A look inside DVB-T2 - CSI Magazine · Although DVB-T2 does have the tools to enable ‘mobile’...
Transcript of A look inside DVB-T2 - CSI Magazine · Although DVB-T2 does have the tools to enable ‘mobile’...
Released by the DVB Project in March
and to be made publicly available in the
middle of the year, the simple goal of DVB-
T2 is to improve on a ten-year old standard
and put terrestrial broadcasters in the 21st
century. It is the second generation digital
terrestrial transmission (DTT) standard,
succeeding DVB-T, which was developed in
1995 and is used by most of the DTT
implementations around the world.
The common consensus is that mass-
market DVB-T2 products will be available
by mid-to-late 2010, although prototypes
are expected to emerge in 2009. Chip
vendors are said to be working on it
already, and simulating the options in their
programmable arrays.
The headline figure is that DVB-T2
promises at least a 30% improvement in
capacity. According to Peter MacAvock,
executive director of the DVB Project Office,
simulations have shown that up to 60%
capacity improvement is attainable, although
exact performance gains depend on the
parameters that are chosen. Significant
innovations of the new specification
compared to DVB-T include new forward
error protection, high-order modulation
modes, variable modulation and coding of
services and increase in the number of
OFDM carriers.
Principally, the increase in data
throughput is achieved by employing
enhanced forward error correction (FEC)
schemes, such as LDPC schemes, as
opposed to just BCH convolution used in
DVB-T. It is essentially the same LDPC that
DVB-S2 introduced to improve DVB-S. This
suggests around a 2dB improvement, which
can be converted to capacity either through
using less error correction or moving to a
higher order constellation, such as
256QAM (DVB-T allows up to a 64-QAM
constellation per carrier).
As Darren Fawcett, chief technologist at
Pace Micro, points out, FEC schemes
ultimately improve the robustness of the
signal, but it's down to the broadcaster how
they choose to use LDPC. "They can use it
to improve the resilience but not choose to
exercise higher order modulation. Given the
drive is to increase the efficiency and
capacity of the network, it's likely they'll
move to higher order modulation."
Significantly, there is a 32k carrier mode,
in addition to 2k and 8k for DVB-T.
According to Simon Gauntlett, technology
director of the Digital Television Group
(DTG), the benefits this option brings is the
ability to use single frequency networks
(SFN) and increased robustness to impulse
interference. "This makes the 32k mode
extremely important as part of the DVB-T2
spec. The UK could in theory be covered by
a network of 100 or so transmitters using a
single frequency and so it's potentially much
more efficient that the current network."
Despite the benefits on offer, Mike
Brooks, head of R&D at Arqiva, sounds a
note of caution. "Nothing comes for free;
you can't get something for nothing and if
we're changing parameters we need to
understand what those impacts are. There
are consequences for doing that which
mean that, in the same way as DVB-T, you
have to pick and mix from the suite of
options that will be available within the T2
spec. The big thing about the DVB
standards is that they allow flexibility - they
are a toolkit of options that enable you to
trade robustness, and hence the number of
transmitters, against the data throughput.
And that philosophy is continuing with T2."
A competitive platform?
DTT HD services are the main driver behind
the specification, and it is likely that the
transition to DVB-T2 will be linked to roll-
out of these services. This makes
commercial sense. As Mathias Leutiger,
head of product marketing of broadcasting
T&M products at Rohde & Schwarz, points
out, one of the biggest challenges for a
successful introduction of new technologies
is adding value for and being accepted by
TV viewers. "Only when they perceive a
page sixteen www.csimagazine.com Cable & Satellite International march-april 2008
Terrestrial transmission
A look inside DVB-T2The new standard will empower DTT broadcasters and help them fight back against competing platforms,but many compromises are involved, says Goran Nastic
clear advantage will viewers be willing to
accept the added expenditure."
So while DVB-T2 doesn't mandate
MPEG-4, there are good reasons for
combining the two to ease transition.
“What’s being considered for the delivery of
HDTV is an MPEG-4 box so you may as
well add DVB-T2 to make that jump as far
as possible and avoid lots of mini
switchovers. That’s why the tie between
DVB-T2 and MPEG-4 is made. It makes it
more attractive to the consumer, as you
can’t really sell technology for technology’s
sake,” says DTG’s Gauntlett.
For this reason, it is likely that all T2
boxes will be H.264 based. It’s a matter of
bundling new technologies together at a
convenient point, with a new consumer
device that has enough of a value
proposition for them to change
The question is whether DVB-T2, even
hand in hand with H.264 encoding, will be
enough to make DTT a viable competitor to
cable and satellite, particularly with just one
multiplex as proposed by Ofcom in the UK?
According to Tim Sheppard, senior
business development manager of
broadcast at Cisco, how many HD channels
are possible varies hugely by country, but a
realistic target would be at least four HD
channels in a terrestrial multiplex. Of
course, MPEG-4 encoders continue to get
better as well, he notes. “Also if you put
four into a multiplex you can take advantage
of statistical multiplexing between the
encoders, meaning that one can go as high
as five or six HD channels in a mux leading
to a position where you have as many
H.264 HD channels as you used to have in
MPEG-2 SD.”
A similar argument is made by Simon
Frost, business development director at
Tandberg. “The 30% increase in bandwidth
gives you around 30Mbps to play within an
8MHz RF channel. Taking 5-6Mbps as the
average bit rate for HD in H.264 in a
statmuxed environment by 2010, you are
looking at maybe five or six HD services in
a multiplex. For some countries, that might
be enough.
“You can also pack many more SD
channels through with T2 and H.264. If
your proposition is payTV, such as Malta,
Netherlands and Sweden, then it will
definitely be a technology that you will want
to use. For example, we see a very clear
split in the countries that want to do payTV
in terrestrial in that most are using MPEG-4
even though the STB costs are currently a
little higher. In the public service broadcast
model, which is more prevalent in DTT, the
majority are MPEG-2 and less aggressive in
the amount of channels they are trying to
pack into a multiplex. Ultimately, it is the
business model that drives the technology
choices, but MPEG-2 and MPEG-4 AVC
remain critical in terrestrial,” he adds.
In the UK, a splinter group from the DTG
proposes to use the spectrum efficiency of
SFNs, claiming Ofcom’s proposals could
seriously compromise DTT’s long term
viability They see eight HD channels as the
minimum level to keep DTT competitive,
and argue that up to 40 HD channels are
possible on DTT using this approach. In
addition, it would also free up spectrum for
other uses, such as portable TV, as can
already be found in SFN areas like Germany.
It is worth pointing out here that DVB-T2
is not exclusively about HD. The debate in
the UK is mainly HD-driven, and
significantly, that is where most of the
urgency seems to be coming from the UK.
Elsewhere, interest is in ultra mobile
(portable TV in trains, cars etc), as well as
SD capacity, according to DVB’s MacAvock.
page eighteen www.csimagazine.com
Terrestrial transmission
To MIMO or not to MIMO?Whenever a MIMO (multiple input
multiple output) approach is used, it is
reasonable to expect an improvement in
robustness and a reduction in
destructive interference in SFN.
Arqiva has, in fact, been working with
the BBC and National Grid Wireless
(NGW) on a trial in Guildford looking at
the benefits of MIMO and says there will
undoubtedly be capacity increases; in
some cases it effectively doubles the
capacity of a single channel.
The technology, however, falls foul
of one of the commercial requirements
of T2, which stipulates that the T2
broadcast will be compatible with
existing DVB-T (and by implication
analogue TV) domestic installations,
ie single antenna, single downlead.
The general consensus, driven
by the commercial module, is that
existing antenna system in peoples’
homes shouldn’t be disrupted,
and these legacy issues preclude
a MIMO type implementation.
There is an option to deal with
MISO (multiple input single output), and
there are techniques such as Alamouti
schemes which could allow that to
happen so those are incorporated,
according to Arqiva.
There is work to enable future
development of a MIMO technology to
be allowed, but MIMO will not be
incorporated from Day 1.
IPR issuesGiven the problems the DVB had with
MHP, any potential property rights
issues have to be considered upfront.
Indeed, MHP kicked off a whole
reinvestigation of DVB’s IPR policy.
As a part of the DVB-T2 work, a new
approach to the existing rules was
suggested in that an early formation of
the patent pool was being encouraged
so that any IPR issues would come to
light well in a advance of anybody
deploying a network.
The aim is to avoid licensing issues
down the line.
“Simulations have shown that up to60% capacity improvement isattainable, although exactperformance gains depend on theparameters that are chosen.”
Cable & Satellite International march-april 2008
Cable & Satellite International march-april 2008 www.csimagazine.com page nineteen
As some sort of guide to the level of
interest, MacAvock points to the fact that
while some 30 companies formed a subset
of actively involved members in DVB-S2,
this number is closer to 60 with T2, with
Russia, China, the US, and Korea to name
but a few all involved.
Although DVB-T2 does have the tools to
enable ‘mobile’ TV better than DVB-T, T2’s
commercial requirements are aimed
primarily at rooftop aerials, and the work
that supports handheld mobile phone type
TV is outside the scope of the T2 work,
according to Arqiva’s Brooks. “There is an
initiative within the DVB called Next-
Generation Handheld (NGH), which will be
looking at this aspect,” he says.
Tandberg’s Frost notes that T2 had the
option of a mobile focus, but there was
general lobby against confusing it with an
emerging DVB-H market. “Having said that,
a DVB-H and T2 mix could be interesting.”
The road to IP
DVB-T2 further cements the arrival of IP in
the broadcast industry by including GSE
(Generic Stream Encapsulation), an
evolution from the traditional method of
putting IP over the MPEG Transport Stream
(TS) over physical layer. The use of GSE -
while not mandatory - contributes to the
30%+ gain in capacity, especially when
applied to IP-based services.
“IP over MPEG TS has never been that
efficient; it’s a little clunky in terms of
encapsulation overheads. One of the
evaluation areas was to look at a native IP
transport layer that could exist inside T2, so
that you may be able to handle IP data
more efficiently and easily within the overall
transport. It may apply to even transport
layers of interactive applications perhaps
that have previously been carouselled.
There is, however, no shift in concept that
the MPEG TS is dead and everything will be
streamed in IP,” says Frost.
This brings into question whether
terrestrial broadcasting will eventually adopt
an all-IP approach that includes
transmission. Arqiva’s Brooks acknowledges
that IP does provide a common layer across
all sorts of platforms and makes distribution
and contribution circuits easier to
implement. “So I can foresee a situation
where we may see an IP infrastructure
being deployed in the future. That said, I
have not heard of any specific plans to
make everything IP-based.”
Even Sheppard’s Cisco agrees that an all-
IP approach isn’t necessary, despite some
inherent benefits. “I’m not sure it’s
necessary to remove the TS from the core
video. But you can imagine services like
push VoD and interactivity besides the core
video and audio, which would logically
make sense to go over IP because that’s
the format that’s now contributed in,
distributed in and browsed in.”
Cisco commissioned an independent
study that forecast that 65%+ of spend in
three years time on video networks will be
IP-based, and Sheppard is confident that
some of that convergence story will leak
over into the end transmission side.
DVB-T DVB-T2Error correction BCH LDPC
Modulation QPSK, 16-QAM, 64 QAM Addition of 256-QAM
FFT sizes 2k and 8k Addition of 32k mode
Physical/link MPEG- 2 TS MPEG-2 TS or IP Baseband frames
layer interface
Bandwidth 5, 6, 7, 8Mhz 5, 6, 7, 8Mhz
Multiplexing Single TS per Multiplex Potential for bonded UHF channels
Transport ~24Mbps ~32.4Mbps (44Mbps+ with MIMO)
rate/RF
Codecs MPEG-2, MPEG-4 AVC, VC-1 MPEG-2, MPEG-4 AVC, VC-1
Source: Various; CSI
Terrestrial transmission
CSI
Time frequency slicingTFS is a technique that allows services to
be spread over more than one multiplex,
maximising the benefits of statistically
multiplexing more services, a harder issue
when moving from SD to HD. As with
MIMO, the concept of TFS is extremely
compelling, but it looks like it has not made
its ways into the original specification given
its broader complexity from a chipset point
of view. The compromise decision in the
DVB-T2 committee is to have the TFS as an
optional part of the spec.
Tandberg’s Frost believes because the
availability of silicon in the tuner and the
commoditisation of the pricing are critical
factors, TFS is unlikely to feature in
commercial products, at least for the time
being. “If it hasn’t made it in as a
mandatory profile there isn’t going to be a
lot of interest in adding that.”
For Luc Ottavj, chief scientist officer and
co-founder of UDcast, typical questions to
be checked are what the exact overhead is
for TFS compared to the statmux gain, and
- due to the network plannings - how many
UHF channels having exactly the same
footprint can be TFS-grouped in a network.
These are mere semantics, however. In
reality, it requires having more than one
DVB-T2 multiplex, which doesn’t fit with the
current Ofcom proposal.
CSI understands that while a number of
companies were known to be pushing for
this technology to solve their transmitter
problem, TFS has been dropped due to the
ambitious timescale of the Ofcom proposal
and that the BBC, the standard’s champion,
feared that including this method would
delay deployment.
These developments have led to
accusations that DVB-T2 has been modified
to suit UK interests. Because parts of the
Uk are already concerned and nervous that
DVB-T2 might impact the DSO, it arguably
would have been better to implement the
technology following DSO’s completion.
With this in mind, despite its many
advancements, it is in some ways hard to
escape the conclusion that DVB-T2 could
have been much more than the sum of its
parts, at least in its first incarnation.
CSI