by Ben Worthen


Mar 15, 20022 mins
Data Warehousing

Deep inside a mountain outside Pittsburgh, there’s a climate-controlled vault that could be some superhero’s secret fortress. It’s actually the storage facility for the National Geographic film library, which includes more than 25,000 hours of footage such as Jacques Cousteau’s first television show in 1965 and the output of critter-cams, which are cameras mounted on the backs of whales and other creatures. Formats include 16- and 75-millimeter film, Betamax and digital video. Soon, however, everything will be available to download at the click of a mouse.

More than 2,000 hours of footage have already been digitized and loaded into a 5-terabyte database, says Matt White, National Geographic Television & Film’s vice president. National Geographic researchers give each shot meta-data tags that catalog the actual content, as well as other attributes. For example, a shot of a tree in the middle of the desert would come up under a search for the tree’s species, and under a subjective category like “lonely.” The 2,000 hours of video, which contain more than 300,000 individual clips, are available in Real Player and ASF formats so researchers don’t have to run to the vaults to physically check film clips to ensure the footage fits their needs. White says local television stations looking for clips of exotic animals, such as the recently discovered giant African crocodile, and researchers who want to count the number of times a critter-cam captured a whale’s fin flipping have already downloaded footage. White plans to work his way through the archive, converting 2,000 hours of footage a year for the foreseeable future. The archive, however, recently acquired thousands of hours of footage from the World Bank and intends to acquire more. “It’s a never-ending process,” he says.