BBC1 – Bright New World
by Brian Mason
From Eng Inf No.20:
The Quarterly for BBC Engineering Staff
is reproduced here with thanks to the BBC.
Recently many viewers
will have noticed a new look to BBC 1. For many years the channel logo has been
the familiar 'rotating world', generated from a mechanical model. Originally
this was produced by a remotely controlled camera, affectionately known as Noddy,
which was also used for the old mechanical clock and fault captions. Most of
Noddy's functions have been replaced by electronic generators, but the World was
still being produced by a caption scanner, followed by a colour synthesiser and
PAL coder. This required regular maintenance, and alignment of the video
processing to produce consistent results.
On February 18th a new
Symbol was introduced. The generating equipment is all electronic, using the
latest digital techniques. Some digitally generated pictures suffer from an
effect known as 'aliassing'; this is most noticeable on sloping lines and
circles as small steps, which show the graphic being made up of discrete lines.
The new equipment is fully 'anti-aliassed' on both the logo and the captions,
which substantially improves the overall quality.
By using internal frame
stores there 's no restriction on the colours, which can be properly shaded.
Since the colour information is internally stored there is no need for an
external synthesiser or clipper, thus reducing both day-to-day and long-term
variations. All National and Regional Centres which opt-out of BBC 1 have been
provided with their own equipment whose outputs, apart from their customised
caption, are all identical and of consistent quality.
The symbol generated by the equipment is a
rotating image of the world with a caption displayed beneath. The image is
larger than its predecessor, and there is no reflecting mirror, but the detail
and accuracy are much greater.
The caption is customised
for each of the various regions.
The symbol of the world
comprises three coloured parts. A gold shell, which is painted black on the
inside, with a shaded blue disc behind it. The sea areas are etched away,
leaving the land masses highlighted in gold on the outer surface, and black on
the inner surface. Where the shell is completely transparent, that is where
there is sea on the front and back, then the shaded blue disc is seen. The
outside is highlighted to make it appear as though it is lit by a spotlight
above the viewer.
Graphic Design at
Television Centre undertook the artistic design of the symbol. Clearly this had
to be done in very close conjunction with the engineers in Designs Department.
Over the last few years, with the design of several electronic graphic devices,
a good relationship has developed between the two disciplines. While neither
party fully understands the restrictions and principles of the other, each now
has a good grasp' of one another’s limitations. Interestingly this even
transcends the use of jargon.
developed the principles of the system. The most important part of this was the
data compression format for storing the map of the world at all angles. The
hardware development included EPROM based frame-stores, and digital processing
of the video. Software had to be written for the departmental VAX-II computer
for validating and processing the data into a suitable form for direct
programming into memory.
A third group in the
project was Computer Graphics. They have been extensively involved in past
projects, and for the BBC 1 symbol they wrote large amounts of software,
especially to generate the compressed data for each view of the world. This
processing and the transferring of data between their Quantel and VAX-II
computer took a long time and was mostly run over weekends, but the time had to
be carefully chosen so as not to interfere with election coverage in 1983 and
later the Olympic games.
Computer Graphics also
organised some early feasibility studies to evaluate the graphic principles. In
particular this showed that a 3D effect could be obtained even with an
The digital standard used
is that specified by the EBU for a digital parallel interface. While the device
has only analogue outputs, to suit present installations, the EBU specification
defines the required sampling rates and levels for luminance and chrominance. A
further advantage is that the Quantel Paintbox handles data in a form quite
close, but not identical, to this format.
To produce the overall
effect two full frame-stores are used. One is the foreground store which holds
the highlighted gold shell, and the other is the background store for the shaded blue
disc and the captions. These stores hold only a single frame, and there is no
restriction on their content. They are full colour and can display any picture
produced by the Quantel Paintbox. These stores are generally known as the
The main store of the system is known as the
'sequence' memory. This holds the data for the map of the world for each of the
600 fields that are displayed. This data is compressed by a coding system which
combines the benefits of traditional run-length coding, with the advantages of
pixel definitions. There is physical space for up to 7.5 Mbytes of memory,
although the addressing can access up to 16 Mbytes. For this application a field
of data is stored in less than 8 kbytes of memory space, as opposed to over 400
kbytes for a full field store.
The starting point for
the sequence data was a purchased data-base of a Mercator's projection map of
the world. This was edited by Computer Graphics to remove all political
boundaries, and transferred from their VAX-II computer into the Quantel Paintbox.
This was used for a 2:1 size reduction which incorporated the anti-aliasing
algorithms. The data was then transferred back to the VAX for encoding into the
Designs Department data compression format.
The two sets of fixed
data, the gold shell for the foreground and the blue disc with caption for the
background, were 'drawn' on the Paintbox. This data is properly anti-aliased at
source and was also transferred to their VAX-II. Further processing ensures that
no degradation occurs.
Both the fixed and
sequence data was transferred from Computer Graphics to Design Department on
magnetic tape, using the internal post!. This proved to be an extremely
efficient method of data interchange, far exceeding the earlier methods of paper
tape and floppy disks.
The memory structure is
similar to that used for the recently introduced digital Test Card F generator,
but each card can hold more data, and can be a part of a larger data-base. The EPROMs used are 27128 16 kbyte devices, although the memory card can take 27256
and even 27512 devices as, and when, they become available. When fully populated
with 27128s each card holds 0.5 Mbytes.
The controller unit has a 24-bit sequence
address bus giving access to 16 Mbytes, but the equipment has space for only 7.5
Mbytes, and is fitted for 5 Mbytes. The controller also addresses the fixed
memory through a 19-bit address bus. This memory uses identical cards to the
sequence memory and four are in parallel. A result of this is that the
customising for each region affects only the data on two cards in the system. In
fact, since only the caption is different for each, only 16 EPROMs are
The controller also decodes the sequence data
from its highly compressed format into a usable 13.5 MHz data stream, and
distributes timing information to the rest of the system. Timing control is
useful, since it can eliminate the need for external synchronising. This
equipment's output can be varied from over 6µs early to over 3µs late
relative to the mixed syncs input reference. Digital multipliers are used to key
the map onto the foreground and background data streams. These are full 8 x 8
bit devices, and correct scaling is incorporated to ensure unity gain where
necessary. The two keys are processed to prevent any excess amplitude after
The two data streams are
added digitally, before being blanked. Normally digital blanking needs to be
shaped to conform to PAL system I, but since all the signals are generated
internally, correct shaping is naturally included within the data. (In any case
the start and end of all lines are black.)
A new, triple-video,
analogue-to- digital converter has been designed to provide the YUV outputs and
an analogue matrix used for two sets of RGB outputs. Both these units employ
close tolerance components to minimise drift, and hence regular alignment. A
test waveform is included within the system for checking output levels and
The system also includes
a large power supplier, based on a commercial unit, and a BBC designed clock
generator which is common with other digital equipments.
The detailed design of
the system began in early 1984, with the requirement that the new symbol should
be ready to go on air by 1st January 1985. This included the slight complication
that not only London, but also eleven regions had to have their equipments
delivered and installed well before Christmas.
Taking a certain amount
of risk, some manufacturing had to be initiated before the prototype was fully
operative. This particularly. applied to the making of over 200 memory cards.
Existing designs and the large quantity of memory units were made by Equipment
Department, while the rest, including the crates, were made by the Production
Unit of Designs Department. The manufacturing process and system testing went
quite smoothly and all the regional units were delivered in November.
All the digital cards
that were designed for the system were prototyped and proven using the wire-wrap
technique. They were then laid out for pcb's using Designs Department's Racal
Cadet CAD system. Without this, it would probably not have been possible to
obtain the component and interconnection density. Certainly the accuracy of the
computer-generated artwork ensured high quality products requiring the minimum
of fault finding during test.
Perhaps the most time
consuming part was the programming of nearly 6000 EPROMs. This was facilitated
by the purchase of two specialised MOS PROM programmers.
One could be connected
through a terminal to receive data direct from the departmental VAX computer,
while the other could copy ten devices from a master EPROM. Without the VAX
facility, or something of similar power and capacity, it would probably not have
been possible to complete the software processing in the required timescale. For
example, more than five days of CPU time was used to process just the background
data for all regions.
This complex digital project has been
successfully completed on time. We were, to a certain extent, fortunate in being
able to obtain the required number of some components despite shortages. In
particular we received a great deal of help from Valve Section, who managed to
buy 6000 EPROMs in a depleted market being scoured by the hungry computer
The end result is a successful example of
co-operation between artists, programmers and engineers. The mutual respect
between these parties has been fostered over several projects in the past, and
it is to be hoped that it shall continue in the future. The techniques developed
for this project have applications for the future, and we can look forward to
new devices using similar hardware and data preparation.
Return to Technology Article list