Video Conference Article

Real improvements on the way Future technologies by Fiona Harvey Published: February 14 2003 16:02 | Last Updated: February 14 2003 16:02

The problem with videoconferencing is that it is not very real. With most of today's systems, though the picture quality may have improved markedly from the early days of jerky movement and lips out of synch with the sound, it is still impossible to forget that you are staring at a screen with a representation of a person many miles away, rather than engaging in a natural conversation with someone in the same room. Some people shy away from videoconferencing because of its unreality: recent research from ACT Teleconferencing, in the UK, found that fears about how they would look inhibited many people from being comfortable with the technology. For this reason, computer scientists have been working on a new generation of videoconferencing that would go a stage further than faces on a screen and voices from a microphone. Their goal is tele-immersion - a technology that would create three dimensional real-time images of people that would be so real they could fool our senses into thinking our far-off interlocutors were in the same room. It sounds like something from the Holodeck of Star Trek. But with advances in compression technology, faster networks, and ever increasing processing power, researchers believe they will be able to create fully immersive virtual reality systems - though it might be two decades before they come into widespread commercial use, according to Jaron Lanier, a renowned virtual reality expert who heads up the US's National Tele-Immersion initiative. Tele-immersion systems are under development at several universities and research institutes around the world, but they tend to share the same characteristics. They take up entire rooms, where the subjects of the conference can be surveyed from a variety of angles by a system of digital cameras, with the information relayed to a central processing system capable of reassembling the images. As well as capturing the images through cameras, the systems may attach sensors to parts of the users' bodies, such as their heads and hands, which relay accurate information on their movements. As the amount of data such cameras and sensors collate can be huge, the systems use finely honed compression techniques, and need very fast networks. The systems also may demand users to wear special headsets or goggles, which present one image to one eye and a slightly different image to the other eye - mimicking the way that our eyes see objects from slightly different angles, so that the brain then interprets the image as three-dimensional. Some industry observers remain sceptical. Daniel Rasmus, vice-president at Giga Information Group, believes the technology is still much too intrusive: "The technology itself is distracting, because you have to have so much stuff to feel that you're immersed that you can't feel that you're immersed. Like you have to put on a pair of glasses, so that's a bit false." He says that current videoconferencing is good enough, and people use its shortcomings merely as an excuse because they do not wish to get to grips with technology. However, he agrees that the next generation of business people may be more ready to take

up fully immersive teleconferencing: "A 50-year-old executive today would not be willing to put on goggles, but 50-year-old managers 20 years from now will be willing to do it." Another major problem for immersive systems remains to be overcome: the sheer amount of brute processing force required to run these systems, which may generate data at the rate of terabits per second. Tele-immersion systems that require supercomputers seem unlikely to hit the mainstream soon. That is why a group of researchers at the University of Pennsylvania in the US wants to use the technology over a distributed network, using the same principles as "grid computing", which yokes many cheap low-power computers together to share their power. They say their work could deliver tele-immersion systems at a fraction of the price of current systems under development. However, Dr Lanier believes that the future of video conferencing on grid-style networks will be limited, because of latency, or the time lag between data being sent and its successful reassembly at the other end. He argues: "With any dispersed network like that, working across the internet, you're inevitably going to be looking at significant latency, however you slice it up." Latency will render real-time tele-immersion impossible, and make it difficult to avoid the lip synch problems that have plagued current video conferencing systems. Accordingly, he stakes his hopes instead on unified hardware systems with large amounts of processing power. At present, these may still be expensive, but as the cost of computer hardware continues to fall rapidly, while processing power improves, he believes it will become possible to develop such systems at a reasonable price. Yet one problem that remains intractable is the speed of light. No matter how good the networking technology, fibre optic cables still can only transmit data at the speed of light. Conferencing over very long distances is thus subject to a short delay, and any delay longer than about half a second can be picked up on by the human eye and ear. Thus the pretence of someone far away being in the same room may be impossible. In the meantime, several intermediate technologies promise to bring us closer to the goal of immersion. For instance, the European Union has brought together scientists from Sony, BT, and several universities in its Virtue project. Virtue stands for Virtual Team User Environment, and consists of a desk with a large plasma display, and four cameras mounted around it. The speaker must sit in front of the screen, and can conference with a maximum of two other people at the same time. The real images from the physical cameras are combined through an image-processing software system to be made to appear as if they were coming from virtual cameras, placed behind the screens. This placement means people can look straight into the screens to find what seems to them to be eye contact with their fellow speakers. Sony hopes to have products available based on the technology in about two to three years' time. Teleportec, a Manchester-based company, takes a more ambitious approach. Its technology creates life-size three-dimensional "holograms" of the speakers standing at lecterns or sitting at boardroom tables. A system of small cameras capture the subject, then transmit the digitised images to the remote location over a high speed line. At the receiving end, the images are funnelled through a beam splitter, then projected on to a pane of glass from several angles, which gives a near 3D effect, so that the speaker's image appears to hang in the air behind the podium, before the audience. With high speed networks allowing a good data rate, alongside the company's compression technology, there should be little problem with lip synch, according to the

company. The system is already in use in several schools, enabling virtual teachers to take classes, and in companies including UPS, Compass and Sky TV. Likewise, Texas-based Zebra Imaging specialises in creating 3D holographic images. These have been used to create virtual 3D models of cars for Ford Motor Company, allowing executives in remote locations to collaborate in examining virtual new car models. If these intermediate technologies turn out to be good enough, will we still need teleimmersion? Dr Lanier believes so: "The technology to make it possible is coming together, and the experience will be so rich that people will demand it. As people get more used to video conferencing in various forms, their appetite for tele-immersion will only grow." © Copyright The Financial Times Ltd 2003. "FT" and "Financial Times" are trademarks of the Financial Times.