The British Film Institute’s (BFI) National Archive recently started a preservation project to convert 3PB of DPX film scans into FFv1 Matroska video files using automation scripts written by staff not formally trained as developers. This involves the use of RAWcooked, a lossless compression software produced by MediaArea, a consortium of developers and archivists producing tools with preservation interests at their core. Using RAWcooked the BFI has seen significant data size reductions which in turn has led to reduced network impact, provides an opportunity to review film scans via the Archive’s media asset management platform and will provide a long term benefit to the BFI in cost reductions for preservation storage. These benefits complement the primary driver for the project: digital preservation using a lossless, open, standards-based format that is increasingly adopted by public archives. This talk will draw on a recent blog which shares in full this project’s bash scripts, tools and approaches. Our mass digitisation project continues to reveal complications associated with upscaling automation of such diverse collections, and our scripts continue to respond to these changes. The open source preservation community actively develop scripts and software together, freely sharing knowledge to better improve the skills base of the sector. This methodology works against some historical tendencies for institutions to rely on broadcast solutions or vendor developed closed technologies. This talk reflects on this open source ethos, and the complications and benefits encountered by collections seeking stable futures by divesting from the built in obsolescence of commercial alternatives.
