the dubious distinction of being the most controversial item of new
technology to be introduced in recent times. Enmeshed in long running
arguments over basic technology, price and market access, HDTV is having
a painful birth worldwide, and the Australian experience differs little
from that overseas. While the focus of the domestic arguments in many
respects differs from the focus of the international arguments, the
common factor is that the standard evokes loud and partisan argument.
Wherein lies the truth ? In this month's feature we will
take a closer look at the basic technology in HDTV, what its likely
consequences will be in the wider market and attempt to shed some light
on the broader issues.
Analogue TV in its modern form is late 1940s technology,
and in most respects a well entrenched artifact of the industrial age.
When television was introduced, it incorporated the very state of the
art in wireless broadcast and modulation technology.
The video or picture information was presented using an
interlaced raster scan. Synchronisation pulses were attached at the
beginning of every line in the picture, and special lines were used to
synchronise the frames of the picture. The video signal used intensity
modulation, but inverted in polarity.
The modulation technique was also advanced for the
period, using Vestigial SideBand (VSB) Amplitude Modulation. This
technique almost doubled the efficiency of a transmitter, since one
sideband of the modulation envelope and much of the carrier were
The sound channel used the best technology available for
high quality analogue transmission - Frequency Modulation (FM). With a
high ability to reject interference, FM proved to be an excellent
Early television was not without its warts. Birth
defects and radiation injuries due to X-ray emissions through the glass
Cathode Ray Tube faces, and very nasty glass shrapnel injuries from tube
implosions both proved to be issues. The modern tube today uses a
toughened glass face, with X-ray absorbent dopants in the glass. By
today's standards an early television receiver was a dangerous
contraption. Reliability was also an issue, since the vacuum tube
technology of the day was only capable of several thousand hours of
operation before the tube cathodes became exhausted. This made early
television expensive to maintain and created a massive industry of tube
jockeys whose principle technical skill lay in guessing which tube to
swap to eliminate a particular symptom.
The US were the first to introduce colour television.
Their NTSC standard used Quadrature Amplitude Modulation (QAM) to encode
two colour difference channels on to a single sub-carrier. The
instantaneous amplitude and phase of the sub-carrier encoded the hue and
saturation of the picture. By demodulating the colour sub-carrier, and
combining the monochrome video signal with the two colour difference
channels, the Red, Green and Blue signals could be reconstituted to
drive the three gun colour tube.
NTSC quickly earned itself the nickname of Never The
Same Colour, since the QAM technique encoded colour hue in the phase
angle of the sub-carrier, and that phase angle could be distorted by the
ugly phase characteristics of the cheap and nasty booster amplifiers,
front end amplifiers and intermediate frequency stages of the day.
The Europeans followed the US with colour. Germany
developed the PAL (Phase Alternating Line) standards, derived from NTSC
but with the addition of a phase reference burst in every sync pulse,
this burst being alternated in phase line by line to compensate for
phase distortion in transmission. The final PAL-B standard was widely
adopted in most of Europe and is the scheme used in Australia. The
French, true to form, decided to go it alone and developed the SECAM
standard which used FM techniques to encode the colour. SECAM was
adopted mostly in French speaking nations, but also became the communist
bloc standard. Today it is mainly used in France and Russia.
Analogue TV has evolved very little since the
introduction of colour. We have seen teletext added, and stereo sound
transmission. Neither represent significant enhancements to the basic
Analogue TV has several serious limitations. The first
is its sensitivity to interference and self-interference via the
multipath propagation of signals, the latter a bigger issue in cluttered
urban environments. The second is its poor resolution, since it is
limited by a fixed modulation scheme which is locked to the picture
scan mechanism. With at best the equivalent to an 800 x 600 interlaced
monitor image, it is hardly an icon of picture quality and does not do
any justice to modern productions made for cinema presentation. With
almost no ability to accommodate further technological growth, analogue
TV is showing its one half century old origins. While the modern TV set
is a technological marvel by the standards of 1950, it falls very much
short of what can be achieved in the digital age. Herein, however, also
lie the roots of much of today's controversy about HDTV. The incessant
chorus of complaints about HDTV costs are firmly centred in comparisons
with an evolved and tired 1950 period technology. Reality check for
HDTV critics: compare the cost of a modern HDTV set to your annual
salary, and compare that to the cost of a television set against annual
salaries in 1955. You might be disappointed !
The DVB and ATSC HDTV standards which are now being
introduced in Europe, Australia and the US, build upon nineties
technologies such as MPEG lossy picture encoding and advanced digital
modulation schemes. They are not the first foray into HDTV, the Japanese
having made an earlier attempt to leapfrog the pack with a satellite
delivered analogue HDTV standard. With prohibitive demands for
bandwidth, analogue HDTV never quite made the grade.
Digital HDTV is clearly the path of the future, since it
provides superior quality, can fit into the tight 6 to 8 MHz TV
broadcast channel bandwidth of established analogue TV, and provides the
power and flexibility of a digital transmission channel and encoding
scheme. To better appreciate the longer term implications of HDTV it is
very useful to explore the basic technology.
HDTV Standards and
The two principal families of HDTV standards currently
penetrating the marketplace are the US ATSC (Grand Alliance) and the
European DVB standards. The Americans were the first to make a serious
commitment to developing HDTV during the early nineties trade war with
Japan. US TV manufacturers had been almost wiped out by the previous two
decades of competition with the Japanese, who has managed to displace
them from most of the US domestic and export markets. HDTV was seen as
an opportunity to make a new start and also revive the US industrial
base for manufacturing commodity video RAM, DRAM and consumer market RF
components. The Grand Alliance, comprising several US manufacturers and
industry groups, decided to exploit the capabilities of the new MPEG
lossy compression video encoding scheme, which removed the bandwidth
problems which plagued HDTV schemes using analogue encoding or
uncompressed digital video.
Considerable effort and money was expended in the
development of the standard and the design and evaluation of trials
hardware and systems, to identify what were likely to be problem areas
in the technology.
The result of the Grand Alliance effort was the ATSC
(Advanced Television Systems Committee) standards package scheme
(www.sarnoff.org). It combined the use of MPEG-2 digitised video with
the 8-VSB (8 level trellis coded Vestigial SideBand) modulation. The 8
level (3-bit) trellis code is a convolutional forward error control
scheme which was specifically chosen for good resistance to white noise
and thus good system performance with weak HDTV signal. The ATSC scheme
takes the MPEG encoded video and sound, and randomises it with a
scrambler, after which it is encoded into 187 byte packets using
Reed-Solomon (R-S) error control coding with 20 parity bytes. The latter
is used to overcome the limitations of the trellis code in handling
burst noise. The R-S encoded data stream is then interleaved to further
reduce sensitivity to burst noise, and synchronisation sequences are
added. This stream is then fed into the trellis encoder and then the VSB
modulator to produce the signal fro transmission. For cable TV use, a
16-VSB scheme is available which trades noise immunity for better data
rates. A data rate of the order of 20 Mbps is needed for rapidly
changing scenes such as sport or action cinema.
The ATSC standard has at this time been adopted by the
US, and with minor modifications, by Canada, South Korea and the
Phillipines. Japan is fielding its own ISDB standard. Europe and
Australia have opted for the DVB-T (Digital Video Broadcasting/Digital
Versatile Broadcasting - Terrestrial) standard.
The DVB-T scheme is defined by the ETSI EN 300 744
standard (www.etsi.org). It employs a very different approach to
modulation, using a COFDM (Coherent Orthogonal Frequency Division
Multiplexing) scheme, combined also with MPEG-2 encoded video and sound
(ETSI ETR 154). COFDM is derived from researched published in the mid
nineties, and post-dates the early US HDTV research.
Like the ATSC system, the DVB-T system starts with an
MPEG encoded data stream. Packets of 187 bytes are fed into a randomiser
and then a Reed-Solomon encoder, using 16 parity bytes. The R-S coded
data is interleaved. At this level both the DVB-T and ATSC systems
differ little in concept, although many of the parameters are slightly
different. The systems diverge radically at this point. The DVB-T
system uses Quadrature Phase Shift Keying (QPSK), 16 or 64 Quadrature
Amplitude Modulation (16-QAM or 64-QAM) applied to 1512 or 6048
individual carriers. Rather than modulating a single carrier at a very
high data rate, COFDM modulates a very large number of carriers, spaced
at either 1.116 kHz or 4.464 kHz, each with a very slow symbol rate.
The most trivial comparison is that COFDM transmits data in parallel,
against 8-VSB which transmits serially. To aid in receiver
synchronisation, COFDM continually transmits 17 or 68 pilot carriers.
Figure 1 shows the characteristically flat COFDM spectrum (ETSI EN 300
The COFDM scheme has the nice property that established
Fast Fourier Transform algorithms can be used in the demodulator, and
inverse FFT in the modulator. Indeed the DVB-T standard specifies these
as the preferred technique.
There is considerable argument in broadcast engineering
circles as to which of the two standards, 8-VSB ATSC or DVB-T COFDM, is
the better. While the US and European authorities have pretty much cast
themselves into concrete on the issue, loud public arguments are raging
over the export markets for their respective domestic manufacturers.
There are two bones of contention here.
Trial tests and theoretical predictions indicate that
COFDM performs significantly better in the presence of multipath
interference (self-interference by the signal being reflected off large
terrain features or other objects, the symptom of which is heavy
ghosting in analogue TV). At high carrier to noise levels, COFDM
exhibits a 3 dB advantage over 8-VSB in multipath rejection. On the
other hand, where multipath interference is low, of the order of -20 dB,
COFDM demands a carrier to noise ratio typically 4 dB higher than
8-VSB. In the presence of impulse noise interference, COFDM may require
up to 10 dB higher signal levels than 8-VSB to perform properly.
COFDM advocates argue strongly that most TV receivers
are in dense urban areas where multipath is the main problem, whereas
8-VSB advocates point out that cheaper antennas and lower transmitter
power suffice for good 8-VSB reception, compared to COFDM reception. A
city viewer is thus likely to do better with COFDM, whereas a rural
viewer in an area without large hills would probably do better with
8-VSB. The reality is that under signal conditions where good analogue
reception exists, both standards will perform well.
While both sides can score some technical points, the
reality is that the real issue in the debate is that of whose
manufacturers get the first and biggest bite of the market. This style
of acrimonious argument over which way the pie gets carved up differs
little from Australia's HDTV debate.
What Do I Get With
As interesting as the technical debate may be, it is now
appropriate to explore the consumer side of the equation.
One of the biggest issues which has percolated to the
surface in the HDTV debate, especially in Australia, is that of what do
I get for such an expensive product ? The mass media have not been short
of opinion on this issue, not surprisingly very often precisely
reflecting the commercial agendas of their respective owners.
The big difference we will see with HDTV is a dramatic
improvement in the quality of the picture and the sound presented.
Advocates of HDTV correctly point out that HDTV picture and sound are
cinema quality. Never mind the woeful content !
The first improvement we will see is that digital
transmission eliminates the noise and ghosting artifacts we have put up
with in analogue TV for half a century. This is tremendous leap in
technology, no different from the introduction of error correcting
The second improvement is that the 3:4 aspect ratio 25
frames/sec picture is replaced with a range of possible formats, and
more flexibility in picture frame rates. One of the objectives of the
Grand Alliance effort which set the context for the whole HDTV
development process was to accommodate the various standard cinema
aspect ratio and frame rate formats, so that the full range of celluloid
stocks could be presented without the nasty problem of frame rate
conversion and cropping to fit the 3:4 aspect ratio tube. A minimal HDTV
requirement is to support a 16:9 aspect ratio picture.
The third improvement is that HDTV uses a frame buffer
scheme and can thus present a very sharp and stable picture, since its
display hardware has more in common with a desktop computer than a
classical analogue TV set.
The fourth improvement is sound, and one not to be
scoffed at. Australian HDTV was to use the Dolby Digital Broadcast (DDB)
scheme (www.dolby.com), A.K.A AC-3 or Dolby-D. The Dolby model is based
upon the idea of adaptively allocating more bits of the transmission
channel to those portions of the signal which are more readily
perceived by the human ear, to achieve a 15:1 compression ratio. The
DDB system can provide for up to 3 forward channels, two rear channels
and a Low Frequency Effects channel to drive a woofer with 3 Hz to 120
Hz signals. A 24-bit dynamic range is supported (cf CD at 14 bits).
The fifth improvement is the potential for significant
greater intelligence in a TV set, due to the use of a wholly digital
platform with a common frame buffer for the screen display. A HDTV set
will be able to double up as a web browsing platform or computer display
simply due to its basic architecture. Many of the proposed interactive
TV/web schemes rely fundamentally upon this idea.
This aspect of HDTV raises other important questions. If
you put a HDTV quality monitor on a PC, fit a HDTV receiver/decoder card
into it, is it a TV or a computer ? Recent experiments performed in the
US involved the transmission of an ATSC encoded MPEG-2 HDTV picture
stream over a TCP/IP channel, layered on top of an optical fibre telco
link. As decent HDTV quality requires only around 20 Mbit/s throughput,
a household wired with a 100 Mbit/s LAN could probably support several
HDTV channels concurrently. Contention over the family room TV ? Banish
the dissenters to their study to watch it on their computer.
The common thread in the domestic and international HDTV
debate has been the issue of cost to consumers. When HDTV was conceived,
one of the background aims was to push large area display technology
into the high volume consumer marketplace to push unit costs down. This
was seen to be pivotal to reducing the costs of high resolution large
area displays for computers. Another background aim was to facilitate
the penetration of TFT LCD displays, and similar low voltage flat panel
display technology, into the volume consumer market. The computer
industry would simply ride on the coat-tails of the consumer boom and
benefit by using HDTV display technology on the desktop.
I for one would be very agreeable with the idea of a 30
inch diagonal 2:1 aspect ratio TFT LCD on my desktop. The ergonomic
gains are considerable.
To date the primary cost driver in HDTV sets has been,
not surprisingly, the display technology. If the set uses a CRT
solution, it will require the same technology as used in top end
graphics monitors which as we all know, are not cheap. If it uses a
plasma or LCD display, the poor production yields and low volumes in the
current computer market also mean that it will not be cheap.
It is a classical chicken and egg problem. The computer
market cannot support the volumes required to drive the cost down to
something appealing to HDTV users. The slow build up in the consumer
HDTV market has made it difficult for many manufacturers to amortise
large display R&D and tool-up costs.
However, once some critical mass in HDTV set sales is
crossed, all of this will begin to change very rapidly, since an
affordable HDTV set also amounts to an affordable large screen computer
Other important pieces of technology will also
contribute to the HDTV and computing synergism. DVD technology and high
capacity digital tapes will at some stage be capable of supporting the
required 20 Mbit/s data rate at an affordable cost. Whether the
computing market or the HDTV market drives this technology to that point
first remains to be seen.
What is clear at this stage is that both the HDTV and
computing markets stand to significantly benefit from technology being
developed for either market. Indeed a future device with a high
performance CPU and HDTV display and decoder hardware will be difficult
to accurate label, especially if it uses an industry standard operating
system and can accommodate a mouse, keyboard and network connection.
Perhaps the biggest problem HDTV has is that of
unrealistic expectations. This is true, arguably, of almost every single
group which has contributed to the debate. HDTV has the potential to
benefit consumers and industry, but for that potential to be realised,
all players must contribute their bit. So far this has not been