So far there has been an eerie silence over the penetration and acceptance of the only MediaFLO live mobile TV service, that from Verizon in the US, marketed as VCast TV. We have assumed it meant that the service was failing or at least reaching fewer numbers than expected.
So when Faultline asked Qualcomm for some guidance it was nice to end up chatting with MediaFLO CTO Kent Walker and Omar Javaid, Senior Director of Business Development at MediaFLO this week.
The big question, the one which decides whether or not MediaFLO becomes a $20 billion behemoth, leading paid US mobile TV services for the next decade, is the number of people that have so far taken the service, and Javaid is not slow to say that this number is secret and that it belongs to Verizon. The other things that Qualcomm won’t address are the legal action with Nokia over this technology and the precise reasons why Verizon only offers 8 TV channels on its service right now. But that’s about the end of the coyness.
MediaFLO is now rolled out in 40 markets across the US, whereas it started life with 19 and Verizon is subsidizing virtually all the handsets, so this service is unlikely to encounter the problems with uptake that we have seen in places like Italy (3 Italia DOES subsidize and has 750,000 customers while Vodafone and Telecom Italia Mobile don’t and have just 10,000 each). It is also scheduled to be launched by AT&T in Q1 of 2008.
The market rumor is that Verizon has only low tens of thousands of customers for VCast TV, and it hasn’t been strong on advertising the service. However Verizon was very slow in the early days of its FiOS IPTV service, but that is now a raging success, so the same process may be happening here, with the early customers really an extension of market trials, and the first few months being all about tuning the offering.
‘The acceptance has exceeded our expectations,’ said Javaid. ‘We thought customers would be snacking on short form content. Everybody told us to expect only short sessions, but that’s not what’s happening. People want TV they know and already use, and they are watching in 30 to 40 minutes bursts, and taking in full length TV shows.’
In a way it’s amazing that anyone still thinks that mobile TV is about ‘snacking.’ That was purely a phenomena of the early streaming services, where its quality was so low that to watch it for any longer, risked your eyesight. All DVB-H trials around the world found similar outcomes.
‘Viewing is evenly spread,’ said Javaid, ‘and peak times are like traditional TV, except that we are seeing a late night peak between 11.00 pm and 2.00 am in the morning.’
Interestingly the morning rush hour peak that is witnessed in Europe is missing, as many workers travel to their place of work on public transport and watch TV then, whereas in the US, there is far more travel by car in the mornings. There is also no post workday peak as there is in Europe.
‘The demographic is younger too and across both sexes,’ says Javaid. ‘We had expected it to be more male dominated and in the 18 to 34 age group, but this is from below 18 as well as from 18 to 65.’
Over the past few years sniffing around radio engineers among rivals and partners for MediaFLO the word has been clear, that the actual live performance of MediaFLO and DVB-H ‘ two violently opposed rival standards for control of the mobile TV agenda, are more or less, dollar per bit, identical. The only company that continues to disagree with this statement is Qualcomm.
You can almost hear Qualcomm VP of Technology and CTO for MediaFLO Kent Walker, shrugging at the other end of the phone. You can certainly hear him sigh. ‘There have been 4 or 5 trials of MediaFLO where we have confirmed it, that with the same spectrum, the same transmitter power and the same bits per second per hertz, there is a 4.5 dB link margin benefit in favor of MediaFLO,’ and he goes on to list the three main differences between the two.
‘The key difference is the use of Turbo code in the inner code. When DVB-T was created over a decade ago, Turbo coding wasn’t possible. Now it is the combination of Soft Decision Codes which means that each channel in OFDM gets less multipath interference. Soft decision codes take the transform of the signal, and flatten the spectrum out by rating whether or not the signal is any good or if it has interference. It can place more trust in the channels with better signal to noise ratios. Turbo codes were an academic study when DVB-T was designed and DVB-H has inherited its form of OFDM. Since then it has been used in 1ex EV DO handsets and its now pretty well understood.’
Another difference claims Walker is the fact that DVB-H bursts less often and for longer, which means it is more likely to have to use error correction codes to reconstruct the signal, and that has an effect on battery life ‘ MediaFLO.
But we can remember back to demonstrations of the superiority of IBM’s OS2 operating system over Windows, and could bring up a hundred other instances where supposedly superior technology did not win a market war in the technology sector.
The key question for MediaFLO is its lack of acceptance outside of the US, and despite having public trials at PCCW in Hong Kong, BSkyB in the UK, KDDI in Japan, and with Taiwan Television, and Maxis and Astro in Malaysia, there is little prospect of these coming to market first within any given country, complete with spectrum, and certainly not with a network build commitment from Qualcomm, which is what won the day for Qualcomm in the US.
PCCW is wedded to launching MBMS services next year, although it will perhaps also use a broadcast technology alongside it; BSkyB has no spectrum and there is only L Band spectrum to be auctioned there; KDDI has experienced the huge success of ISDB-T in Japan, with around 18 million mobile TV handsets shipped. Both Taiwan and Malaysia have long lived DVB-H trails with commitments from the local cellcos. So outside the US it becomes apparent that MediaFLO has had to chase the 5 years lead that DVB-H had by going after broadcaster dominated trials for the most part, instead of the leading cellular operators that were already too far down one track to change strategy.
‘We have a few more trials out there that are not public’ Javaid reminds us, but really the big questions for MediaFLO are whether or not it can maintain dominance in its native US.
Here there are specific challenges not the least of which is Nokia’s legal challenge over intellectual property, which is an across the board Qualcomm problem anyway. Separately will DVB-H or DVB-SH ever gain a foothold in the US? and finally public perception of the usefulness of mobile TV and the fact that the cellular operators there don’t seem to believe that there is a true business model, at least not for paid mobile TV.
If that is the case then the mobile TV market could be left to companies like the existing Satellite Radio companies, already stretching their offering to include video or to the local broadcast TV owners which are relying on the new ATSC M/H standard. The cellular operators could put in the least amount of effort to develop mobile TV, let the broadcasters carry the brunt of the popularizing work and then step in with paid TV variants once the case is proved. But that could mean a three or four year wait.
Interestingly Kent Walker attends ASTC committee meetings, the organization which is in the process of selecting a standard for this, and Qualcomm, like Nokia, has one of the 8 submissions to the ATSC committee back in June, the others being from the original ATSC developer LG subsidiary Zenith, and from Coding Technologies, Coherent Logix, DTS, the Mobile DTV Alliance, Micronas Semiconductor, Nokia and Thomson
‘The ATSC is working on its own physical layer and that cannot be the same as MediaFLO, but the upper layers are a different matter, and it could be interesting to see upper layer commonality between the ATSC work and MediaFLO,’ which is the exact same rationale behind Nokia’s push to make the ATSC M/H standard upper layers similar to DVB-H which Faultline revealed two weeks ago.
‘Do I see a handset in the future that can offer MediaFLO and ATSC on the same device ‘ certainly I see that,’ says Walker. But what if Nokia wins that battle to dominate the upper layers of organization on ATSC M/H standard?
‘It’s just a matter of convenience really. It would be more convenient if MediaFLO won that battle, but it doesn’t make it unlikely that a MediaFLO handset or a DVB-H handset or both, will emerge in the US market regardless of which way ATSC goes.’
The ATSC is working towards placing a mobile TV signal into the existing SD digital broadcasting channel, and wants it in available equipment in time for when the 700 MHz signals in the US are freed up from analog use in February 2009. So unless MediaFLO can unlock the volume pay TV market for mobile by then, it looks like the US mobile TV avalanche may be put off until well into 2009.
However it is important not to get confused between MediaFLO and Qualcomm here. Not all of Qualcomm’s mobile TV eggs are in one basket. It has built a chip which offers MediaFLO, but also ISDB-T and DVB-H all on the same chip, and it can chase a share of any mobile TV market purely at the chip level. In which case, it is almost bound to add an ATSC chip to the mix, so that it can also get a piece of the free to air ATSC action, if and when it comes around in early 2009.