Rumors of the demise of the Texas Instruments Hollywood mobile TV chip, are exaggerated, says Yoram Solomon, director of Strategic Marketing, Industry & Standards, Mobile Connectivity Solutions, at Texas Instruments.
It was a question that Faultline has raised in print and certainly raised with the TI people that attended IBC two weeks back.
‘I don’t think the development of the chip is on hold, it’s just that the market is on hold,’ is the first thing that Solomon says, who is also the TI point man for its involvement in the Mobile DTV Alliance, the pro-DVB-H club in the US.
Our own Faultline research says that if the market is on hold it’s about to take off, projecting something like 11 million mobile chips to ship this year, with just 50% of these being ISDB-T and T-DMB chips destined for Japan and Korea. Those are healthy chip shipments and the next few years will see a huge number of new services launched. By 2011 we expect 244 million handsets to be deployed in the field that can receive some form of broadcast video and DVB-H should be on at least half of these.
‘Hollywood comes in two versions, one for ISDB-T, which is shipping in Japan, and one for DVB-H, which is also shipping now in volume, although I’d have to admit it is low volumes. Handset manufacturers are in the process of integrating these chips into devices, but none of those devices have yet launched so we can’t tell you yet which manufacturers,’ said Solomon.
ISDB-T 1 seg for mobile TV launched in Japan in April 2006, and although the visual experience is reported as substantially below that of DVB-H, all the operators there are backing the standard for the time being, and volume device shipments have begun in earnest. Much the same has happened in Korea with T-DMB, while over 50 trials in 35 countries are now scheduling the launch of DVB-H services, most of them over the next 30 months.
‘Texas will have a significant market share when those devices come to market,’ promises Solomon, ‘but right now we are more excited about GPS on the handset.’
Back in March TI announced its chip entry into that space, the NaviLink 5.0 GPS chip, built using the Digital RF Processor (DRP) technology which came out of joint research with Nokia.
Rival Broadcom clearly agrees with Solomon and this week launched its own contender for that chip space, the first since it acquired Global Locate in July this year, in what appeared a direct response to the TI move. It cited research which suggests that there will be 47 million stand alone GPS devices and 436 million handsets with GPS technology by 2010, so we can see what they’re getting excited about.
We had thought that TI might have lost interest in mobile TV since the US market was stolen in two deals by another of its arch-rivals Qualcomm, with its MediaFLO service. ‘MediaFLO is perhaps the third biggest hurdle to DVB-H in the US,’ says Solomon,’ the first is that the operators don’t believe that the public wants to watch TV on their handset.
‘The second reason is that the US operators don’t think the public will pay for it. The only major mobile TV services, in Korea and Japan don’t involve any increase in ARPU because they are free to air,’ Solomon pointed out.
Which explains why are both Verizon and AT&T are going with MediaFLO, because it involves little investment on their part. ‘Once they find a proven business model the big US cellular operators will want to own their mobile TV systems,’ observed Solomon. He also said he thought that AT&T at that point would go with DVB-H. ‘Verizon will pay no royalty fees for MediaFLO technology because Qualcomm’s license terms are that if you already pay for CDMA, you can have MediaFLO for free. AT&T will have to pay license fees, because it doesn’t use CDMA, as will its phone makers, unless they are already CDMA handset manufacturers like Samsung and LG.’
That is true if Qualcomm sticks to its public position on license fees, but it may not, and could instead prefer to ask for a percentage of the service revenue in order to keep AT&T away from adopting DVB-H.
This is the second time that we have heard that AT&T would, in the end, embrace DVB-H, with HiWire parent Aloha intimating similar logic to us a few weeks ago. This makes it look like AT&T is just covering the Verizon Vcast TV move and is not really behind MediaFLO as a revenue earner.
It wouldn’t be the first time that operators have done this. In Italy, where DVB-H has taken off with three competing services, 3 Italia has garnered 700,000 customers for its service while Telecom Italia and Vodafone Italy have failed to release customer numbers, and both are thought to have less than 10,000 customers, simply because they have decided not to subsidize handsets.
‘Most of the new customers that are taking Hollywood are putting them into high end and mid-range devices,’ said Solomon, and we do not know of any company that is buying it to put into a device which is not a handset.’
That doesn’t surprise us too much, in that Hollywood offloads the screen driving work to the OMAP architecture and it make sense that it would just be used in a handset that uses TI for its applications processor. This is not true for DiBcom or Siano chips, which are definitely in demand for devices which are not handsets at all, but which can receive a DVB-H signal.
‘There are two things driving handsets these days. The cyclic replacement of all the existing 3 billion handsets out there and the opening up of cellular to people in the world that have yet to buy handsets. The replacement market is where most of these phones are going, as operators try to differentiate their feature set with handset features. But even at the low end, where TI aims its LoCosto handsets, these devices need at least unicast video features. These are not the target for Hollywood, but these cheaper markets are targets today for video in devices that are under $40.’
The implication is that as the price comes down on Hollywood (bill of materials is under $10 now), it may well end up in handsets down in that price range, once those markets get broadcast video services.
The interesting thing about putting out a DVB-H service in a third world country which has low quality, non-digital TV, is that it has the opportunity to become the de facto TV service for that country, watched as much on PCs as it is watched on handsets. This is certainly the case in South Africa, which is expecting that townships will watch mobile TV using a PC as if it was a TV.
The big target for TI is to win the Nokia DVB-H business, since it is the prime contractor and handset supplier in about half of the global DVB-H trials and in some of the launched services. In the past Nokia would work the kinks out of a technology and then hand it over to a chip partner once the feature was ready to go into commodity.
‘It is true that in the past Nokia has gone to a chip partner once a chip is ready for production ‘ we would not say when it is going to become a commodity, but when it is ready to go into a catalog, there’s a slight difference there.
‘While we wouldn’t say that we are delighted that Nokia has just decided to share its chip partnerships more widely, and that includes giving business to our rivals, the positives that we can take out of that strategy change is that Nokia is moving from custom to catalog chips. That means it will buy more off the shelf devices, which can also be sold to others such as Motorola or Samsung. That plays to our strength because with our OMAP applications architecture we can handle multiple protocols and codecs with the same device.’
All of which may mean that as Nokia goes to volume on DVB-H devices it may indeed turn to TI after all, or it is at least as likely to do that as turn to one of the other possibilities, DiBcom, Siano and the outsider Broadcom which to date has no design wins for the DVB-H chipsets it inherited when it bought Athena Semiconductors in 2005.