In most cases, waiting for a search result on a web site involves watching a blinking cursor or a magnifying glass. If you visit the National Film Board, however, it looks a little different: a white silhouette of a movie reel against a white background. This, as it turns out, is the least of the customization behind the organization’s massive digitization project.
For the last five years, the NFB has been busy putting together the technology, the processes and the policies that will fundamentally charge the way films in this country are produced, collected and stored. This includes a collection of more than 13,000 works, from documentaries, cartoons and short films to full-length features.
According to Luisa Frate, the NTB’s director of finance and technology, the digitization project came from a realization that the NFB was now operating in a world in which everyday Canadians have become accustomed to finding video content on demand, on any device, wherever they are. The problem is that unlike the average Hollywood blockbuster, the NFB’s collection needed its own particular set of metadata.
“Each one of our production has about 10 different versions — English, French, short, long, and so on,” she explained. “There are so many different aspects and combinations. We wanted to reflect that in our digitization platform.”
The project lead to an approach whereby every time an NFB production is digitized, partners and clients who request a film can have it practically custom-assembled with multimedia elements such as titles, sound, dialogue, subtitles and images. For example, a client could request an English version of a film with French titles, or the reverse.
Why ‘Store and Forget’ was not an option
Archiving digital content is an ongoing challenge for many organizations. This is not only due to the volume of content in some cases but the fact that formats change, and ensuring the long-term accessibility and quality can be uncertain. Jimmy Fournier, head of the Research and Development department of the NFB, said the organization tries to stay ahead of that by adhering to to four ‘golden rules’ of archiving. These include:
- There must be a process to continually check the integrity of the data which has been stored.
- Open file formats should be used whenever possible, in order to avoid frequent data migrations.
- Obsolescence of the storage hardware should be assumed as inevitable.
- Two copies of all content or media assets should be maintained on different technologies, in different locations.
Fournier said this last rule is probably the most critical. For example, the use of LTO tapes in a tape library is common practice in the film industry, but now the NFB is using ASG’s Digital Archive, which has a life expectancy of 50 years based on Sony’s Optical Disk Array. Backup copies are still stored remotely on LTO, but as analog media degrades over the next two decades the NFB will still be able to ensure its collection is intact.
“In the beginning it was great, because we worked with the actual collection — physical elements we could go back to,” Frate said. “If anything happened in an LTO, we could go back to that physical element. As we produce today, we are completely on a digital file-based workflow. There’s no physical element.”
The most difficult aspect of the project was establishing the metadata and ingesting it into the workflow and the archiving system, Frate added. As a result, the NFB has started to recognize its collection as not merely a set of films but a far more intricate series of valuable components. “The archiving system allowed us to think beyond the film,” she said.
The longer term challenge, she said, might be helping those making NFB projects — from directors to producers and even editors — that they need to think “digital-first” as they create a film.
“It’s a complete cultural change, and it’s not an easy thing to do,” she said. “Creators never want to share their content until everything is completed. This is a very open way of thinking. We can ingest content as we produce it and all kinds of surrounding content, which means more security but also more visibility on productions that are not completely finished.”
