Startup bets on mobile video

With commitment from major handset makers to trial its chips

A small California, US based "fabless" chip design startup is going after a market which is enticing everyone from large semiconductor (chip) makers to major Hollywood studios "" digital video distribution to mobile handsets.

TrueSpan Semiconductor, the startup, has also found a second home, in Bangalore, to drive the development of two chips which it says will be better than the competition's products. And competition in this case are firms such as Texas Instruments, say top executives of TrueSpan.

The startup's founder and chief technology officer, Sanjai Kohli, has set his sights on the European market for digital video broadcast (DVB) that will go to commercial pilot trials next year. Kohli, an IIT Mumbai and Washington University educated engineer, built and sold two successful technology firms, one of which was SiRF, a GPS chipset and software operation. It is now traded on the Nasdaq and did sales of $120 million in fiscal 2004.

Kohli has now raised some $12 million from Tallwood Venture Capital, a relatively new but well respected fund. Tallwood, started by Dado Banatao, an ex-Mayfield venture partner, invests in early stage network infrastructure technology companies.

"We are developing chips for DVB H, which is a broadcast wireless standard for video distribution for mobile devices such as handsets, but can even go on laptops and media players."

This standard was ratified by the industry recently. It's a fairly new standard and there are no commercial deployments today, but he expects the first commercial deployments to happen by May 2006. That is also about the time when TrueSpan's two chips will be ready for trials.

"We decided to focus on the European standard, which we see as being deployed across pretty much the entire GSM footprint. We are developing two chips "" an RF (radio frequency) chip and a baseband MAC (media access control) chip, and all the firmware and the software that goes with the chips." The RF chip is otherwise known as the tuner and the MAC chip the demodulator.

TrueSpan's business model is that of the standard fabless semiconductor firm. It will design the chips, market them, have them fabricated or manufactured by contract foundries, in Taiwan or mainland China, and take possession of the parts after the manufacturing and sell them to the handset manufacturers.

When DVB H goes commercial, initially, it will be in a couple of cities, and "we see widespread demand by 2008-10," he says. Right now there are technical pilots in Pittsburgh in the US, and in Oxford, UK. The US pilot is being run by Crown Castle International, a US based network infrastructure firm, "which actually owns a lot of these cellular towers from which carriers hang their antennas," he says.

Commercial video distributors "are going to use the tower infrastructure plus the spectrum that they bought". Crown Castle, for instance, will itself deploy the network and distribute the content to the handsets, and the operators will become retailers. "That is quite different from the cellular model."

In Europe, a six month trial will start in June in Oxford; there is a trial in Berlin with Vodaphone and T-Mobile and there is a trial going on Finland. Japan and Korea have come up with their own exclusive standards, independent of each other. The Korean standard is actually commercially operational today. The Japanese standards will get deployed next year.

Kohli's second ace, in the race against the big boys, is "commitment by most of the major handset makers to trial our chips provided we meet their specifications". Most handset firms, with large volumes and therefore high risks, "don't want to deal with the startups."

But TrueSpan has "managed to get a fair amount of traction from not just the handset makers but also operators, for carriers are increasingly specifying what goes into a handset," Kohli says.

But what about 3G, the next generation of mobile phone standards? Kohli says, it actually complements DVB H, so the two technologies co-exist. There is also a technical difference between the two that makes the complementarity possible: A 3G network is a two way system." It sends you a packet, the phone responds and says I got the packet." The DVB H is a one way network, like radio.

So, it can support a lot more users. Take football games in Europe, where a lot of people go to the stadium but also carry small television sets. In such a situation, "there is no way you can support video-to-a-handset for 50,000 handsets on a 3G network". But it can be done over DVB H, for that standard "doesn't care whether one guy uses it or 10 million."

So, 3G and DVB H aren't mutually exclusive: The public content will go on DVB H but the premium content will go on 3G so they complement each other. And since the carriers already have a billing system in place, it makes things that much easier to work.

Crown Castle, for instance, has the ability to broadcast some 15 channels. The firm broadcasts nine basic channels and partners three carriers who will each specify two special channels. So out of the 15, a subscriber of one carrier will get 11 channels - nine general (on DVB H) and two special (on 3G). A different subscription will also get 11 but the two special channels will be different.

Standard content could be say CNN or Doordarshan. Premium content can come from the Hollywood studios that have set up "mobile divisions." They have realised that people are not going to watch 30 minute sitcoms on their mobile phones. So the studios are coming up with two to five minute sitcoms to be specifically aired for handset viewing.

Market research shows people are going to watch about 10 per cent of what they are going to see on the TV on the handsets. "So they would want to watch cricket, perhaps, until Tendulkar got out," says Vinod Gopinath, manager for technical marketing at TrueSpan.

Kohli says, ABI Research, a market researcher, forecasts the worldwide demand for "mobile television and digital video content services" will explode from under a million subscribers today to 250 million subscribers by 2010. So the market would grow from $200 million to $27 billion.