Shimmering Images

Toby Howard

This article first appeared in Personal Computer World magazine, January 1997.

ALTHOUGH WE LIVE in a three-dimensional world, almost all of our computer pictures are displayed in two dimensions, on flat screens. Clever display and interaction techniques help bridge this dimensional gap, and they do quite a good job - even after a few minutes of using VRML on the Web, or playing Quake, most people will have the strong sensation of being inside the world on the other side of the screen. But imagine if we could create three-dimensional images that actually floated before our eyes, like the holographic video recordings projected by R2D2 in Star Wars. Researchers at Stanford University are bringing the reality closer, with a new display technology, where true 3D images float in a cube of glass.

Making convincing three-dimensional images has long been a Holy Grail of artists and engineers. In the 15th century Albrecht Durer's pioneering studies of the laws of perspective laid the foundation for the creation of flat images that embodied a sense of depth. And the earliest 19th century photographers experimented with "stereo pairs" -- two photographs of a scene taken from slightly different positions, corresponding to the views seen by the left and right eyes. Viewed suitably, stereo pairs can give a remarkably convincing 3D effect. But it remains an illusion, not a true 3D image.

The technology behind the 50s craze for 3D movies ("See the hideous flying fingers of the swamp monster!") remains popular today. Glasses with green and red lenses enable each eye to see only the appropriate image of the two projected simultaneously on the screen. Although the technique works, there are problems. The images suffer from cross-talk, so that the left eye sees some of the right eye's image, and vice versa. This weakens the illusion, strains your eye, and gives you a headache.

A more modern approach uses polarisation, instead of colour filtering. Left and right eye images are projected onto a screen through polarising filters, the left image vertically polarised, the right image horizontally polarised. The viewer wears lightweight glasses whose lenses are correspondingly polarised. Crosstalk is eliminated, and if the system is well-aligned, you can see a full-colour 3D image floating in front (or behind) the projection screen. It is only when you reach out to touch the image, that you realise there is nothing there.

3D display technologies are now appearing which allow the viewer can see a stereo image without wearing any special glasses. The principle of these "auto-stereoscopic" displays is still to present the eyes with left-hand and right-hand images, but the geometry of the display is such that each eye sees only the image intended for it. Several research groups are working on displays which can also track the location and orientation of the viewer's head, using ultrasound or infra-red beams, adjusting the images accordingly.

Although these techniques give exciting results, they are not creating true 3D images; they are presenting us with pairs of 2D images, and it is up to our brains to construct the true 3D scene they represent. Some people are better than others at doing this, and some people cannot do it at all.

The latest, and most futuristic, development is the "volumetric display", which generates an authentic 3D image. The image hangs in space while you move around and look around it. This is R2D2 territory.

There are currently two main volumetric systems, known as "swept-volume" and "static-volume". In swept-volume displays, the display screen itself is flat, but it is rapidly rotated to sweep out a three-dimensional volume. If the screen is moved sufficiently quickly, the eye can't see it, but if particular pixels on the screen are repeatedly switched on, in sync with the rotation, they are bright enough to be perceived as spots of light suspended in space (rather like the streaks of light you see when twirling a sparkler). A version of this technology, on show recently at the Electronic Imaging Symposium, used a rotating helical surface illuminated by a laser. It was affectionately known as Felix the Helix.

Many researchers believe the future of 3D displays lies with the static-volume display, which is a transparent grid of 3D pixels, or voxels. An unilluminated voxel is invisible, but when switched on, it appears as a spot of light floating in space.

The prototype display recently announced by Elizabeth Downing and her colleagues at Stanford University works on this principle. It uses a block of transparent glass, in which voxels are illuminated when two invisible infra-red laser beams intersect. To create a complete image, the lasers are scanned around the volume, to repeatedly light up all the corresponding voxels, often enough to maintain a stable image. It's a clever technology, based on the strange physics that is never very far away when cutting edge technology is discussed: quantum mechanics.

The voxels in the display cube are created by introducing tiny amounts of rare-earth elements into the glass during its manufacture. Each rare-earth ion normally exists in its lowest stable energy state (call it state A). But if it is illuminated by laser light of a suitable wavelength, it absorbs some of the light and jumps into a more excited energy state (B). It will stay in state B for a short time, before decaying back to state A. However, if it receives another jolt of laser light of a different wavelength while it is in state B, it gets even more excited, and enters state C. This is as excited as it can get, and after a few thousandths of a second it drops back to state A, emitting its suplus energy as visible light. The voxel lights up.

The current prototype of the display is tiny -- about the size of an OXO cube, and monochromatic, the colour depending on which particular rare-earth element is used. There remain problems of scanning the lasers suitably, and creating arbitrarily coloured voxels. But the basis of the technology is proven, and the feeling is that volumetric displays are on the verge of a breakthrough.

Synthetic worlds and their monsters, once safely sealed inside our old monitors, will one day pop up right in front of our eyes.

Toby Howard teaches at the University of Manchester.