Memorandum submitted by TrinityStar Associates
Ltd
INTRODUCTION
Terrestrial television broadcasting, in a form
we recognise today, started almost 60 years ago in the UK. In
that time it has undergone one major transition, from black and
white 405 lines to colour 625 lines, and is now part way through
its second, from analogue to digital. Satellite broadcasting and
cable television have also appeared on the scene and become significant.
Any change to an existing broadcasting system will always be complex
and potentially disruptive because of the large installed base
of equipment. In any transition, it is always attractive to go
for the very latest technology, however, viewers are used to a
reliable rugged system and are not tolerant of their television
misbehaving in the way that computers do with new software. For
that reason the broadcast tends towards the tried and tested rather
than the latest.
The purpose of this submission is to look at
the engineering issues behind analogue switchoff and to explain
these in a way that makes the constraints and compromises that
go into planning a system clear.
BACKGROUND
Television broadcasting was restarted immediately
post war with a black and white system using 405 lines on VHF
Band I with a single programme from the BBC. Near universal coverage
of the UK was achieved with around 50 stations. In 1955, ITV was
born under the auspices of the Independent Television Authority.
ITV also started in 405 black and white but on the much higher
Band III channels, pushing transmission technology to the limit.
This heralded the first "set top boxes" which were to
enable Band I sets to watch Band III.
Colour transmissions started in the late 60s
using new spectrum on UHF bands IV and V. The spectrum use was
planned and co-ordinated at an international conference in Stockholm
in 1960 (a plan that we use today and is the basis for our digital
plan going forward). Two things are important about this:
— Due to the propagation characteristics
of UHF compared to VHF, over 1,000 stations would now be needed
to give universal coverage.
— There was spectrum availability to
allow full simulcast of the 405 and 625 line services, 405 line
finally being closed down in the early 80s.
At the time the UHF stations were built, provision
was made for four UHF channels. This made the building of Channel
Four in the early 80s more straightforward as the antennas, structures
and combiners were all in place and the building programme was
to add the transmitters. Channel 4 launched in November 1982 following
the world's fastest ever rollout of a national television network.
Within bands IV and V, 46 channels are available
for transmission. The network, totalling some 1,150 sites, was
rolled to provide a service down to communities of 200 people.
This added up to 4,600 transmitters, which meant that each available
channel was reused 100 times. Analogue transmissions are very
susceptible to interference from other transmitters from a considerable
distance so this represented quite a feat of planning.
Research and development work on digital started
in the 1980s in the Independent Broadcasting Authority R&D
department and was continued post privatisation in 1990. It is
interesting to note that the possibility of mutichannel digital
terrestrial existed at the time Channel 5 launched and there was
press comment at the time about launching with "yesterdays
technology".
The digital terrestrial system was demonstrated
in the House of Commons in 1995 and paved the way, under the 1996
Broadcasting Act, for the planning and implementation of the present
80-station DTT system. At the time, the UK was the world leader
in this technology.
THE DAWN
OF DIGITAL
TERRESTRIAL
No more spectrum was made available for the
new DTT system, six multiplexes had to be shoehorned into the
existing channels. Furthermore the regulatory constraint was that
there must not be any degradation of the analogue service. The
best that could be achieved was a network using 80 stations using
low power (81 were planned but co-ordination problems meant that
the Channel Islands were never built). Furthermore, restrictions
were placed on some of the transmitters in some directions to
avoid interference, this resulted in some homes receiving a subset
of the possible six multiplexes.
The transmission technology is OFDM (Orthogonal
Frequency Division Multiplex), which is, in effect, splitting
the digital signal up into "comb" of carriers within
the channel. Each element of the comb only carries low speed data
and only a small part of the data resulting in a rugged signal.
At the time receiver chipsets were available only for the 2K carrier
variant (8K has now been adopted in other countries). One of the
disadvantages of 2K is greater susceptibility to impulse noise.
The system went live on 15 November 1998. The
transmission network was delivered on time and within budget.
Much was learnt in the first years; the interference to analogue
was not as sever as feared do the DTT transmitter powers were
increased improving coverage. As compression technology improved
more channels were squeezed into the platform.
The use of the platform and the marketing aspects
are outside the scope of this submission but there is no doubt
that the simple Freeview concept was the key success factor that
has lead to DTT being a close second to Sky.
LOOKING TO
THE FUTURE,
PROJECT GENESIS
Once the 80 station DTT system was on air a
small group started thinking about the "all digital future"
and how to get there. This lead to a project commissioned by the
ITC called Project Genesis. In the first phase we developed a
number of scenarios, looking to beyond analogue switchoff to form
a view of the end game. A broad range of views was taken from
across the industry and then, in the second phase, some of the
scenarios were developed in great detail. It was something of
a first bringing together a regulatory, strategic, engineering
and planning view to build a picture of the future.
The co-authors of this report,[102]
published in 1999, were Dave Brian, Brian Tait and myself with
input from Smith Consulting. The Genesis report forms the blueprint
for the future of digital television in the UK and many of its
concepts have now found their way into EU thinking. It is a tragedy
that Dave was later killed in a road accident, perhaps this submission
will stand as a tribute to his fine work.
The next stage was to look at how to get from
where we are now to the all-digital world. In an ideal world we
would build out digital to match analogue coverage and then withdraw
analogue when digital take-up had reached an acceptable level.
The snag was that the existing broadcast spectrum was tightly
packed and no extra was available, on the contrary the changeover
entailed releasing 14 of the 46 channels.
The conundrum was that we wanted to build digital
in order to be able to close analogue but could not build digital
without closing analogue. From early 2003 a colleague, Brian Haseler,
and I turned our minds to this. The only solutions being offered
in the industry at the time were hugely expensive and technically
unworkable, involving multiple changes of transmitter frequency
at each site. Between us we developed the concept of closing down
one analogue channel at a transmitter site and immediately replacing
it with a high power digital multiplex that contained the analogue
services being withdrawn. From a viewer perspective, they would
lose an analogue channel but would get the service back by getting
a digital box (if they had not already done so). After a period
of time the remaining analogue services would be withdrawn. From
a resource perspective this would break the project down into
manageable chunks and phase the activity over time. The plan rapidly
became known as "the ntl plan". It had its first public
airing at the Digital Action Plan (DAP)Stakeholders Group on 18
February 2004 when Brian and I presented it to the industry. There
was widespread support for the principle but no one was going
to volunteer to be the first off at that stage.
The plan was adopted by the DAP and work started
on the way in which the changeover should be tackled across the
country. A sensible basis was the existing television regions.
It rapidly became clear that the order would be determined by
engineering constraints as there was only a certain order to do
the changeover that avoided one region wiping out existing coverage
in an adjacent, yet to be converted, region. The country divided
in to two blocks, within those blocks the order was fixed but
the order of the two blocks could be swapped. Early work on this
by Brian and myself was taken into and developed by the DAP Regional
Rollout Group. The North/West block was chosen to go first as
it had the lowest risk of being changed in the process of co-ordinating
channels with France/Belgium/Holland.
All this work has now transferred to DigitalUK
and forms the basis of very detailed implementation plans.
ENGINEERING ISSUES
AND CONSTRAINTS
System choices
The overarching specification for DTT allows
numbers of choices so that the system can be engineered to match
the needs of the user. One choice has been mentioned above, that
between 2K and 8K carriers. The UK has decided to transition to
8K which will improve performance and avoid being trapped in a
2K system forever. All but the very early DTT receivers cope seamlessly
with both standards.
Other choices involve balancing capacity against
ruggedness against transmitter power. There is a choice between
modulation schemes referred to as 16QAM and 64QAM. 16QAM represents
a more rugged signal against 64QAM having 20% more capacity. The
difference in ruggedness can be offset by an increase in transmitter
power. Currently both are in use.
Ofcom[103]
consulted on this and other options in February 2005 and concluded
that 64QAK would be used. This was in conjunction with the DTT
power being at -7dB relative to analogue (one fifth the power)
at most sites and at -4dB at some. This would result in a served
population of 98.5%.
Coverage and frequency planning
The actual coverage of a transmitter is determined
by three factors:
— Distance from the transmitter.
— Interference from other transmitters.
— Terrain blocking the signal.
From the above, it is clear why there are few
transmitters in East Anglia and lots in South Wales. Although
distance is ultimately a limiting factor, in practise the other
two are far more significant.
As things stand today, there is a stable situation
with a known population of transmitters in the UK and surrounding
countries. Similarly the interference from the UK out and other
countries inwards is understood and allowed for. As mentioned
above, this is all based on the 1960 Stockholm plan. One possibility
would be to start with a clean sheet of paper and totally re-plan
for digital, however this would be a mammoth task. What has sensibly
been accepted is the concept of "conversions". This
was agreed at an international planners conference at Chester
and allows an exiting Stockholm 60 analogue allocation to transition
to digital. This eases the frequency planning and co-ordination
task considerably but it is still complex. The mechanism for co-ordination
is explained more fully in the next section.
The UK frequency plan should deliver 98.5% coverage,
however, there is a little more to it than this. Expressed in
full the plan will deliver coverage to an acceptable level of
all three public service multiplexes to 98.5% population for greater
than 99% of the time based on a standard receiving aerial at 10m
height. The 99% time reflects the fact that incoming interference
changes with weather conditions and very occasionally the weather
may give rise to increased interference, (this is the same for
analogue). The levels of interference are "worst case"
based on known plans of our EU neighbours and the commercial broadcasters
rolling out to their full allocation of 200 sites. It may well
be that the interference is less, in which case this number served
will rise.
Similarly the remaining 1.5% are not simply
"have nots". Some of the 1.5% will get all three multiplexes
but at a quality that the planners regard as below standard. In
most cases this is recoverable by a better aerial installation.
There will be a subset that will get one or two multiplexes and
a few that get none.
It is not possible to say where those few will
be other than they will not be concentrated on a particular geographical
area and will be scattered house or very small communities.
As a very general statement, anyone getting
a passable analogue signal now should be able to get digital.
Using the same planning tool, analogue delivers 98.5%. The note
of caution is that the two 98.5%s are not necessarily the same.
The differences (which will be very small) arise from inevitable
small changes in transmitter antenna pattern when they are re-engineered
and different levels of incoming interference from the Continent.
Regional Radio Conference
The Regional Radio Conference 2006 will put
in place a plan to replace the Stockholm 1960 plan. Each country
has now submitted its plans and discussions between neighbouring
countries have been taking place to mesh their individual plans.
For the UK, Ofcom leads this activity with support from the BBC
and the two transmission operators, National Grid Wireless and
Arqiva.
The UK, in line with most, of its neighbours,
has put in a plan for eight multiplexes (described in the planning
as "layers"). Six of these represent the six licences
UK multiplexes and the remaining two have been planned in the
14 channels of released spectrum. One of the two new ones is of
particular interest to the broadcasters as it is mainly a conversion
channel rather than a new allocation and so is eminently suitable
for broadcast applications.
All being well the RRC should be a rubber-stamping
exercise but there remains some risk of change to the plan along
the South and East coast up to the conference.
Extent of rollout
In 1993 there was strong feeling in the industry
and government that DTT would only roll out to 200 stations and
that satellite would be the mechanism to deliver service to 8-10%
of the population (5-6 million people). Detailed analysis coupled
with developments in small transmitters reversed this sentiment
and the Digital Replacement Licences issued by Ofcom in December
1994 contained an obligation to build out to the full 1,154 sites
(except the commercial operators who are limited to 200 by spectrum
availability).
From the transmission point of view, the project
to build the DTT network will cost circa £500 million of
which 20% is for stations 201 to 1,154. These sites vary from
quite significant ones covering a large town down to a telephone
box sized building covering perhaps 200 people. In total these
sites cover 4 million people, which works out to an average of
£25/head for conversion (this cost is borne by the transmission
company/broadcaster). The tiny sites represent the worst case,
if we take the smallest covering 200 people, the conversion cost
is around £15,000 which represents £75/head.
There is a cost should conversion not take place
as many of the smaller sites are only used for broadcast and,
if this ceases; they will have to be returned to a Greenfield
site. The cost to do this is comparable with the conversion cost.
The changeover plan
The principles behind the changeover plan have
already been outlined. This section provides more detail and explains
more of the thinking behind the plan.
The task of changeover is dominated by the work
that needs to be done at the main stations; the work at the small
relay stations is largely a box swap. It is for this reason that
the four-year changeover period has remained even though the number
of stations considered has gone from 200 to the full 1,154.
At the main stations, the existing analogue
transmitters and low power DTT transmitters have to be removed
and replaced by high power DTT. In two cases the structure needs
replacing and in many others strengthening. The antenna at the
top of the mast needs upgrading or total replacement. In many
cases, electrical switchboards and generators need replacing.
Most of the analogue equipment is near end of life and due for
replacement.
Of these tasks the most critical is the structure/antenna
work as this can only be done in the summertime.
A typical project plan for, say, a sports stadium,
would have a number of milestones along the way and then a contingency
period at the end before it was required for use. This project
is unusual in that there are firm milestones representing the
changeover date for each region (this would be publicised well
in advance and are therefore firm). Contingency at the end of
the project is therefore no use. The solution in this case is
to build contingency at the beginning by starting the antenna
work two years in advance of the time it is needed. The project
will stand one or two bad summers before serious slippage occurs.
The project is divided down to a main transmitter
and its relays. Two years before changeover the antenna work will
be done. Around six months before changeover the new transmitters
will be installed and thoroughly tested. On the changeover night,
the BBC2 analogue transmitter will be closed down and a high power
multiplex brought up on that channel. Over the next 24 hours all
the relays will be visited and converted. One month later the
remaining analogue transmitters will close and the other DTT transmitters
will go to their new channel and high power and the relay stations
completed.
From a transmission point of view the month
is about right, shorter than two weeks would leave insufficient
preparation time. Longer than one month would mean diverting teams
to another site and then bringing them back.
This sounds like a great deal to be done from
the viewer perspective in a short space of time but it is important
to note that the existing sites reach 75% so it is only those
on the edge of main station coverage and those in the relay areas
who will not have DTT until the first switch.
The Future
There is no doubt that DTT has become a successful
platform, the sales of receivers and the price recently paid for
the extra capacity on the platform are proof of that. The platform
is now full and there are unlikely to be improvements in the MPEG2
compression technology that will allow much more to be squeezed
in.
The domestic trend towards larger screen sizes
will raise awareness and desire for high definition. HD is bandwidth
hungry and the present system could not cope with it. There are
two (not mutually exclusive choices):
— Move to a more efficient coding system.
A simple strategy would be to allow one or two
more multiplexes for broadcast. Using MPEG2, two HD channels could
be put onto a multiplex. A better strategy would be to code the
HD signals in MPEG4. Domestic equipment that could cope would
receive the HD signals; the remainder would not.
A better strategy still would be to consider
how to migrate the whole platform to MPEG4. At present all consumer
equipment for DTT is MPEG2 and suggesting that this should be
thrown away is inconceivable. It is probable that boxes will soon
be dual-standard so over time the problem will disappear with
the natural replacement cycle. The implication of this is to simulcast
in MPEG2 and MPEG4 for a period of years. Again this will need
more capacity than the current six multiplexes. With seven or
eight it would be possible to transmit some HD as well as simulcast
some existing channels. Fortunately this has no implications for
the transmitter network, which will handle both.
Ofcom have just launched a consultation into
the use of the released spectrum, hopefully it will conclude that
some should be used for broadcasting.
CONCLUSION
This submission has set out to give some background
to where we are now, explain some of the issues and give some
insight into the engineering and technical aspects of the transmission
system. Opinions expressed are those of the author and do not
represent any organisation other than TrintyStar.
17 November 2006
102 Genesis Project, Independent Television Commission,
February 2000. Back
103
Ofcom, Planning Options for Digital Switchover, February 2005. Back
| 
