Seeing the Light

There can be little more dour a prospect than a combination of maths and Cambridge. Luckily for the games industry, fledgling middleware developer Geomerics is about the most exciting possible product of the two. The outfit announced this week the appointment of Gary Lewis - a business heavyweight who was most recently Take-Two's COO in New York - as chief executive, giving the clearest indication yet that success won't be a long time coming for founder Chris Doran and his team.

Doran is a leading research scientist with over 15 years experience in applied mathematics and theoretical physics. He's a Cambridge die-hard. He is the author of a major book on geometry and physics and of over 50 papers on subjects including quantum theory, gravitation, and computational geometry.

Doran isn't messing about. He's created what is apparently the first true real-time radiosity solution ("Simply put, radiosity equals 'everything lights everything else'," we're told) for videogames developers - and the bottom line is that Lewis reckons he's onto a winner.

Speaking at London's Science Museum, just before Doran gives a speech about the company's first product, Lewis says, "I think Geomerics offers a huge opportunity within in the industry. I believe it's part of a new breed of companies within the industry.

"We're not a traditional type of middleware company, and what we're going to do is try to deliver different applications to developers and publishers that haven't been done before."

Lewis is no small catch. During his eight year stint at Take-Two, which ended just recently, he took on roles such as president of international in London and global COO in New York. His experience in bringing products to market is perfect for Geomerics, and there's no illusion as to why he's been brought on board: he's there to make the money.

"My expertise is commercial business, and I'm really going to use that," says Lewis. "What we have to do is to clearly explain to people what we can offer the games industry and be different. That's what it's about. We have to be unique and we have to work with the partners and solve areas that they haven't been able to do before."

It may seem odd that someone with Lewis's credentials has swapped Manhattan and a top slot at one of the world's biggest publishers for the job of fronting a tiny developer in Cambridge. Odd, that is, until you see what Doran and the rest of the Geomerics team have been up to for the past 18 months.

Lighting the current generation

Doran's Science Museum talk focuses on how geometric algebra can impact low-level programming and how lighting in games is likely to evolve in the coming years. He shows a video of Geomerics' real-time radiosity solution in action - and Lewis's career choice suddenly doesn't seem so outlandish.

"It's realistic lighting," says Doran, explaining to us in words of one syllable a development solution that is in all probability comparable to rocket science. "If you look round a building like this you see just how complex the lighting is to get right if you want to make a scene look realistic. We've taken a major step in that direction."

So what does that actually mean for a real game? "A typical game might be a first-person shooter, where you're exploring a corridor with a torch," he says. "As you shine the torch at the wall the rest of the room should illuminate, but currently in games it doesn't. The eyes immediately cotton on to the fact that the lighting's not realistic.

"There's this curious thing in a game; it only looks as good as the least realistic thing in it. If you've got this mish-mash between something that's extremely high fidelity and things that look completely unrealistic it actually becomes quite an unpleasant experience. You actually need quite a consistent level all the way through.

"With the development of tools and the art pipeline, you can make pretty detailed geometry and textures, but the lighting often looks very poor. It's not realistic, it's not dynamic, characters don't look right as they move through a scene. The eye always picks up on the most unrealistic thing."

Part of Lewis's first job as CEO is to brand the real-time radiosity product and take it to GDC. There's no news as yet on what the brand itself will be, but Doran is confident the first customers to take advantage of the code will be instantly recognisable.

"I think a lot of [the major developers] might just buy it from us," says Doran, grinning. "For the big studios the economics will still work out in favour of buying from us, because to crack the problem to the depth that we have, they would have to take a lot of their top people and put them on the work program for six to 12 months."

And Doran's work is hardly going to stop with lighting. Once the real-time radiosity solution's out of the door, the next release for Geomerics will be based on animation.

"Our current research is on an animation product to follow on from this," he says. "And that will be a product that does the same for animation that this product does for lighting. It'll be much more realistic, and freely embeds characters in the scene so they're correctly shadowed and lit and all the skinning effects are all sorted.

"The goal with that product is to take motion data from motion capture, or whatever, and to plug our product straight in to give you the character moving realistically around the scene."

All indications are that Geomerics is going to be around for some time to come. Lewis is certainly bullish about the company's prospects. Looks like space-age maths and old English university towns aren't quite so dull after all.

"I am very excited about my involvement in Geomerics," he says. "Being successful in the games industry is about identifying opportunities and approaching these areas differently to others. This results in delivering a greater gaming experience to the end user, and this is exactly what we believe our technology will do."