“WHEN you look at a screen you expect to see something flat,” says Chris Yewdall. “But we can change that.” Yewdall, president of Digital Dynamic Depth, a software company based in Santa Monica, California, is totally serious. On the screen in front of us is a trailer for Shrek – the computer-animated fairy-tale film of a bumbling, green ogre. With a click of the mouse, the hero bursts from the screen wielding a sword that seems to flash in and out of the display. With another click, the 3D character shrinks back to his old 2D self. “Hey presto,” Yewdall grins, “we can play the animation in 3D.”
The screen is specially modified so it can show 3D images, but what is really new here is the software. Just as you adjust the contrast, colour and brightness on a screen, Yewdall’s 3D “volume control” lets you add depth and dimension to a 2D image in real time. And it doesn’t just work on animated films. His software will transform everything from real movie stars to the gun-toting characters in shoot-’em-up computer games.
This software is a massive step towards making 3D a reality, offering a quick and cheap way to create movies and effects that leap out of the screen at you. What’s more, the technology for showing 3D images has undergone a quiet revolution. In Japan, electronics giant Sharp has already sold over a million mobile phones equipped with tiny 3D screens designed to give you a multidimensional view without the usual clumsy 3D goggles. And just weeks ago, Sharp announced the release of a notebook computer equipped with a larger version of this 3D screen. Other manufacturers, such as NEC, say they plan to do the same.
Advertisement
The enthusiasm is catching. More than 120 companies from every corner of the electronics and computer industry are joining forces to push 3D technology from the lab to the high street. The consortium estimates that by 2008, the market for gadgets equipped with 3D screens – mainly mobile phones, handheld devices and computers – will be £17 billion. There are even whispers about 3D TV. This time, experts say, the technology is here to stay.
But haven’t we heard it all before? John Logie Baird first toyed with the idea of 3D TV in the early days of television, and the fad has surfaced like clockwork every decade or so since. The craze peaked in the 1950s with the 3D-specs genre. Back then it was driven by Hollywood’s desire to win back audiences lost to television. A flood of 3D B-movie classics such as House of Wax and John Wayne’s Hondo appeared in cinemas, while a series of experimental broadcasts were aired on television in Los Angeles and Mexico. None was a big success, and although one-offs have continued to appear, the technology has been languishing ever since.
Part of the difficulty is the complexity involved in making and showing these films. Each of our eyes sees a slightly different view, and our brain processes the discrepancies between them to give us a picture of our surroundings in three dimensions. The usual way to mimic this onscreen is to shoot a film using two different cameras and then project the two images together in two different colours that overlap. Viewers then wear a pair of glasses with a different filter in each lens. The left lens only allows through the image from one camera and the right lens lets through the picture from the other. This convinces the brain that the image is actually three-dimensional (see Graphic).
There’s one clear drawback: people don’t want to wear clumsy, uncomfortable glasses. For 3D to take off it needs to ditch the old image and lose the specs. And now audiences can do just that. After 10 years of research, engineers at Sharp’s European research labs in Oxford developed the spec-free 3D display that is already proving such a hit with consumers in Japan.
Sharp’s device is based on a clever use of a pair of liquid crystal displays. A standard LCD is built from a layer of liquid crystals sandwiched between two polarising filters that are set at right angles to each other. Usually these crossed polarisers block light from the rear of the display, but depending on their orientation the liquid crystals can alter the polarisation of light passing between the filters. The liquid crystals are normally arranged in microscopic spirals that rotate the polarisation of light by 90 degrees so that it can pass through the front filter. But if the liquid crystals are forced to lie parallel, on the other hand, the polarisation of the light is unchanged and the screen stays dark.
To create text or pictures on a display, the screen is covered by a grid of transparent electrodes that divide it up into thousands or millions of pixels. The liquid crystals in a pixel can then be made to lie parallel by applying a small voltage across them. Add red, green or blue filters to each pixel and you can create colour images.
To turn a standard LCD into a 3D screen, engineers at Sharp added a second, modified LCD. It uses a liquid-crystal layer and a special polariser called a retardation film to create a series of transparent vertical stripes about 60 micrometres wide separated by opaque stripes about 120 micrometres wide. Sharp calls these stripes a “parallax barrier” (see Figure).
The barrier is designed to control which pixels on the rear LCD are visible to each eye. So the image destined for the left eye can be displayed on the rear LCD pixels in such a way that the opaque stripes prevent the right eye seeing it, while allowing it to reach the left. Similarly, the right-eye image can’t be seen by the left. And this ingenious design has another trick up its sleeve. Switching off the parallax barrier transforms the 3D display into a normal display showing 2D images.
This could be a very cunning marketing move: a dual-purpose screen that changes to suit the viewer might just have what it takes to persuade consumers to part with their cash. Sharp has been quick to exploit that with the 3D mobile phones, which it launched by teaming up with Japan’s biggest wireless provider, NTT DoCoMo. Consumers who buy the latest generation Sharp camera-phone can take snaps, use built-in software to convert them to 3D, and then swap them with friends who have the same phone. They can also see games and screen wallpaper in 3D, and there’s always the option to use 2D mode like any standard display.
The success of these switchable screens has persuaded Sharp to create even larger displays. This month, for example, it announced the release of a notebook computer equipped with its 3D screen, and says more will follow.
Other companies are jostling for a slice of the 3D cake. NEC has announced a notebook computer with a 3D screen that will go on sale later this year. And last March several manufacturers, including Sharp, NEC and Sanyo, joined forces to promote 3D technology and thrash out a set of industry standards to help it take off. This makes sound business sense. The last thing a burgeoning 3D industry wants is a repeat of the VHS vs Betamax showdown, when two competing formats for video recorders were released – and with over 120 companies already in the consortium, consensus is vital.
Into the third dimension
But surely the $64,000 question is: when can we start watching 3D TV? It is still the ultimate goal, admits Ian Thompson, business development manager at Sharp Labs. “There are just a few wrinkles to iron out first.”
The problem is that viewers have to be in just the right place to see the 3D effect. With Sharp’s screen, for example, 3D images are only visible within a “sweet spot” that is fixed by the designers at, say, 50 centimetres from the centre of the screen. The image deteriorates if you move as little as 3 centimetres outside this zone. This means the screen can only be used by one viewer at a time, which is why most companies are concentrating on putting 3D screens into devices used by a single viewer, such as phones and notebook computers, and in specialist applications in engineering, medicine and design.
Researchers at Sanyo have tackled the issue head on with what they call a “multi-view 3D system”. Their screen functions much like Sharp’s, but instead of projecting just two images, Sanyo’s screen uses four. The result is that up to 10 viewers can enjoy 3D simultaneously. However, this reduces the screen’s resolution dramatically – by up to 40 per cent.
Improving the resolution of the LCD screen could help offset this, but it’s all useless unless film makers and TV companies develop the 3D content to watch on these devices. And right now, large studios are not prepared to step in until the screens are up to scratch.
It is here that Yewdall believes his company can help. Yewdall’s team has created software that bypasses the two-lens camera method traditionally used to create 3D images. It instead uses sophisticated image-analysis software to recreate the depth information lost when a scene is filmed with a standard one-lens camera. The team has already demonstrated the results on sequences from Jurassic Park and animations like Antz.
To process a particular scene, the film is digitised and then the first and last frames of the scene are extracted. Every object in the two frames is labelled with shades of grey – objects in the foreground are white while those in the background are black, for example – and an intelligent algorithm then analyses the two scenes and learns the depth relationships between objects. Once it has the depth information for the first and last frames, the software then creates depth maps for all the intermediate frames. You end up with the 2D image unchanged, says Yewdall, but you also know exactly where every pixel should be positioned from the viewer’s perspective.
All this takes time – about 30 hours for every minute of footage processed. But the advantage of Yewdall’s system is that once the depth information has been compressed, it only adds 5 per cent to the size of the signal. So 2D pictures and 3D data could be broadcast simultaneously, he suggests. Those with 3D televisions will be able to decode the additional data and enjoy the multidimensional experience, while those without could watch as normal. In other words, consumers can buy into the technology at a pace that suits them.
While he waits for screen technology to catch up, Yewdall is turning his attention to computer games. In the same way that gamers’ thirst for faster processors has driven developments in the PC industry, he suspects that a desire for 3D games may be the making of this new technology. And since most high-end computers come with powerful graphics cards built in, Yewdall saw an opportunity to put his system to the test.
To kick things off, he has designed a piece of software called Visualizer that can convert computer-generated animations or games in real time. Games already have depth information coded into them. Visualizer sits between the games software and the graphics card and looks at the commands that are being sent to the card. “It says, I understand what you are trying to draw, but I am going to ask you to draw a bit extra,” Yewdall says. The result is real-time 3D for games.
At the moment, Yewdall’s system is used for ad campaigns and trade shows, where companies want to make a splash. The wow factor comes at a price – $22,000 – but as the efforts of the consortium make 3D systems cheaper, Yewdall believes anyone will be able to enjoy the experience. As LCD costs fall, for example, you may soon need to pay only 50 per cent extra for 3D capability on a screen.
However, the main stumbling block for 3D is proving to be the human eye. In the real world, our eyes work together to create a stereoscopic view. The muscles around our lenses tense or relax to focus on objects at different distances, in a process called accommodation. With 3D images there is no need to alter focus as the screen is always a fixed distance away, but our brain tells our eyes to try anyway. This can lead to eye fatigue and headaches.
The way our eyes converge on a point in a 3D image can also cause trouble. As we look about, our eyes change their convergence depending on how far away an object appears to be. But when you try to show 3D images on a flat screen there’s a hitch. Although your brain may tell you a character is coming out of the screen towards you, where should your eyes converge, on the character or on the screen?
For 3D films, the optimum point of convergence is usually decided before shooting, based on an average screen size. So the director might put in a point of convergence based on a 1-metre screen, but if a viewer’s screen is twice that size, the point of convergence moves even closer to you, forcing you to cross your eyes more. In other words, the character seems to be right in your face. “The different screens viewed at the same distance would normally have different convergence points, but the director can only provide one,” says Mark Schubin, a technology consultant to the film and television industry based in New York and a 3D fanatic. Even pulling your chair closer to the screen changes the point of convergence, he says. Viewers could end up cross-eyed after little more than 30 minutes’ viewing.
Sharp and other companies in the 3D consortium are clearly taking the problem of eye fatigue to heart. Sharp and Digital Dynamic Depth have already created software that viewers can use to control the effective depth in 3D films. “These have been developed to allow comfortable viewing,” says Thompson. The consortium has also set up a working group to look for ways to make viewing more comfortable.
It certainly looks like the industry means it this time, but after so many years of hype and let-downs, can companies convince viewers that 3D is more than just a gimmick? They clearly believe they can, but Mark Schubin is less convinced. “There is always an appetite for 3D in the same way that there is always an appetite for a rich dessert,” he says. “But just as we don’t subsist on rich desserts, it is extremely unlikely that we will watch all of our television in 3D.” Only time will tell whether – nearly 80 years since the invention of television – viewers and entertainment bosses are finally willing to think outside the box.
