Your browser is not supported. Please update it.

16 September 2005

The four options for mobile television set out their stalls

For a man who is likely to be building a UK-wide DVB-H (Digital Video Broadcasting-Handheld) network for delivering TV to cellular phones, Simon Mason of Arqiva, speaking at last week’s International Broadcasting Conference in Amsterdam, was surprisingly open handed about the technology choices.

Mason brought the entire debate over whether to deliver mobile television using DVB-H, or the alternative of technologies derived from the Digital Audio Broadcast (DAB) standard ‘ such as DAB enhanced packet mode and T-DMB – down simply to how many channels operators want to offer. DAB does a good job of offering up to five, he said, while DVB-H could take that up to 20.

But the main thrust of his enlightening speech was that it could take as many as 5,015 transmitter sites to cover an area the size of the UK just to bring mobile TV to 85% of all cellular subscribers using DVB-H. Since there are already thousands of DAB transmitters offering audio services already, that route could be very much cheaper.

Mason’s job at Arqiva is to make its Oxford-based DVB-H trial work, and to learn as much as he can about what that means for a nationwide network. In his presentation, he looked at three alternatives in spectrum terms and showed the mobile TV audience at IBC how Band III in 220MHz, UHF in 600MHz and L-Band in 1470MHz spectrum would be likely to play out. He adjusted for height loss, receiver gain, the fact that live transmitter sites cannot always be found exactly where you need them and accounted for adjacent channel blocking, and took findings from real live trial data, and concluded that as many as 5,000 transmitters might be needed, using 7.6MHz of bandwidth, for good UK coverage.

He said that the only practical modulation scheme was QPSK and also called for regulators to be careful in considering their spectrum options, and to not even consider putting high powered handheld services next door to low powered services. The holes in coverage that the new service creates means that booster sites are needed as ‘make good’ transmitters for the weaker services, driving up costs further still.

The Oxford trial was conducted with O2 and Nokia and has recruited 350 participants between the ages of 18 and 45 to test the service over a six-month period, with Arqiva (once part of NTL) building just eight DVB-H transmitters to broadcast 16 channels.

There was silence after Mason’s speech, but it wasn’t because people found it boring. He had injected some much needed reality into the proceedings, which had threatened to become a half day of contradictory advertising for the four main protagonists in mobile TV ‘ DVB-H, DAB Enhanced Packet mode, T-DMB and Qualcomm’s proprietary, but startlingly clear, MediaFLO.

Given that it was a European audience, more credibility was given to European standards, but Mason reminded everyone that, although T-DMB was being used in Korea, it was invented in Europe as part of the EU-funded Eureka Project and that, far from being a proprietary extension, it was just another layer placed directly on top of DAB.
Dr Chan Yeob Yeun of Korea’s LG followed with the key point that broadcasters and cellular operators both want to own mobile TV but that potentially, they speak different languages and do not share the same aims.

This was later highlighted speaking to Qualcomm’s Rob Chandock, vice president of engineering and market development for MediaFLO, who commented: ‘I visited an official in SARFT (State Administration of Radio Film and Television) in China. They didn’t want to know anything about MediaFLO because they thought it was simply cellular and nothing to do with them. Once they realized it was a broadcast system, the SARFT official said, ‘Then I must regulate it’.’

Yeun’s experience in Korea was that the broadcasters would take control of advertising and own much of the broadcast revenue, but that the cellular operators were handling billing and adopting aggressive m-commerce models built around product placement. ‘A viewer can flip the red button and then click to buy anything on the screen, what the actors are wearing, any items in the room, anything.’

This revenue clearly is going through the cellular operator in Korea where it is important that the services can automatically drop into time shifting (which implies quite a lot of disk or flash storage) so that the viewer misses nothing when conducting such a purchase transaction.

Trials for T-DMB have begun in Finland, Italy, France and Norway, he said, claiming that DMB costs five times less than DVB-H to build out and offers channel change times of 2-3 seconds against DVB-H at 6-10 seconds.

Ulrich Reimers, chairman of the DVB technical module, countered with the claims that DVB-H would be capable of the same performance as car receivers using its terrestrial cousin, DVB-T. But apart from the benefit of being able to use the odd existing DVB-T tower and perhaps sharing the space on the head end encoding processors, he never really addressed Mason’s points about the number of new transmitters that would be needed.

Instead he insisted that DVB-H was ready for introduction, that it could scale to 30 channels each using 250Kbps, and that its powerful time slicing, where the phone is using battery for receiving for only 0.4 seconds out of every four seconds, means that it is battery efficient and ready to roll. He said that it could even manage a 6Mbps shared signal at Mach 1.0, implying jokingly that it could even be used on an aeroplane.

Reimers presided over a press conference later that day where four German TV networks said that they would be committing content to DVB-H and he told the audience that the Germany top soccer division would be offering a trial in the lead-up to the soccer World Cup, to offer coverage over mobiles using DVB-H protected by Microsoft’s DRM system. Presumably that means there are limits on what type of phone they will run on, but it wasn’t mentioned.

MobiTV, represented by Ray DeRenzo, vice president of business development for EMEA, held up the end of the cellular-delivered television technology, which uses the basic 3G network. MobiTV recently signed a deal with Orange in the UK.The company offers a software download that can work in either in pure Java J2ME or on a Brew (Qualcomm) or Symbian (Nokia) platform, which plays video at between 10 and 15 frames per second. His company offers an end-to-end service from content aggregation and content ingest, sending video as an IP transport stream into a wireless operator’s IP gateway which the phone application transcodes, decodes and decrypts and decides how best to render to a particular mobile screen.

DeRenzo said, ‘The typical behaviour of a mobile TV viewer is to take 3-5 minute ‘snacks’ of video and over the entire month they might use 40 to 45 minutes of video. This gets better as you improve things like video quality and viewing experience, which you can do by moving to a UMTS phone or if you improve the electronic programming guide, but it doesn’t increase linearly. If you improve something 100% you might get a 25% improvement in viewing figures.’

The services that MobiTV provides for Orange are at 128Kbps, which we can see is around half the speed that DVB-H is targeting in Europe. But likewise MobiTV has delivered on GPRS networks at just 60Kbps and says that some operators want to use as little as 40-50Kbps.

‘We will end up with a hybrid with some TV delivered over the cellular link and some broadcast,’ concluded DeRenzo.

Perhaps Chandock of Qualcomm summed it up best when he said: ‘There’s no cookie cutter for this market.’

He went on: ‘DAB benefits from 1MHz to 5MHz delivery on existing transmitters, but its downside is scale. Elsewhere there are spectrum issues. I don’t think a lot of hard decisions about taking services to market have been made yet.’

Qualcomm insists that it is talking, if somewhat quietly, to all of the European operators and broadcasters. ‘They’d be crazy not to make sure that MediaFLO isn’t the right option for them. They have to at least look, and we never leave a potential customer without their technology group agreeing that MediaFLO is a better technical solution. The final decision is down to the political aspects of broadcast and wireless.’

He continued: ‘But don’t be fooled, most of the services people are talking about are QCIF (176 x 144 resolution). We don’t think that you can watch sport and read the score on anything less than QVGA (320 x 240). If DVB-H delivered that, it would only be able to deliver 10 to 12 channels, against our 20-plus. The standardization of MediaFLO will be announced in the next six weeks and the US service we are building will be up and running by October 2006.’

Qualcomm is gradually adding service as the spectrum it has bought in the US is vacated by analog TV. In Europe it has no such luxury as owning a spectrum itself to prove its service. Chandock said: ‘In Europ it is all going to depend on which players get access to spectrum and that’s still down to decisions that regulators have yet to make.’

Qualcomm doesn’t like to confess that it is one voice singing against the choral hordes that have voiced support for DVB-H – but then Chandock tilts an as-yet unlaunched Sling Shot concept phone over for us to view, running video from FLO. The resolution IS compellingly accurate, unlike anything else that is being shown at the show.

This analysis was originally published by our sister research service Faultline, which covers triple play models and broadband digital media markets.