Issue 2.03 - March 1996

3-D Viewing Goes Glassless

By Chris Chinnock



Long the dream of the imaging industry, three-dimensional displays are finally emerging from the lab. These systems promise to aid doctors in visualising patients' bone structures before surgery, help scientists make sense out of mountains of data and, of course, let kids shoot at bad guys with breathtaking realism. However, most current techniques require users to don a geeky-looking pair of special glasses to see the 3-D effect. Only recently have practical methods been developed that allow for natural three-dimensional viewing without special eye-wear. These techniques are blowing the 3-D display market wide open. But hurdles remain: most of the glassless methods are expensive and work only when the viewer sits perfectly still.

The key to human 3-D vision is our ability to discern depth cues. Because we have two eyes, we see objects from slightly different angles, allowing us to determine whether one object is in front of the other. This is known as parallax, a critical ingredient - along with factors such as shading and shadowing - in allowing us to perceive our three-dimensional world.

The idea behind many 3-D display methods is to mimic human vision by producing separate left- and right-eye images. Perhaps the best way to do this is by taking pictures with a stereoscopic camera. These cameras feature two optical channels that record separate images of the same scene from slightly different perspectives. For example, Imax Corporation of Toronto has developed a US$1.3 million (£833,000) camera that simultaneously exposes left-eye and right-eye images onto separate pieces of film.

In the movie theatre, viewers must wear glasses that alternate their eyes' reception of left and right images. On the screen, first the left image is displayed, then the right image. The user wears liquid-crystal-display shutter glasses that are synchronised to the sequence of images: the glasses open the left-eye shutter when the left image is on the screen and block transmission to the right eye. Images alternate between eyes at 48 frames per second (24 frames per second for each eye), so users perceive a 3-D image without experiencing any flicker.

To eliminate the glasses, manufacturers are working on autostereoscopic devices, hoping to perfect optical methods that automatically project the left and right images into the corresponding eyes. A number of devices under development - including image splitters, lenticular lenses and holographic optical elements - can perform this spatial separation.

To separate images spatially, Sanyo in Osaka, Japan, uses a small LCD and a simple image-splitting technique called parallax stereogram. An optical filter that consists of a series of narrow slots is placed in front of the LCD screen. By displaying the left image on odd columns of the LCD, the picture passes through the slots in such a way that it is directed towards the viewer's left eye. The right image, displayed on even columns, travels a different light path through the slots to arrive at the viewer's right eye. Each eye sees a different image, producing a 3-D effect. It's like viewing the world through Venetian blinds.

The difficulty is that the viewer's head must be in just the right place to observe the 3-D effect. That's why researchers at Richmond Holographic Research and Development Ltd. in London use electronically steerable optical elements to direct light to a viewer's left and right eyes. Used with a head-tracking device, this system allows the 3-D image to follow the viewer's movements.

Because LCDs are not widely available in sizes above 12 to 13 inches, developers of large 3-D screens are turning to alternative technologies. For 40-inch and 70-inch displays, Sanyo uses two CRT projectors - each driven by left and right video images - and double lenticular lenses (a series of long, vertical half-cylinders) to emit spatially separate left and right images. The two projectors illuminate the lenticular lenses from the back at slightly different angles, projecting the images onto a specially designed screen. The result is a spatially separated stereo-image pair. However, as with the LCD system, the 3-D viewing area is limited, and viewers will notice that the 3-D image shifts when they move their heads.

A different kind of glassless 3-D viewing can be achieved with volumetric displays. These fill a volume of space with a three-dimensional image. A volumetric display of a sporting event, for example, would allow viewers to see plays from different perspectives simply by moving to the other side of the display. You might think a football receiver was in bounds, but a viewer on the other side of your living room might have a better angle and call the player out.

Researchers at the US Navy's NRaD Laboratory in San Diego use a spinning helix as the surface on which they create the volumetric image. The helix, which resembles a spiral staircase, spins inside a see-through box. An image is projected by lasers onto the spinning helix one point at a time. When this is done fast enough, the eye perceives all of these points as a continuous image. The viewer can walk around this display observing the image from every possible angle.

While all of these different approaches work, they are all also quite expensive: typically from $50,000 (£31,250) to more than $100,000 (£62,500). As a result, they will probably first show up in medical applications and in arcades or other entertainment-based locations. But it seems that every month brings new and improved techniques for glassless 3-D viewing. We may be able to break out of flatland sooner than anyone expects.

Chris Chinnock (76061.3671 @compuserve.com) writes about emerging and optoelectronic technologies.