Your browser is not supported. Please update it.

2 April 2008

Dish and Qualcomm duke it out in 700MHz TV; AT&T’s options still open

Just what satellite operator Dish Networks/EchoStar thinks it is up to buying one-way 700MHz spectrum that is just right for mobile TV terrestrial broadcasting, we’re not quite sure, but we have one outlandish idea for it. By combining the 700MHz delivery from the ground with a satellite delivered parallel signal, somewhat like the system EchoStar is helping the Chinese authorities build, it could come up with a high penetration, low cost network build, or a service that is more efficient than any of the existing technologies – in essence a version of the Chinese CMMB STiMi technology. Maybe.

The only two auction winners (see separate item) that are specifically addressing mobile TV (one of the ‘sweet spot’ applications for 700MHz) are Dish Networks and Qualcomm, which spent $712m and $558m respectively for unpaired 6MHz spectrum, carving up the video-oriented E Block between them. Unpaired spectrum doesn’t fit easily into the existing cellular network model, at least for now, so it is more likely that this spectrum will be used for broadcast, and the bet is on that it will be focused on mobile TV.

Qualcomm already has its MediaFLO mobile TV network in channel 55, so buying a piece of E block, which is channel 56, seems to make perfect sense, except that it has only won it in the larger commercial centres like Los Angeles, New York, Philadelphia, Boston, and San Francisco, and has probably overpaid. The rest of it has gone to the Dish Network. Block E almost seemed earmarked for Qualcomm, but Dish seems to have spoilt the party.

If you believe that both of the major cellular operators are comfortable with the concept and business model for mobile TV (and we don’t think they are), then you might easily believe that AT&T bought 12MHz of unencumbered 700MHz spectrum from Aloha partners for $2.5bn – a deal which only completed in February – and then bought a further 12MHz of spectrum in 176 markets for $6.6bn at auction, and will eventually build two networks, one a cellular network perhaps using LTE technology to augment its cellular holdings and introduce mobile web, and the other to send out mobile TV.

If, on the other hand, you believe that AT&T is a sceptic on the subject of mobile TV, then it may be using ALL of that restriction-free 24MHz of spectrum on two-way wireless networks, giving it more capacity than Verizon. Then perhaps it will also resume negotiations and eventually buy Dish Network, and acquire its 700MHz spectrum in the process, leaving it a 6MHz slice to compete directly with MediaFLO, but with its own network, probably using DVB-H technology.

Our take is that AT&T remains undecided on what it will do with all this spectrum, having multiple exploitation plans, which depend as much on what its arch-rival Verizon does, and how well its MediaFLO service is accepted, as it does on the price it has to pay for bandwidth. The one thing we can say for sure is that AT&T and Dish are not supposed to have colluded, or remained in acquisition talks, while bidding for spectrum. And that means that they will each have a plan of how to proceed with the spectrum they have, as separate businesses, which they may remain, despite persistent rumours.

It is not impossible, as we said earlier, for 18MHz to be used by AT&T cellular devices, and just 6MHz used to apply to mobile video services. In fact, there are many ways you can cut this 700MHz spectrum. But what bothers us about this is that AT&T is most worried at auction time, not about introducing new services, but about keeping Verizon at bay. It doesn’t want to give it an inch, and preferably it wants to get one over on it.

That is the entire rationale of AT&T signing a deal with MediaFLO, not because it has chosen its mobile TV technology, but that it has to have an offering that can go head-to-head with Verizon in any market, because the MediaFLO offering it has described (and which is said to be on the point of launch) is virtually the same as Verizon VCast TV.

And any view of AT&T and its spectrum usage needs to be tempered by considering in detail the other spectrum that it holds, like all that Bellsouth 2.3MHz bandwidth, which was supposed to be built out in WiMAX.

Of course, AT&T could simply bypass specialist broadcast video networks and go straight to a really fat wireless internet pipe to the handset, over which web video could be viewed. But then that’s more like the open network that Verizon will have to build, so why pay more for the spectrum? Still, anything’s possible. So right now we won’t pretend to reach a conclusion on how AT&T will use its brand new spectrum, let’s just wait and watch developments. However at Qualcomm and Dish, we might be able to get closer to their intentions.

Qualcomm is straightforward. It has an almost-national license for one 6MHz slice, and now in major metropolitan areas it can add another. That gives it 12MHz to use in each of Los Angeles, New York, Philadelphia, Boston, and San Francisco, and the implication is that with very little change to its existing MediaFLO network transmission equipment, it an offer double the number of channels in those markets as it can elsewhere.

The big question this raises is double ‘what’ number of channels. Currently the Verizon VCast service, supplied by MediaFLO, has just eight TV channels. AT&T is due to launch it any day now with the same eight channels plus one or two more, of its own devising. Qualcomm claims that its system can deliver 20 TV channels in 6MHz. Well if that were true, then before it ever bought any more spectrum, surely it would be offering more TV channels. It is a very different business proposition offering two chunks of eight or 10 TV channels, with a full bouquet of just 16 to 20 channels ‘ it is another entirely to be able to carry 40 TV channels in some metropolitan areas, but only have eight used up so far. Which is it?

We tend to believe that 20 channels IS the capacity of MediaFLO in 6MHz, and what is holding back Qualcomm isn’t so much the lack of capacity, as the lack of content that broadcasters are prepared to make available. Virtually ALL of the MediaFLO content is ‘made for mobile’ shortened or abbreviated material such as Fox Mobile or ESPN Mobile ‘ channels that were put together for cellular streaming services, and which have yet to accept that viewing times on broadcast mobile TV are 20 to 30 times longer than on cellular streaming and require entirely different content, more like traditional TV channels.

But if Qualcomm had a problem contracting enough of the right content for MediaFLO version 1, then with double the capacity its problem would be doubled. Certainly the lack of enough of the right kinds of TV channels has affected the uptake of VCast TV, and will likely have the same effect at AT&T if it ever gets around to launching a TV service based on MediaFLO.

None of which bothers Dish in the slightest. It already carries cable networks and free to air networks and has carriage agreements to send them virtually anywhere in the US, so it has all the content it needs. What it doesn’t have is any experience building terrestrial networks, and given that its spectrum purchases are not entirely nationwide, there doesn’t seem any way that this footprint can be served by a satellite, without it spilling over and interfering with Qualcomm’s signal.

Actually there’s one way it could do that, not that it currently has any transponders that we know of that could send a 700MHz signal, and that’s if it was sending exactly the ‘same’ signal as Qualcomm. We can’t see this happening, and we shouldn’t really put it forward as an idea but if Dish was to use the FLO technology for a mobile TV service, and perhaps send some of the same TV channels as the MediaFLO service, then their signals would augment each other. Of course that’s NOT what’s happening, because it would be so much easier for Dish to subcontract delivery of its content to the existing MediaFLO network, and once again it would require collusion between two separate bidders, and that can’t happen.

Other suggestions have been that Dish is planning to use the spectrum to somehow support its Sling Media acquisition, but this is of course not likely. First off, Sling is owned by EchoStar, now at arm’s length from Dish. They will help each other, but Dish helping EchoStar is like the tail wagging the dog. Secondly Sling needs a two way broadband line for one to one communication, which we already pointed out is not the strength of this particular piece of spectrum, and to make that any different Dish would have to spend billions of dollars in build-out costs, which it would give to Clearwire if it wanted that and had that kind of money.

And it’s this that has everyone baffled. The network that Dish builds has to be terrestrial in nature or risk interference with MediaFLO. Dish has no experience with terrestrial networks, but is suddenly prepared to get involved with them or has planned to partner someone who was not in auction, to help it build a US wide mobile TV network, with a likely cost of not much more than $600m-$700m. The build time would be around 18 months, and the spectrum would be ready in 13 months, so a partial network might be ready the day after the spectrum is switched over. That doesn’t quite have the right feel to it.

If AT&T goes on to buy Dish, as we know it has been considering for some time, it might take on such a mobile TV network as a service it can offer, or it might choose to offload it to Dish’s sister company, EchoStar, and perhaps later lease some of its services, so a future acquisition won’t badly affect either the price of Dish or the potential for the service.

All that remains is to work out what technology Dish is likely to use to build such a mobile TV network? Would it be FLO, DVB-H or something proprietary. It can’t use the approaching ATSC M/H standard because that’s not about putting mobile TV channels in one 6MHz slice together, but about incrementally adding them to each TV broadcast.

Well once again, if it was interested in a FLO network it could have had that in partnership with Qualcomm’s MediaFLO organisation. If it offers a DVB-H service, and remains unattached to AT&T for the foreseeable future, it risks AT&T coming out with its own DVB-H network. After all, like Dish, AT&T’s U-Verse IPTV service means that AT&T already has access to a few hundred channels of content.

What you have to look for with Dish is a technology which leverages from a satellite, but the use of S-DMB in Korea (where EchoStar is a shareholder) and DVB-SH (which EchoStar’s new investment Terrestar might compete with at McCaw company ICO Communications) in 700MHz would both mean risking interference with rival MediaFLO.

But when we remember that EchoStar (which provides services for Dish) has first hand experience of working on the Chinese CMMB STiMi standard, and look at how that works, and we can see immediately how a US wide mobile TV service might be pulled together with this spectrum. STiMi stands for satellite and terrestrial interactive multiservice infrastructure.

What that means is that in China they use two separate frequencies to deliver something that operates like a single frequency network. First they build the terrestrial frequency in 700 MHz, and spray it out from about 40 major cities. Then they build a 2.6GHz transmitter and send a signal which is harmonized at the data level with the 700MHz signal. If your receiver catches a tiny piece of one weak signal, and a tiny piece of the other, it can bring them together to compare and construct a single strong data stream. It’s the same principle as a single frequency network, but smarter, and using two frequencies.
The beauty of this system is that the satellite spectrum can come first, or be added later, and components that support this type of approach are just coming onto the market. US companies like NextWave do something very similar in their MXtv system for WiMAX TV (they use much closer spectrum, different channels of the same chunk of 2.5MHz, which are next to each other) and we are sure Alcatel ‘Lucent could be convinced to perhaps offer DVB-SH which works in two frequencies, not just one.

In China there will be three 700MHz channels, and around 30 TV channels delivered in this way, using its own codec technology, and its own design of mobile TV system. Now it would be a huge risk to take the Chinese technology and try to export it, because outside of China, some of it might be considered another company’s intellectual property. But when has that ever stopped Dish trying out a new idea ‘ which is perhaps why it is where it is with TiVo.

It is no coincidence that EchoStar has just gifted its L-Band spectrum to a new investment on mobile satellite, Terrestar, with some cash in return for a substantial shareholding. And during 2009 Terrestar has just the kind of satellite that might be required for this type of operation, ready to take off. In the mean time it can quietly build out in a few major markets, keeping the lid on its precise technology, and then suddenly, just as the war for Mobile TV takes off in the US, in mid-2009, it would have a significant presence, with proven satellite TV content. Also its build out would be cheaper, because it would only need to provide a weaker support signal from any terrestrial end.

And if in the meantime the company comes out with specialist devices for reception of placeshifting, and these also have receivers for such a mobile TV system already built in, that might be game set and match in mobile TV for Dish and EchoStar in the US.

This analysis was originally published in Faultline, Wireless Watch’s sister title, which offers weekly insights into the digital media markets and the economics of the quad play. Please contact Peter White on [email protected] for more details.