How I Learned to Stop Worrying and Love Digital Archives: Digital Archiving Practices at the Swedish Film Institute
- How I Learned to Stop Worrying and Love Digital ArchivesDigital Archiving Practices at the Swedish Film Institute
Digital archives are often perceived by film archivists with some skepticism regarding the reliability and trustworthiness of digital technologies for long-term preservation (more than twenty years). Today, there is urgency to digitize our film heritage in order to give access to film to as wide an audience as possible, and short-term access is not considered a problematic practice. However, it is perceived that storing digital data on tapes is unreliable, risky, and expensive. Film archivists are operating mostly in an analog environment, and a big question mark hovers over digital-born films that do not have the guarantee of being stored in a safe, cheap, and cooled analog archive.
Yet, there are people convinced that it is possible to build up a reliable, long-term digital archive with an adequate initial investment, some technical knowledge, and appropriate personnel training. The Swedish Film Institute launched its digitization project five years ago, and the technical supervisor of the digital archive, Per Legelius, has no doubt that, following the best practices, there is nothing to fear. I met with Per to learn more about the practices followed at the digital archive of the Swedish Film Institute.
So, what is the purpose of your digitization project?
We digitize both for access and preservation. Our digitization project is a two-headedmonster!
The digital archive leaves behind film cans and dusty shelves. What are its elements? How is it organized?
We save in our archive raw scans and raw audio files, the restored version, a DCP, and a ProRes file. All that is packed into the tapes, throughout a “subformat” used to pack data into the tapes. The tape library, also called “robot,” consists of the physical enclosure with all the tape slots and a sort of arm that physically grabs the tapes. Then, we have software controlling all the operations, with database pointers.
What kind of technology do you use for the storage?
We use two different tape technologies: Oracle T10K and LTO. We deliberately decided to use two different technologies for safety reasons. We are using new drives, to be able to put more data on each tape; so right now, we are in the middle of a migration process, from LTO6 to LTO8, in order to store up to twelve terabytes per tape. The tapes are formatted with the Oracle DIVArchive system. We chose the Diva system because it is very reliable, it fulfills all our needs, and it is specifically designed for media files.
What format do you use to store the files?
We were very careful in choosing our file format for the raw scans and the restored versions. Moreover, we decided that the Master Archive Package (MAP) best fits our needs. MAP is a container developed a few years ago by Fraunhofer IIS, and we jumped on that because of the several advantages it gives us. First of all, it uses Jpeg2000 lossless, open source compression, which permits saving a massive amount of storage. Furthermore, the way the files are stored is almost identical to a DCP: we have one track for each reel, one file of audio and one of image, so that we do not get one file for each picture, as it is for a DPX, and the image and the audio can be kept synchronized. But, unlike the DCP, we are not limited in resolution; hence, we have no boundaries and no color encoding: we can put the same technical characteristics of our scans in the MAP format. Conversely, in a DCP, the color space conversion is never 100 percent transparent. If it happens that, for some reason, we must modify a film’s section, we can easily restore the MAP files, convert it in DPX, take out the area that has to be modified and reencode [End Page 144] it back to MAP. Due to it being lossless, we can do it as many times as we want. I do not know what other archives are doing, but...
