ÐÓ°ÉÔ­´´

The consumer electronics industry wants next year to mark the beginning
of the end for print and paper publishing. Instead, it foresees text, pictures,
animated graphics, moving video, encyclopedic text and sound – collectively
known as multimedia – being ‘published’ on interactive compact discs. Factories
will mass-produce the discs for well under £1 each, using the same
pressing machines as they do for music CDs. They will come in two sizes,
8 and 12 centimetres, depending on the amount of information they hold.
A player, connected to a TV screen and a stereo sound system, will unravel
the multimedia mix under the control of a simple hand-held unit.

There will be mains-powered, tabletop units looking like today’s CD
players, for home and office use. There will also be battery-powered portable
units for people on the move. The discs will offer entertainment, education
and what the industry now calls ‘edutainment’, an undefined mix of each.

First, though, the industry must sell domestic consumers the concept
of multimedia, while at the same time fighting a standards battle over the
technology. There are already half a dozen multimedia systems ready to compete
for the prize of becoming the de facto world standard, and as many more
in the pipeline. Most are incompatible.

The technology was supposed to be on sale in the shops five years ago
but so far only business users have been able to take advantage of it. The
delay occurred because it proved far more difficult and costly than expected
to develop software that does justice to the hardware. For example, an encyclopedia
system must do more than simply transfer text and illustrations found in
a book onto the screen. It should let users search through quickly, using
key words to find the topic they want with the level of detail required,
aided by spoken explanations, music, illustrations, animated cartoons and
sound effects. Business software is less demanding, principally because
it does not have to be entertaining to a wide audience.

Another reason for the hold-up was that the industry wanted to sell
units capable of playing 12-centimetre discs, which store more than an hour
of digitally coded video and sound, known as Full Motion Video. The technology
for this attractive extra feature existed but manufacturers were reluctant
to use it until there were international standards covering the recording
of the digital code on the discs. But even when the standards did emerge,
they ensured that manufacturers worked from the same ground rules, not that
the ensuing products would be compatible. For example, the VHS and Beta
analogue video systems met the same recording standards but, in operation,
were not at all compatible. Only VHS survived.

Digital coding standards were agreed late last year and the multimedia
units due to go on sale in Europe early next year will offer FMV. In the
US and Japan, however, the units could be on sale much earlier. In an attempt
to secure a market for their products, manufacturers say they will be in
the shops before the end of the year. Though these units will not offer
FMV, manufacturers have promised that consumers will be able to upgrade
the machines at a later stage. They have not disclosed how much this will
cost.

Unlike analogue video tape or disc, a digitally recorded FMV disc can
be played anywhere in the world, regardless of the local TV standard. Although
further advances in technology are needed before a 12-centimetre disc will
be able to store a full-length feature film, FMV should be able to offer
similar quality pictures and stereo sound. FMV discs will also be more compact.

Despite these advantages, the multimedia industry is still looking for
what it calls a ‘Killer App’, a feature or application that will persuade
us that we can’t live without a unit. This irresistible piece of software
will probably be a thoroughly nasty video game that uses FMV to blend live
movie action with animation, music, sound effects and dialogue. Players
will be able to simulate launching missiles at real cities, ambushing Rambo,
challenging Rocky to a title fight, hijacking a plane, robbing the Bank
of England or cheating Wall Street brokers with some insider trading. The
industry excuses itself by insisting, rather grandly, that once people have
indulged their fantasies with a Killer App, they will turn to more worthy
software.

The first multimedia system simply buried text and graphics in a music
CD. The standard for audio CDs, as defined by Philips and Sony 10 years
ago in the so-called Red Book, sets aside 97 per cent of the digital code
for music, and the rest for data. The data not needed to control the player
displays text and crude graphics, such as the lyrics of a song and biographical
information on the artist. The system, called BackGround Video or CD plus
Graphics, was never a commercial success. It seems that consumers were simply
notimpressed enough.

In the mid-1980s, Philips and Sony produced another standard, the so-called
Yellow Book, for using all the storage space on the disc, around 650 megabytes,
for data. The disc thus became a read-only memory with the capacity of 1000
floppy discs. A CD-ROM can hold 200 000 pages of text, 10 copies of a 20-volume
encyclopedia, a celestial atlas of the northern or southern hemisphere,
all the telephone directories for a European country or all its ordnance
survey maps.

Although the Yellow Book standard lays down strict rules for the method
of coding the data on a CD-ROM, it is deliberately vague about how computer
software should arrange and locate the coded data. This is to give the software
industry the freedom to find the best ways of doing it. Nor is there a standard
for the way a CD-ROM player, or ‘drive’, is connected to, and controlled
by, a computer. This means that many systems are now incompatible. Despite
the promises in the glossy publicity for CD-ROM products, a personal computer
must often be ‘retuned’ to play different CD-ROMs. For instance, an IBM-compatible
PC system laboriously set up to use British Telecom’s Phone Disc, which
holds all the telephone numbers in Britain, refuses to play the CD-ROM storing
last year’s issues of The Independent newspaper.

* * *

All work and no play

Some companies welcome the fact that CD-ROM is not a ‘buy and play’
system, capable of using any disc, because this stops office staff playing
games on a computer installed to run business applications. But people who
want to use a wide range of software on the same simple machine, without
wasting time tinkering with it, have been scared off CD-ROM technology.

In an effort to allay these fears, Philips and Sony decided to distinguish
between professional CD-ROM and interactive systems for the home. They followed
the Yellow Book with the Green Book, which defines the standard for CD Interactive.
CD-I was designed from scratch as a buy and play system for the home. The
standard requires the player to use a Motorola 68000 series microprocessor,
using the OS-9 operating system developed for fax machines and scientific
devices. The player plugs into a hi-fi system and a TV screen, which displays
commands that are selected with a cursor moved by a remote control. For
more serious applications, the player can also be connected to a computer
keyboard.

But making the distinction between users in the home and the office
proved a tactical error. The new standard hindered the development of CD-ROM
and professional users of interactive systems turned up their noses at CD-I,
branding it a ‘Mickey Mouse’ games format. Nevertheless in March this year,
Matsushita, the world’s largest consumer electronics company with brand
names such as Panasonic and Technics, formed the CD-I Consortium with Philips
and Sony to develop the technology. Sony, however, now seems to be having
second thoughts about the viability of CD-I as ‘edutainment’.

Three years ago, in an effort to forestall the crisis of confidence,
Philips and Sony began another venture with Microsoft, the American software
designer and the originator of the MS-DOS operating system used by most
PCs. The three announced CD-ROM-XA (extended architecture), based on both
the Yellow Book and the Green Book and designed to provide a ‘bridge’ between
CD-ROM and CD-I.

The XA Bridge, as the hybrid standard is known, defines how to store
pictures and sound on a CD-ROM disc, along with the text and graphics for
which the system was designed. It also specifies the layout of data, which
includes pictures, sound, text and graphics, on CD-I discs. This will allow
future CD-ROM systems, and CD-I systems when they become available, to play
so-called Bridge Discs and, eventually, should bring all the systems together
in one compatible format.

Another advantage of XA Bridge is that it has let Kodak and Philips
develop Photo-CD, a variation of CD-ROM but one that is compatible with
CD-I. Photo-CD allows photographers to shoot their snaps on film and, after
the film has been processed, store the images on a 12-centimetre CD for
about £10. This disc, which holds about 100 images, will play on either
a Photo-CD player, a CD-I player, or future CD-ROM system, which will display
images on a TV screen. However, the picture quality stored on disc is higher
than the quality available from the best TV screen, even a high-definition
TV screen; to appreciate the real quality of their snaps, photographers
will still have to make paper prints of the pictures stored on disc.

Kodak plans to launch Photo-CD next year, with players made by Philips.
Even if amateur photographers show less interest in the system than Kodak
believes they will, professionals already see its potential. For instance,
Photo-CDs will allow picture libraries to send clients a selection of low-quality
images down a telephone line for them to choosefrom. Kodak is also convinced
that pre-packaged Photo-CDs will follow, with collections of prize-winning
photographs or stills from films. As a bonus, the Photo-CD player will play
audio discs through a hi-fi. So will the CD-I player. But the Photo-CD player
will not play CD-I discs.

The industry’s promotion of Photo-CD is particularly significant because
the system relies on recordable disc technology, for which Philips and Kodak
have been drafting a standard, the Orange Book. This standard, which is
nearing completion, will define the way the recorder stores the data on
a blank disc and puts a directory or table of contents at the beginning
of the disc, so that the player can quickly recognise the type and location
of data stored.

But the Orange Book will also make it possible to copy the recording
from one CD to another, thus opening the door to piracy on a much greater
scale than that which affects audio recordings. For instance, while music
CDs sell for around £10, BT charges £2200 for its Phone Disc
CD-ROM. Recent agreements to compensate record companies for piracy offer
little consolation to multimedia software designers. In the US, the electronics
industry will soon add between 2 and 3 per cent to the price of digital
recorders and blank media but there is still no mechanism for distributing
the revenue within the music industry. The Serial Copy Management System,
adopted by the music industry along with the ‘compensation tax’ in a series
of agreements over the past two years, will be of even less interest to
the software industry. SCMS circuits in a digital recorder stop it making
a digital copy of something that is itself a digital copy of an original
commerical release bought in the high street. But SCMS can do nothing to
stop someone making a string of copies of the same original onto different
blanks.

If the multimedia industry wants to prevent piracy, it will need to
devise its own protection system. So far there is no sign of any policy.
Instead, there are individual responses that appear clumsy. For instance,
when someone buys a copy of BT’s Phone Disc they must ring a ‘help line’
for a password that applies for a limited period and to only one disc and
one CD-ROM player. Clearly, this cannot work with mass-marketed buy and
play software.

When the first CD-I standard was set, manufacturers assumed that the
discs would never be able to store video. Simple arithmetic seemed to bear
this out. Audio CDs, on which CD-I discs were based, were designed to play
for up to 75 minutes, which fixed both the size of the individual data pits
on a disc’s surface and the speed at which a laser read them. These factors
limited the rate at which information could be read to around 1.5 megabits
per second. While this is just fast enough to reproduce stereo sound, it
is nowhere near the speed required for a sequence of moving pictures. For
instance, European TV screens display the equivalent of 200 megabits per
second, made up of 25 pictures each containing 8 megabits of information.
Even this rate, however, compares unfavourably with the amount of data in
a 35-millimetre slide of about 80 megabits.

Early multimedia discs compromised by displaying crude motion using
a few images per second in a small area of the screen. Later discs, such
as one for improving your golf, skilfully blend animated figures with a
stationary background, though no one would dare call it ‘video’.

In early 1987, engineers from RCA’s Research Centre in Princeton, New
Jersey changed the rules of the multimedia game by demonstrating full-screen
video from a 12-centimetre CD-ROM, which they called Digital Video Interactive.
The breakthrough came when the researchers discovered they could reduce
the data rate of video to less than 1.5 megabits per second, the maximum
for CDs, by coding only the changes in a sequence of 30 images per second,
which is the broadcasting standard in the US. They broke down each image
into a mosaic of 250 000 individual picture elements, or pixels, and then
compared the pixels of one image with those in the next. Intel, the manufacturer
of the microprocessor chips used in IBM PCs and compatible machines, bought
the rights and refined the system. This year, Intel began selling plug-in
circuit boards that enable standard PCs to generate multimedia displays
from data stored in solid state memory, magnetic discs or a CD-ROM. Though
the system is used only in industry, with DVI circuit boards costing around
$5000 (about £3000), Intel plans to produce a cheaper system for
consumers, perhaps by the mid-1990s.

Following the DVI breakthrough, groups of researchers around the world
began to devise their own ways of compressing data to break the 1.5 megabits
per second barrier – and nearly 20 methods emerged. All had basic features
in common but differed in detail. Companies were concerned because a mess
of conflicting standards already existed in the closely related field of
still picture compression. The telecommunications industry was also worried,
because it relies on data compression to send still and video pictures over
telephone lines and satellite links.

As a result, three years ago the Joint Photographic Experts Group and
the Moving Picture Experts Group of the International Standards Organisation
(ISO) began work on a world standard for compressing and coding still and
moving pictures. With the core of the standard now agreed, electronics companies,
led by C-Cube and Motorola of the US, have started to produce microchips
for CD-I players that will code and decode images.

Under the new standard, video data is compressed in two ways: spatial
compression works on the individual stills and temporal compression on a
sequence of stills.

For large areas of a still image without spatial detail, such as sky
or sea, the digital code needs to define only the boundaries of the area.
(Facsimile machines already work in this way, taking less time to transmit
the blank lines between text than the text characters themselves.) An image
is divided into blocks with each block usually made up from an 8 by 8 mosaic
of 64 pixels. The coder takes an average reading for the brightness and
colour of the whole block and then logs the differences at each pixel. It
also logs the difference between the average values of adjacent blocks.
This reduces the code needed for a still picture by a factor of about 50.

Temporal compression can triple this gain by comparing the same parameters
for the content of each block in successive pictures and coding only the
differences. Some pictures must be fully and accurately coded to act as
references. Usually one fully coded picture is followed by seven pictures
coded only with information about the differences. To reconstruct images,
a player compares two fully coded pictures with the seven partially coded
pictures in between.

Together, spatial and temporal compression can reduce the video data
rate to around 1.2 megabits a second, which leaves 0.3 megabits for sound.
With the latest audio compression techniques, such as those developed for
digital audio broadcasting and Philips’s new digital compact cassette, this
is sufficient for storing stereo sound of near-CD quality.

* * *

Going for a new image

There is a hot dispute over the picture quality that the multimedia
systems will be able to produce from FMV discs. Pioneer, whose 30-centimetre
Laser Disc stores two hours of analogue video, says that a 12-centimetre
CD will not rival its product this decade. Philips, the inventor of laser
disc technology, used to say the same. But after three failed attempts in
10 years to sell analogue video discs under the names Laservision, CD Video
and Laserdisc, it has now changed its tune. Unofficially, Philips hints
that it will have developed a 12-centimetre digital CD to rival 30-centimetre
analogue discs for playing time, quality and cost, by next year.

Another clue to the future comes from news of a further round of meetings
on data compression standards for video, starting in November in Japan.
These meetings will examine ways of improving the quality of FMV coding
to match broadcast television standards.

The Moving Picture Experts Group of the ISO must now define a standard
for a datastream running at between 4 and 6 megabits per second, four times
the current maximum rate for CDs. It is not an impossible objective. For
at least three years Nimbus, the British CD manufacturer, has been looking
at ways of quadrupling the storage capacity of CDs. In Japan, Denon and
Sony have already announced that they have developed prototypes of the ‘quad
density’ discs.

These systems rely on halving the 1.6 micrometre pitch of the spiral
track of pits recorded on the disc and halving the speed at which the disc
rotates. As a result, the size of the pits reduces by a factor of four,
which quadruples storage capacity. But it means that the optics in the player
must focus the laser beam into a much smaller spot than that formed from
the infrared lasers in existing CD players. Solid-state blue lasers can
do this but they are not yet commercially available. So quad density CD
systems use higher quality optics to tighten the focus of a red laser beam,
which has a shorter wavelength than the infrared beams currently used.

No one knows yet when quad density discs will reach the consumer market,
nor whether the interactive CD players due to go on sale next year will
be able to play them when they are available.

Commodore, the American computer company, saw the delays in the development
of CD-I as an opportunity to gain a foothold in the multimedia market with
a simpler system. In June last year, it announced Commodore Dynamic Total
Vision, or CDTV, which combines a Commodore Amiga computer with a CD-ROM
drive and looks like a video recorder or CD player. But the company’s rush
to the market has already backfired. CDTV players and discs have been slow
to reach the shops and those that have offer little more than the much cheaper
Amiga computer and software.

The real breakthrough for multimedia CD may well be the development
of portable players, such as the Data Discman that Sony started selling
in Japan last year with text discs for dictionaries and encyclopedias. The
machine, which goes on sale in Germany in November for DM998 (about £350),
has a built-in screen, sits comfortably on the lap or on a desk top and
is ready to use with the press of a single button. But the system is deliberately
incompatible with CD-I and CD-ROM to stop users making electronic copies
of its discs.

People who used to read while travelling now either listen to music
on a personal stereo or play with a pocket video game. The games may be
puerile but to many people they are addictive, and interactive machines
would be even more so. Both Sony and Philips have struck separate deals
with Japanese games company Nintendo. In addition to CD-I, CDTV and DVI,
at least half a dozen CD-based games systems are ready or in the pipeline.

But history tells us that in the consumer electronics market there is
no room for two (or more) competitive systems that are incompatible and
rely on pre-packaged software, irrespective of technical merit. The 45 revolutions
per minute music disc was launched to challenge the 33.3 revolutionsper
minute LP, but it survived only as an extended single. The compact cassette
killed off the eight-track cartridge. Optical video disc killed off RCA’s
Selectavision and JVC’s VHD. The four quadraphonic sound systems of the
1970s failed in face of Dolby’s cinema and video technology. The audio CD
killed off competition from a system developed by Teldec in Germany and
a Japanese variation of JVC’s VHD. The VHS home video tape system killed
off both Beta and V2000. Sky’s satellite service killed off BSB.

The battle to create a de facto standard for multimedia CD looks likely
to be as bloody as the most vicious video game. A video game based on the
interactive format war might itself become the elusive Killer App. The big
question then will be whether the winning format progresses beyond the stage
of playing time-wasting games to become the replacement for paper that the
electronics industry so confidently predicts.