‘Data
Model’ Carries Advantages
Demystifying Digital Cinema
Part 2: Less Can Be More
by Michael Karagosian
NATO Digital Cinema Consultant
Last month
we discussed how, in the realm of digital cinema, “system
architecture” is a good place for a manufacturer to start
its “value engineering.” Because the world contains
so few moviehouses, digital cinema products will not be manufactured
in the enormous quantities we associate with consumer goods
like TVs and CD players. This lack of a sizeable market dictates
that expensive research and development for digital cinema
needs to be avoided wherever possible, as it would significantly
add to the cost of the final product.
In the
same article, we presented the concept of the “broadcast
model” (which puts the movie “player” with
the server) and the “data model” (which puts the
“player” with the projector). What we didn’t
get into last month was how moving the “player”
from server to projector could simplify the “link,”
or “network,” connecting server and projector. The
simplification comes about because the required capacity of
the link and the need for stringent timing characteristics
across it are reduced by moving the “player.” This
simple move allows the link between server and projector to
be implemented with off-the-shelf computer technology.
Reducing
the cost of the link is not the only benefit of the “data
model.” This model also eliminates the need for a technology
called “link encryption.” To best understand link
encryption, we need to understand the security side of digital
cinema systems.
|
Digital
Cinema Event
REG Screens Live Football
DENVER – Some 90 people in Longmont, Colo., paid
between $6 and $10 to watch a historic live Sept. 7
college football match.
Backers
say it was historic because it marked the first time
a digital cinema system had projected in high definition
a live, American football game for a paying multiplex
audience. The UA Twin Peaks 10-plex in Longmont was
one of four Regal Entertainment Group (REG) venues in
the Denver and Los Angeles metropolitan areas to exhibit
the match, which pitted Colorado State University against
UCLA.
The
UA Pavilions 15-plex and the UA Greenwood Plaza 12-plex,
both Denver venues, and the Edwards Irvine Spectrum
21 in Irvine, Calif., were the three other REG cinemas
exhibiting the game.
REG
subsidiary Regal CineMedia (RCM), which oversaw the
theatrecast, conducted a similar alternative content
experiment in June when it screened a live concert by
the hard rock band Korn.
“Clearly,
the whole mission statement of RCM is creating different
uses for our theaters,” said RCM CEO Kurt Hall,
who also serves as its parent company’s co-CEO.
“Now that we have the digital capabilities, this
is the obvious place to start as certain sports programming
is very compelling on the big screen.”
DirecTV
Veteran
NewCo Names Ordway CTO
HOLLYWOOD – Hughes Electronics vet Walt Ordway
was formally named chief technical officer of NewCo
Digital Cinema (NDC) on Aug. 9.
“The venture makes great progress with the selection
of Walt Ordway as CTO,” noted NDC CEO Chuck Goldwater.
“Walt has a rich working knowledge of digital cinema,
and was in charge of the early Hughes Electronics Digital
Cinema project. Walt has the respect and admiration
of all the studio members as well as the engineering
community involved with digital cinema.”
“NDC holds the unique position of being able to
see a complete digital cinema system architecture specified
to the satisfaction of the studios, the exhibitors and
the manufacturing companies,” said Ordway. “I
am honored to be able to assist Chuck and the studios
in the development of digital cinema standards.”
Ordway most recently served as vice president in the
engineering organization of Hughes’ DirecTV. He
was also director of the engineering development lab
of DirecTV International. A member of SMPTE, he chaired
the Digital Cinema DC 28.4 Study Group, and is a member
of LFCA (Large Format Cinema Association). He holds
a master’s of science degree in electrical engineering
from Penn State University.
NDC was created last March by seven motion picture companies
– Disney, Fox, MGM, Paramount, Sony, Universal
and Warner Bros. – to “establish and document
an open architecture for digital cinema components that
ensures a uniform and high level of technical performance,
reliability and quality control.”
|
When a
digital movie is shipped from studio to exhibitor, it will
not be in a form that can be played in your computer or in
your home theatre system. The most fundamental reason for
this is that the digital movie will be “encrypted”
before it’s shipped. Encryption is the digital process
of making perfectly playable movies into digital data that
is complete nonsense. To make it useable again, one applies
a “decryption key” to the encrypted digital data,
allowing the player to convert it back to the original movie.
Yes, theft can be attempted on such systems by “hacking”
the encryption to discover the key. But if the movie is encrypted
well, it would take today’s computers many decades (and
with some methods possibly several millennia) to bust into
an encrypted file and recover the original movie data.
Because
of the movie encryption process, we can assume that the digital
movie is secure while stored as encrypted data in the digital
cinema system. Even if someone were to walk away with the
movie data, the encyption would render it useless to its possessor.
When it comes time to play the movie to an audience, the system
applies the decryption key to convert the nonsensical data
back into the movie.
A decrypted
movie is still not ready to play, however. Once decrypted,
the digital movie data must be decompressed by a decoder.
The output of the decoder is a synchronous stream of digital
movie data, ready for viewing. By synchronous, we mean that
the data follows a particular cadence, or timing.
Here’s
where the “data model” differs significantly from
the “broadcast model.” In the “data model,”
the decryption and decoding processes described are situated
in a “player” within the projector. In the “broadcast
mode”l, they are situated in a “player” within
the server.
From a
security perspective, the “data model” ships movie
data with the original encryption (and original compression)
from the server to the projector. Because the original encryption
is still intact, the data is secure as it travels across the
fiber or wire. Since this data is also still compressed, it
requires less bandwidth, typically by a factor of 20. This
makes it easier for the “data model" to utilize
off-the-shelf network technology to ship data from server
to projector.
The “broadcast
model,” however, has already decrypted and decoded the
data from the server before it leaves the server for the projector.
If nothing more is done, the movie data will leave the server
at full bandwidth and with no encryption. Thus, the movie
data will be unsecured, and vulnerable to theft. At the same
time, the broadcast server link requires a bandwidth that
is typically 20 times greater than that of the "data
model."
To
make this data secure as it leaves the broadcast-style server,
a process called “link encryption” is applied. In
this process, the movie data is re-encrypted before it leaves
the server, and then decrypted again at the projector. Thus,
the “link” between server and projector is protected
by specially encrypting the data that travels across it. Figures
1 and 2 depict the differences described between the “data
model” and the “broadcast model.”
While
“link encryption” helps secure the “broadcast
model,” it also introduces a few problems. One of those
problems is that it adds cost to the system. Additional modules
to support link encryption and link decryption must be added.
Also, a more expensive link between server and projector may
be required due to the higher bandwidth that must be supported.
There is also a danger in the “broadcast model”
in that the movie data is exposed, or “in the clear,”
after the original encryption is removed. This exposure creates
a vulnerability that a hacker could exploit. For instance,
a broadcast-style server that decrypts data in computer memory
could be hacked by using a “memory resident program”
– which could enter the system as a virus – to make
an “in the clear” copy of the movie data between
encryptions.
Are there
other cost benefits to the “data model”? By moving
the “player,” along with the accompanying decryption
and decompression modules, to the projector, a relatively
cheap server can be constructed for the theatre.
Are there
any cost penalties associated with the “data model”?
When comparing the projector used in the two server models,
the “data model” projector has only one more module
– the decompression module – than the “broadcast
model” projector. But since a decompression module must
exist somewhere within the digital cinema system, there is
no cost penalty to the system.
In summary,
the “data model” eliminates link encryption, and
allows the use of off-the-shelf network technology for shipping
movie data from server to projector. It not only reduces the
system cost somewhat in this way, but it also is inherently
more secure than the “broadcast model.” Thus the
title of this article: less can be more.
In our
next installment of this series, we’ll discuss how the
system can be designed so that it is forward-compatible with
future projection technologies. We’ll see once again
why it is important to stick to the “data model.”
