Abstract / topic summary
What does one do when tasked with preserving two years of continuous video material from the Library of the University of Zurich’s AV holdings, but have neither the budget for custom software nor the time for manual processing? You MacGyver a solution.
This presentation outlines a real-world, get-the-job-done workflow that turns an off-the-shelf computer into a powerful assistant for our digital lab. By combining Python, a suite of open-source tools (like FFmpeg or OpenCV), and the patient, iterative assistance of AI, I was able to automate some otherwise prohibitive manual grinds. We will explore how to build custom pipelines for automated quality checks, rapid codec comparisons, or metadata extraction; crunching through years of data in mere days - and how to avoid some pitfalls (aka "blind trust in AI will lead to blank disks").
While AI-assisted DIY is no replacement for human-designed software, it offers a welcome third way for the budget-deprived archivist with specialized needs. When the choice is between letting a sprinkle of AI help you or leaving your collection to languish in obscurity, the results of these experiments prove that a little bit of code goes a long way. (Plus, your boss might start calling you a "coder.")
Speaker summary
Martin Weiss
Martin Weiss has always been fascinated by technology and visual storytelling. Starting as a freelance photographer in high school, he studied communication in Berlin and film production in London, working as a cinematographer with 16mm and early digital formats. In Norway, he and his wife ran their own production company, and from 2011 he helped establish digital preservation workflows at the National Library of Norway. A research project on film colors brought him to Zurich, where he was part of a team that built a multispectral scanner and co-founded the spin-off Scan2Screen. With 30+ years across production and archiving, he uses pragmatic automation to overcome resource constraints and rescue legacy media. His primary role is as an Audiovisual Specialist at the University Library of Zurich, ensuring AV collections remain preserved and accessible, alongside ongoing work with Scan2Screen, the National Library of Norway, and independent projects.
