such imbalance could manifest itself as colour tint, or brightness fluctuations, as the centre of the lenticular segment continuously shifts along the discrete rows.