Using a Mixture of Technical and Diplomatic Skills, IBM Researchers and
Executives Helped to Create a Single New Format for Digital Versatile Disc
Technology, Thus Preventing a Rerun of the VHS-BETA Format War
In Brief:
During the past two years, a group of IBM researchers and executives has successfully negotiated the end to a potential format war among producers of digital video disc technology. Using a combination of diplomatic and technical persuasion, the team first convinced two consumer electronics consortia - led by Toshiba/Matsushita and Sony/Philips - to agree on a common format, called the digital versatile disc (DVD). Later, the group promoted an agreement between movie makers and the computer industry on a means of protecting intellectual property on discs without encumbering computer companies' ability to develop their technology freely. As a result of those efforts, the first DVD players went on the market late last year.
The success of a new technology sometimes depends as much on diplomacy as it
does on technical excellence. Had it not been for the behind-the-scenes work of
IBM engineers and executives, the digital versatile disc (DVD), which promises
to supplant compact discs, would never have reached the market. Consumers might
instead have been forced to choose between two incompatible storage
technologies. When one technology emerged as the winner, those who chose its
rival would end up losers - just like the unfortunate individuals who backed
Betamax technology rather than VHS two decades ago.
The latest format war threatened in the first weeks of 1995, when Toshiba
and Matsushita announced their plans to market a new type of compact disc.
Called the Super Density (SD) format, it would be capable of holding a
full-length feature film with digital-quality sound. At the same time Sony and
Philips, the companies that had invented the compact disc, announced their new
storage format, the Multimedia CD (MMCD). The two formats were remarkably
similar. Both were exactly the size of conventional CDs; players that could play
the new format would also read CDs. The formats also held a similar amount of
data. But they were completely incompatible with one another.
IBM had become involved with the new high-capacity discs in 1994. Philips
had realized that its new format held at least as much promise for the computer
industry as for the entertainment business. The majority of the 60 million
computers now sold each year come equipped with CD-ROM drives, which suggests
that the computer market for enhanced CD drives will far outstrip the market for
dedicated video players. The fact that the CD format was conceived for digital
audio caused some limitations when it was migrated to computers. Since Philips
wanted to avoid such problems with its new format, the company asked IBM, along
with a group of other major computer companies, to make sure that the new format
would be easily adaptable to computer use.
A committee called the Technical Working Group (TWG) was formed, initially
with representatives from five computer companies: Apple, Compaq,
Hewlett-Packard, Microsoft and IBM. The chairman was Alan Bell, an expert in
data storage systems and technology at IBM's Almaden Research Center in San
Jose. Today more than ten computer companies attend TWG meetings.
Surprise announcement
Bell and his colleagues on the TWG were analyzing the MMCD format when in
January 1995 they first learned that Matsushita, Toshiba and Time-Warner had a
format of their own. The announcement came as a shock. Technically, both formats
looked pretty good. The computer industry would have been happy with either,
given minor modifications. The problem was the existence of two incompatible
formats.
The first and most urgent recommendation of the Technical Working Group was,
therefore, not a technical one, and it did not make the big consumer electronics
companies very happy. "We need only one format; we don't want to have to
choose," Bell said. "Implicit in making a choice is that you may in
the end have selected Beta. We want just one. Work it out and present that to
us."
The example of the VHS-Beta war haunted everyone's thoughts. IBM in
particular, which sells millions of computers a year, did not want to be in the
position of backing the wrong horse. "If two years down the road the battle
is resolved and we end up having chosen 'Beta'" says Bell, "all
those consumers who don't get as much software as they expect to get won't call
Sony or Toshiba. They'll call IBM."
Modified modification codes
In September 1995, Bell flew to Hawaii with Victor Jipson, then manager of
the SSD optical storage business unit at Tucson, to listen to representatives of
both sides. The point of contention concerned what is known as the modulation
code.
The user's digital information is encoded before it is recorded onto an
optical disc as a series of pits. The encoding step is necessary because certain
combinations of 1s and 0s, which might occur naturally in the user data stream,
can make accurate playback difficult or even lead to the mistracking of the
laser beam.
The modulation code avoids these pitfalls by taking the stream of 1s and 0s
from the user and converting it to data that obeys certain constraints, such as
restrictions on the maximum and minimum number of 0s between two successive 1s.
CDs use a modulation code called 8-17, since it takes 8 input user-data bits and
converts them to 17 modulation-code bits on the disc. Sony and Philips proposed
an 8-16 modulation code, which clearly obeyed all constraints and gave a
one-seventeenth improvement in efficiency over the CD code. Toshiba and
Matsushita went one better with an 8-15 code. "It was more efficient, but
it wasn't clear that they could address all the constraints as well,"
says Bell. Their scheme would probably have worked, but it was less proven.
Bell and Jipson listened to both sides for two days and then conferred with
each other until three in the morning before contacting Patrick Toole by
telephone with their technical recommendations. The saving of one bit offered by
the Toshiba/Matsushita modulation code amounted to about a quarter of a gigabyte
more storage capacity-5 versus 4.7 gigabytes. "So the natural question
was: does anybody care about the quarter gigabyte?" Bell explains.
The answer was no. The 4.7-gigabyte capacity was enough to meet the
entertainment industry's requirement that the disc must be able to hold a movie
lasting 135 minutes, the time limit for 95 percent of all films. And while in
the world of computers more storage is always better, it was much more important
to the computer industry to have a single, reliable format. "We got away
from "yours doesn't work and ours does" Bell says. "It was
more an engineering judgment. In the end, even though there are a lot of
business and political pressures, there is a technical truth at the foundation
of our recommendation."
Ten days later, on September 15, 1995, both sides issued a press release
announcing that they had agreed upon a common DVD format. The official
announcement of the format followed on December 8. Minoru Morio, executive
deputy president of Sony Corp., acknowledged the role of the IBM team. "IBM
made a great contribution to this unified format," he said. Alan Bell
looked forward to returning the pager he had acquired during the long
negotiations and getting back to his office and the world of research and
development.
Invitation to piracy
This respite proved to be short-lived. In early 1996 Bell received a call
from Bob Lambert, an executive at Disney. Digital discs, as the computer
industry has long known, invite piracy and unlawful copying. DVDs, which by
design can be read as easily on computers as on dedicated players, were causing
nightmares for the people at Disney and other entertainment companies who try to
control unlawful copying of their intellectual property, i.e., the motion
pictures. Lambert told Bell that he thought they had worked out a solution to
the problem of copy protection and invited him down to Los Angeles to hear about
it.
As chairman of the TWG, Bell had got used to chairing meetings of 25 people
or so. However, all the participants in those meetings were engineers. Although
they came from 10 different companies, they were all more or less on the same
side. In Los Angeles he found himself in front of 50 people-half of them
lawyers-from huge entertainment, consumer-electronics and computer
companies.
Unlike adjudicating the merits of competing storage technologies, copy
protection is not a purely technical issue, because no purely technical solution
exists. Bell likens the situation to money counterfeiting. " You can do all
kinds of things to a banknote to make it harder and harder and harder to copy,"
he explains. "But it will never be impossible, because the incentive and
reward for breaking the scheme is so great." Technical solutions have
therefore to be buttressed by legislative barriers.
The scheme that Lambert and his colleagues described to Bell, called the
Copy Generation Management System, was very simple. Every recording would have
attached to it a simple two-bit code that could be read by every recording
device. If both bits were 1s, the device would simply refuse to make a copy. If
one bit was a 1 and the other a 0, the device would make a single copy but would
change the attached code to two 1s, and hence prevent subsequent copying. If
both bits were 0s, unlimited copies could be made.
As simple as the scheme may have been, it was unacceptable. Why? The Motion
Picture Association of America (MPAA) and the Consumer Electronics Manufacturers
Association (CEMA) wanted it extended to computers, requiring all computers to
have a dedicated chip to detect the two-bit code. Computers without the chip
would be illegal devices. "The fact that this would imply that by law
certain technologies would have to be present in all computers was completely
unacceptable to the computer industry," Bell says. "The unencumbered
and free development of technology is central to our industry."
In April, Bell explained to a 100-person meeting the objections of the
computer industry. The entertainment industry could have ignored the objections
and taken the fight to Congress. Instead, the MPAA and CEMA allowed the
technical people 30 days to find a better solution. They immediately formed the
Copy Protection Technical Working Group (CPTWG), with Bell as one of the
co-chairmen.
Erecting a speed bump
The group began meeting weekly, and within a month had come to some
conclusions. The first was that the best way to protect the data was to encrypt
it. Encrypted data can still be copied, but is useless unless decrypted. The
decryption algorithm would be licensed, so that computer manufacturers wishing
to provide playback capability for copy-protected content, as well as
manufacturers of DVD players, could include the necessary decryption hardware or
software module in their products. "From the technical viewpoint, we think
of this as a speed bump," Bell explains. "But it's also an approach to
which you can attach legislative power."
Once decrypted, the data would be further protected by compliance with the
terms of the decryption license. For example, should the digital data be
converted within the computer to an analog video signal for displaying on a
television, the video output signal is required to include analog copy
protection features, such as Macrovision. The Macrovision signal features
effectively prevent the recording of viewable VHS copies. A fundamental
advantage of making the decryption available under license is that it is a
voluntary scheme. Only computers intended to be able to view video would need
special decryption chips.
The scheme is not perfect; no scheme could be. But it does accomplish the
main objective set by the MPAA of preventing casual copying by consumers. "The
studios recognize that this copy protection scheme will not prevent piracy,"
Bell says. What it will do is keep honest people honest, and prevent them from
easily making copies for their neighbors. The scheme is a convenient hook to
which anticircumvention legislation can be attached. Such legislation is
currently being drafted by policy specialists representing the interested
parties.
Matsushita presented a detailed embodiment of the CPTWG's proposal. The
consumer electronics industries accepted the scheme, but at first Intel and
other computer companies objected.
Intel's concern was that the decryption added too much of a computational
burden to its processors. Already a processor would have to decompress the
digital video, which is compressed by an algorithm known as MPEG-2 to save space
on the disc. If the processor also had first to decrypt every frame of data, it
would have no computing power left for any other activity, such as receiving a
fax or printing a document in the background. The fragile alliance was once
again threatening to fall apart, causing further delay to the DVD rollout.
Within a few days, IBM engineers found a solution that was satisfactory to
all parties. Better still, it required only the smallest modification to the
existing scheme and could be implemented very quickly. What Bell realized was
that not every frame had to be encrypted to make the picture unviewable. The
digital data on the disc is stored in sectors, with about 10 sectors of data
needed on average for every video frame. Encrypting just one in every four to
six frames is enough to scramble the video data thoroughly. Since it needs to
decrypt fewer than a quarter of the frames, the processor has more time for
other tasks.
IBM researchers worked quickly to demonstrate the effects of partial
encryption to the CPTWG. As a result, the MPAA was satisfied with the
effectiveness of the picture scrambling, and Intel accepted the more
computationally effective method. The solution was announced to the world on
October 29, 1996. By early November, the first DVD players were available to
consumers in Japan. At the Consumer Electronics Show in January 1997 several
studios announced their intention to release motion pictures on DVD, signaling
their approval of the overall solution.
Beyond altruism
The efforts of IBM scientists and executives helped to create
the DVD format, and thus avoid a format war with consumers as the ultimate
victims. Their efforts also helped IBM. "The fact is, we're really in the
DVD business," points out Dan Sullivan, director of licensing development
at IBM's headquarters and a key player in the licensing and legislative
discussions relating to copy protection.
IBM currently does not plan to assemble finished DVD players.
However, the company holds patents on the optical technology that allows players
to read multilayer discs. To achieve their high storage capacity, DVDs must
record on two sides. Just like phonograph records, such discs would have to be
flipped midway through playback. In the multilayer approach, pioneered by IBM,
both sides of the disc are read from one side, without flipping. Since the discs
are transparent, the laser that reads the top layer can be refocused to read the
bottom one; the out-of-focus pits on the top layer do not interfere with the
reading of the bottom. Refocusing can be achieved in milliseconds, making this
approach ideal for computer applications.
IBM has also pioneered digital watermarking technology to
strengthen the DVD copy-protection scheme, as well as to extend it to other
content distribution channels such as satellite and cable. This technology
stamps each frame with a mark, invisible to the naked eye, that can be detected
electronically. The mark remains embedded in the content throughout transmission
and processes such as compression or digital-to-analog conversion. Future VCRs
and DVD recorders and players will contain the mark-detection module that will
prevent recording of the signal whenever the mark corresponds to "do not
copy," or playback of unlawful copies should they somehow be made on a
recordable DVD disc.
Alan Bell never did return his pager. He has become the program
director of IBM's newly formed DVD Project Office, and remains actively involved
in coordinating DVD-related activities around the company. What does he see as
the secret of his diplomatic success? "There's been no trick other than
stamina and trying to keep things simple," he asserts. "People
sometimes get angry and excited and say the wrong things because the issues are
so complex and frustrating. To succeed, you just keep returning as much as you
can to the basics."
Bruce Schechter is a science writer based in Los Angeles. He is
currently writing a book on the life and times of Paul Erdös.
