It wasn't too long ago that keeping track of content involved a
videotape label and a Magic Marker. But now digital technologies, which move
content around in video files instead of on tape, have taken video tracking
well beyond the Post-It note.
For content owners, that requires “media-asset management” (MAM), a
multimillion-dollar proposition for stations and cable and broadcast networks.
To date, however, most organizations have been slow to embrace MAM because
it's still an ill- defined technology.
MAM means different things in different parts of a media organization.
Take the typical news department, for example. For the production side of the
house, content management begins before the news is even gathered: What crew
and reporter are shooting the story? Where and when is the story taking place?
What is the deadline? All that information is collected into a database that
will eventually be married with the incoming video to make it easier for the
reporter and editor to find the video footage.
The sales and traffic departments also rely heavily on basic content
management. For instance, syndicated programs and commercials include metadata
tags to identify the content, let users know when it should run, how many times
they can use it, and with what company the rights agreement was made.
Beyond those limited tasks, though, MAM inspires little interest. The
industry wants to change that.
Smaller operations usually decide it's too expensive. “Individual
stations are not pursuing MAM,” says Sarah Foss, head of media-asset
management for Harris Broadcast. They think it's a spending sinkhole.
But she sees it gaining traction with station groups. “We've seen a
lot of them talking about content management,” she says. Terrestrial
networking technologies are increasingly used among sister stations, providing
the backbone for easy file sharing. “They can share a program like
Oprah without all of them having to bring it in via
satellite,” Foss says, “and they can also collaborate more easily across
departments. MAM is really the glue and the middleware between desktop
operations and the IT department.”
One reason for embracing MAM is to make content retrieval easier.
Typically, that's done by tying closed-captioning or script information to
the video so that a user can enter text into search fields and pull up a
low-resolution proxy video or stills of clips.
Historically, however, such technologies have been an expensive
non-starter. In the late 1990s, a number of companies entered the MAM market
with products that used face- and voice-recognition technologies to facilitate
content cataloging. But the systems cost hundreds of thousands of dollars and,
more important, yielded little return on investment.
That history is one reason Harris now offers a MAM starter pack, at a
relatively low price point—about $60,000. “It allows broadcasters to do
some rudimentary low-resolution proxy generation and browse an archive,” says
Foss. The starter pack allows an organization facility-wide access to content.
The company can scale up from there, adding users, features and
Generally, MAM systems can range from $120,000 to the high six figures.
“It depends on how big and all-inclusive you want the system to be,” she
Those big price tags typically scare off broadcasters. So Virage, one of
the leading asset-management providers, has begun to offer affordable
Virage hit the market nearly 10 years ago with its Video Logger system,
which draws on audio tracks, time-code information and keyframe analysis to
demark scene changes. It also takes information included in lower-third
graphics that might identify a subject.
Virage also now incorporates speech-to-text technology from its parent
company Autonomy to make its metadata creation even more complete.
But probably most important, it's much less expensive than it was even
two years ago. Then, a Video Logger system cost $30,000. Today, the price is
$15,000. The VS Archive server, which stores the content, is priced at
“We can now scale to the size of the market,” says Hal Feldman,
director of sales engineering for Virage. “A small station with a couple of
hours of raw material a day can have one, two or three Video Loggers while a
large-market station can scale beyond that.”
Not only is the Video Logger cheaper than it used to be, it's also
better, Feldman says. The use of Autonomy's Softsound speech-to-text
technology (which automatically creates a script from the audio on a tape) is
available for 12 languages. “Within that product is the ability to home in on
a particular vernacular, like medical terms that may not be found in a standard
dictionary,” he says. “That helps give highly accurate results.”
Improvements have also been made on the search side of the system. Video
Logger can now provide what is called conceptual matching. In essence, it works
like a smarter Google. For example, in the past, someone who wanted to find a
story about a lost golden retriever that might be in a shelter might type in
the words “lost dog” in the search fields. But if those words weren't
used in the story, a match wouldn't be made. The new system, however,
understands that a golden retriever is another way of saying dog and shelter is
where lost dogs are kept.
Producers are finding a MAM system very useful. In the old days, they
would have had to call the librarian and try to find a story by saying
something like “Can you find a comment Bush made about Kerry? I think it was
in October and it might have been on ABC News, but I know it was on the White
“Now,” says Foss, “the librarian can do a search and get the
content immediately out to the traffic system.”