Modoff- MediaFLO Round 2
Mobile-TV on the horizon
The fight for mobile TV standards has begun. On the far right is Qualcomm with its
proprietary MediaFLO solution, the far left, the rest of the world and DVB-H. We
dug into both the standards and present some of our findings here. We also
present some of the issues that DVB-H could face in the US, given that Crown
Castle is planning to launch DVB-H in the 1.67 GHz spectrum. Physics 101: The
higher the spectrum, the smaller the cell radius, the more number of towers
required to cover an area.
Round two:
Just when one thought that the battle between Qualcomm and the rest of the world was
over, a new and potentially prolonged front is being born. The new battle field is the lucrative
market that could have a disruptive impact on wireless communications and the way
consumers receive information.
The mobile industry realized early on that streaming live TV on a wireless network (over the
precious limited spectrum) had limited potential for growth. Although this ability to stream
media from the network to mobiles is a powerful idea, streaming live TV (which is generally
bandwidth intensive) presents many challenges and often performs poorly. Thus, the other
option is to stream/broadcast live TV independent of the wireless network. In this scenario,
mobiles would receive TV via a passive receiver (designed into the handset) on an
independent frequency band (that is separate from the wireless carrier frequency). In the US,
Qualcomm and Crown Castle each have spectrum in the 700 MHz and 1.67MHz (L-Band)
bands respectively, and plan to rollout mobile-TV in 2006.
Most technologies that exist in the market are variants/ derivatives of an existing television
broadcast format and have been fine tuned to address the mobile multimedia requirements.
A majority of these technologies use OFDM (except S-DMB, which uses code division
modulation). Also convolutional coding and Viterbi decoding17 are used in all existing
standards except for FLO which utilizes turbo codes instead.
FLO vs DVB-H
Given that MediaFLO is not an open standard like DVB-H (although Qualcomm has launched
efforts to standardize FLO by creating a consortium of handset vendors and operators),
studies that compare the performance of these two competing technologies are sketchy.
Qualcomm has published a variety of comparisons, although any granularity around the
methodology is currently not available. We have decided to include these results to make our
readers aware of the performance numbers currently available, and hope in the future,
Qualcomm offers further details regarding these claims.
Some of the main points we want to highlight from Qualcomm’s whitepaper are the
following:
1) DVB-H performance relative to FLO in 1bps/Hz is 3db – 4db worst.
2) FLO claims much better switching times than DVB-H. FLO average channel
switching time is approximately 1.5 seconds vs DVB-H with an average of 5
seconds. As we mentioned earlier in our write up, we have noticed this issue
(where DVB-H switching times were significantly longer than FLO, and that FLO’s
switching time was clearly in the 1-1.5 second range)
3) FLO claims that it could support as many as 20 channels (3 channels per mega
hertz) in a 6MHz band vs DVB-H’s 9 channels (1.5 channel per mega hertz). If this
claim is true, then FLO can support double the number of channels in the same
capacity compared to DVB-H, a very important distinction that could give FLO an
edge over competition.
We note that even if these figures hold up in the field, there are no guarantees that
Qualcomm’s purported superior technology will be more widely deployed than DVB-H. Given
that there are many more companies participating in the standardization and development of
DVB-H, it may see wider deployment rates starting with GSM/WCDMA operators first. On
the other hand, if FLO visibly outperforms its competition by higher uptakes among
consumers, then wireless operators may be forced to embrace the better performing
technology.
Crown Castle’s unique challenges in the US
The first likely clash of FLO and DVB-H is the US. Crown Castle International Corp., a
company that manages a network of more than 10,000 wireless towers purchased 1,670- to
1,675-MHz spectrum (5MHz) for the purpose of building a national mobile-TV network in the
United States. The company describes its offering (which it plans to sell wholesale to cellular
carriers) as cable of the wireless world. Crown Castle plans to offer 10 channels (including
news, sports and entertainment) at a QVGA-resolution (320 x 240) of 15 to 25 frames/second.
The service will also include about 20 audio channels at 32 kbits/s. Motorola, Nokia and
Samsung are expected to participate in its trials, and Microsoft will provide its media player
and digital rights management software.
Clearly, Crown Castle has the necessary tools to launch service: it has the spectrum, the
towers, a standard driven by the wireless industry and core group of technology partners
committed to delivering this network. Yet, there are many issues that work against the
company. Firstly, the company is limited to a relatively low, 2-kW maximum transmission
power on a high, 1.6-GHz frequency. This is in comparison with Qualcomm, which can blast
50-kW in the lower 700MHz spectrum. As we all know, higher frequencies translate to
shorter propagation distances, and lower power wattage means even smaller footprints. In
essence, Crown Castle would require many more towers than Qualcomm. More
interestingly, Crown Castle’s current trials use towers that radiate a mere 400 W. Clearly,
FLO would employ two or three high-power transmitters in urban areas, where Crown Castle
would need to use a significantly larger number(we have heard as high as 1000 to 2000
towers), implying lower costs of service and management for FLO.
Additionally, Crown Castle has a 5 MHz band while FLO has 6Mhz. If Qualcomm’s claims
hold up in the field, then FLO would be able to pack 20 channels into 6 MHz of spectrum,
compared 8 channels for Crown Castle in 5 MHz.
The downside for Qualcomm is the 700-MHz spectrum is laden by the large number of
analog UHF TV stations broadcasting at or around channel-55 crowded conditions that could
inhibit access to as many as 180 million users, according to Crown Castle. Qualcomm is
already working with many of these local broadcasters to gain access to the spectrum, but it
may have to wait until 2007 (as mandated U.S. deadline for the shift to digital TV), before it
can gain access to its spectrum and build a nationwide FLO network in the 716- 722 MHz
band.
Although it is too early to predict the outcome of this new battle, it is clear that Mobile-TV has
gained traction and is the focus point for many heavy hitters in the industry. The battle field in
the US is clearly tilting towards Qualcomm’s favor, given the simple laws of physics and
cost. Since it has access to a lower frequency spectrum, and can essentially blast power 50
times more than its competition, it would need far fewer towers to deploy and manage its
network. This in turn translates to lower cost of service (a very important point in the
commoditization of mobile-TV). Also, it has a good 1Mhz additional spectrum than Crown
Castle, which means it can offer more services. If Qualcomm’s FLO technology is superior to
DVB-H (which we believe to be true), then FLO can outperform DVB-H in terms of power,
switching times and capacity. The tougher question is whether this would translate to higher
consumer attach rates.
AI
