Your browser is not supported. Please update it.

14 September 2005

DVB-H gulps before leap of faith to build out Cell TV networks

For a man that is likely to be building a UK-wide DVB-H network for delivering TV to cellular phones, Simon Mason of Arqiva at this week’s IBC conference, made it sound like he’s not quite sure of his technology choice.

Mason brought the entire argument about DAB derived mobile video technologies such as DAB enhanced packet mode and T-DMB and DVB-H down to just how many channels operators want. DAB does a good job of offering up to 5, 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, that route could be very much cheaper.

Mason’s job at Arqiva is to make its Oxford DVB-H trial work, and to learn as much as he can about what that means for a nationwide network.
He looked at three alternatives in spectrum terms and showed the Mobile TV audience at IBC how Band III in 220 Mhz, UHF in 600 MHz and L-Band in 1470 MHz spectrum would likely 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 7.6 MHz of bandwidth, might be needed 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 a day of contradictory advertising for the 4 main protagonists ‘ 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 Mobile TV standards, but Mason reminded everyone that just because T-DMB was being used in Korea, it was invented in Europe as part of the Eureka Project (European funded research) and that far from being a proprietary extension, it was just another layer placed directly on top of DAB and he pointed us to the diagram below (in pdf verion).

Dr. Chan Yeob Yeun of LG, from Korea almost turned the event into a product parade, three times taking his TV phone from Korea ‘where we already have it working’ out of his pocket and swivelling the screen into 4:3 mode.

Technically Yeun’s presentation was without value, but he highlighted a key aspect of the future of cellular TV, the fact that broadcasters and cellular operators both want to own it and that potentially, they speak different languages and do not share the same aims.

This was later highlighted speaking to Qualcomm’s Rob Chandock, VP engineering and market development of MediaFLO, when he told Faultline, ‘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.

Yeun also claimed that DMB costs 5 times less than DVB-H to build out and offers channel change times of 2 to 3 seconds against DVB-H which was 6 to 10 seconds.

Ulrich Reimers, the chairman of the DVB technical module, was clearly incensed at the blatant salesmanship of the LG speaker and said so, and went on to talk about the high speed ability of DVB-T car based TV receivers and to claim that DVB-H was capable of the same, and that it would be highly compatible with DVB-T.

But apart from using the odd existing DVB-T tower and perhaps sharing the space on the head end encoding processors, he never really addressed the issues thrown up by any of the previous speakers, especially 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 250 Kbps, and that its powerful time slicing, where the phone is using battery for receiving for only 0.4 seconds out of every 4 seconds, means that it was battery efficient and ready to roll. He said that it could even manage a 6 mbps shared signal at Mach 1.0, implying jokingly that it could even be used on an airplane.

Reimers presided over a press conference later that day where 4 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 in the form of Ray DeRenzo, VP Business Development for EMEA, held up the end of cellular delivered Mobile TV, speaking to Faultline about its recent deal at Orange in the UK.

The company offers a software download that can work in either in pure 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 and then 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 to 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 to 50 kbps.

‘Broadcasting to a mobile phone won’t kill off the use of the cellular channel for delivering video,’ says DeRenzo and although he may be convinced, I’m not sure Faultline is, unless the cellular delivered services improve dramatically in data rates and that’s a decision for the operators.

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

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

He went on ‘DAB benefits from 1 to 5 MHz 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.’

‘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 6 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 gradually vacated by analog TV.

‘In Europe,’ says Chandhok, ‘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’s one voice singing against the choral hoards that have voiced support for DVB-H, but then Chandhok tilts an un-launched 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.