‘Broadcast
Model’ or ‘Data Model’?
Demystifying Digital Cinema
Part 1: Why Architecture Matters
by Michael Karagosian
NATO Digital Cinema Consultant
In many
ways, the business issues regarding digital cinema are more
clear-cut than the technical ones. Normally, when a new technology
develops, the price tag gets more attention than the bits
and pieces that lie under the hood. But at the current stage
of digital cinema, it’s what’s under the hood that
will affect long-term pricing.
The goal
of this article, and of the short series to follow, is to
bring the issues of “system architecture” for digital
cinema into focus, as these issues will ultimately affect
the business of exhibition.
How does
technology affect business? One way is through the cost of
that technology. If a technology becomes necessary but unaffordable,
the options available to the business owner utilizing that
technology will significantly diminish. While one might think
that achieving the lowest cost is everyone’s top priority,
this is not always the case.
Many companies
have a stake in a particular technology, either because they
own intellectual property associated with their product, or
because they have made a significant investment in a particular
style of doing things and it’s too expensive for them
to change. For some, acknowledging that a better method exists
might eliminate the need for their product in this market.
The technology game is a high-risk game. Low cost isn’t
always the prime motivator.
Couple
this situation with the smallness of the digital cinema market.
There are only about 150,000 movie screens worldwide, and
it’s unlikely that only one vendor is going to be providing
digital projectors for all of them. Unfortunately, one cannot
develop technology for digital cinema and expect to reap the
huge profits often associated with the high-volume consumer
market. In a low-volume market, costly research and development
must be avoided if product costs are to be kept low.
The
disparity between low- and high-volume pricing can be dramatic.
At consumer volume levels, specialized personal computer chips
can be created for less than $100 – and sometimes less
than $10. It isn’t possible, however, to develop complex
semiconductor chips at such cheap prices for a low-volume
market. Generally, low volumes cause custom chips to run thousands
of dollars per unit. If digital cinema systems are to be sold
at $40,000 or $50,000 each (the stated goal of some manufacturers),
a different approach must be taken.
It follows that when considering the practical issues in achieving
low-cost technology, digital cinema cannot be a technology
leader. Digital cinema has to be a technology follower, employing
high-volume low-cost technologies that were developed and
paid for by other markets.
Herein
lies the conundrum for digital cinema. How does one introduce
a new technology whose image quality exceeds that of the home
system, while at the same time using commonly available technology?
The answer lies in good “cost engineering,” an attribute
of a good system architecture. In today’s prototype systems,
we may not see much in the way of price difference between
architectures; today’s once-in-a-while sales volume requires
that all prices remain high. But when rollout begins and the
volume of sales goes up, system architecture will be a factor
in defining how low prices can go.
In
the field today, two digital cinema system architectures have
been introduced. One is the “push model,” also called
the “broadcast model.” The other is the “pull
model,” or “data model.”
The “broadcast
model” gets its name because it has its root in broadcasting.
It relies heavily upon a device called a “server.”
You’ve probably heard the term “server” used
by your information technology (IT) department when describing
the central computer of your office network – the root
of all evil when your e-mail system goes down. But in the
broadcast world, “server” takes on different meaning
in that this server also contains a “player.”
The “player,” regardless of system architecture,
is the device that “plays” audio and video. In terms
of your desktop computer, you can think of this “player”
as being similar to Windows Media Player or RealOne Player.
The “player” is the device that accepts “raw”
digital data and produces a “synchronous stream”
of video and audio – video and audio you can look at
and listen to. The word “synchronous” indicates
that all signals follow a required cadence. The word “stream”
is used to imply a flow – one may think of a stream of
water, but in the digital world, the stream is a stream of
bits. “Synchronous streaming” is important in the
broadcast world, as it’s required of your television
set in order to produce image and sound in your living room.
In other
words, your television can’t accept “raw” data,
as it doesn’t have a “player” in it. As a result,
the “player” has to be located in the broadcast
server back at the transmission studio. In the broadcast model
of digital cinema, a synchronized image stream leaves the
broadcast-style server and goes to the projector, while a
synchronized audio stream goes to the audio processor. (See
figure 1.)
With the
“data model” of digital cinema system architecture,
the “server” does not include the “player.”
Instead, the “player” is located at the receiver
end, typically in the projector. (See figure
2.) The data server will produce a stream, but
unlike the broadcast server, it is not synchronous. You can
think of the data server as producing “raw” video
and audio data. In a home system, for example, the data server
can be located anywhere in the world while feeding the “player”
in your desktop computer with “raw” video and audio
data over the Internet. Your television set cannot play this
data, however, since it doesn’t have a “player.”
Just as an ordinary television set cannot play the “raw”
data from a data server, a data server in a digital cinema
system will not work with an ordinary video projector. A projector
with an internal or outboard “player” is required.
So is
an exhibitor better off with the “data model” or
the “broadcast model”? Before answering that question,
we have more issues to touch on. These additional issues are,
in fact, so complex we’re saving them for next month’s
column.
To close
this first installment of our “demystification”
series, it’s important to note that the manufacturer
identities behind the different models discussed is not important.
Certainly, if one architecture shows promise over another,
the companies involved have the option of introducing the
most competitive product they can. By discussing these issues
openly, however, we help bring an important topic in digital
cinema to light: the long-term goal of producing affordable
digital cinema products for the exhibition market.
Next
month: Demystifying Digital Cinema – Part 2: Less Can
Be More.
