Inside the Archive #50: World Digital Preservation Day 2025 part two

A couple of weeks ago Joanna White (Knowledge & Collections Developer) and I, had the pleasure of attending the ninth No Time To Wait (NTTW9), an annual three-day conference highlighting the importance of digital preservation and open-source software. The conference is an amazing opportunity to meet with those who have had a major impact on our workflows in the BFI National Archive, as well as hearing about their recent work. This year, the amazing minds of the NTTW community congregated together in the beautiful and historic building of the National Library of Ireland in Dublin. We were even given a tour of the reading rooms.

For my work in the Digital Media Specialists team, the Audio QC for Re-Formatted Materials workshop presented by the Center for Puerto Rican Studies (CENTRO) proved very interesting. This was particularly relevant for acquisitions that require many hours of quality control (QC) work. This was very fitting as the Liz Truss Lettuce livestream was about to be acquired into the archive. They spoke specifically about using the sampling technique for when a start-finish or 100% QC inspection is not possible. For example, if they had 100 x 60-minute audio tapes, that would be 100 hours of potential QC work. They used a Python script and FFMPEG (a free, open-source tool for handling audio and video) to create random samples from the digitised work, which gave them 20 x 6-minute samples which would cut their QC down to only two hours. I will be doing something similar for the Lettuce video, which is around 100 hours of streaming footage, and will be QCing ‘samples’ of the file but will not be using any automation to create them.

Thursday was the first full day of talks. We learnt about Paradata (data about the process by which the archive materials were collected) and Provenance Metadata (information that records the origin, history and lineage of data) from Lode Scheers of meemoo; types of checksums (cryptographic and non-cryptographic) and their run speeds from Reto Kromer; the development of AV Spex, a currently macOS-only application written in Python which is designed to perform fixity checks, create access files for digital AV media and so much more.

The love for physical media in these conferences is also very present, and you’re unable to mention NTTW without the mention of magnetic materials. Libby Hopfauf, Executive Director of Moving Image Preservation of Puget Sound and Audiovisual Archivist at Seattle Municipal Archives, and Brianna Toth, Video Preservation Specialist, discussed the urgent need to digitise and preserve magnetic media before 2027–2032 to prevent degralescence. Their advice: start digitising your VHS tapes and cassettes now.

On the final full day of talks on Friday, Justine Provino presented her case study with Leontien Talbloom on Agrippa (A Book of the Dead) (1992). The work of art is an auto-destructive book, made to degrade with each page turn until it is no longer consumable. As a book conservator, Provino’s main question was: How do you preserve a book that self-destructs?

There are multiple copies of the book in six private collections and seven public institutions, including eight diskettes. The diskettes contain a poem and are programmed to encrypt themselves after a single use. In 1992, there was an event held in New York for a preview and reading of the poem on one of the diskettes. A group of people secretly recorded the event and released the recordings. As a result, the contents of the diskette has been preserved by the internet. The hackers preserved the poem online, and libraries share custodianship of the book with the internet.

As I am working on the contemporary collecting strand of Our Screen Heritage, a project supported by the Screen Heritage Fund, awarding National Lottery funding, this was incredibly exciting and very fitting for the work we are doing to collect 400 online web videos. You can see an example of the Agrippa files here: AGRIPPA (A Book of the Dead): The Poem Running in Emulation

Dublin has an incredibly rich cultural history, and we made sure to experience as much of it as possible between the digital preservation talks. We toured Christ Church Cathedral, marvelled at its architecture and the mummified cat and rat on display in its crypt, and enjoyed traditional Irish music sessions in authentic pubs.

Of course, no visit to Dublin would be complete without sampling a pint or two of Guinness – even if Joanna surprised a few locals with her Guinness shandies!

The NTTW conference continues to be one of the most inspiring events for all things digital preservation, and next year will mark 10 years since it was founded. For that milestone birthday, it will be held at the National Film Archive – Audiovisual Institute in Warsaw, Poland. I am very much looking forward to what next year’s conference brings and to celebrating NTTW’s 10th anniversary with everyone.

– Iris Mathieson, Digital Media Specialist (Our Screen Heritage)

As we celebrate World Preservation Day, I’m thankful for communities like NTTW. Their approach to collective knowledge development encourages archivists to share successes and failures openly in a supportive environment.

This year’s theme, “Why Preserve?”, is a question I rarely grapple with as Knowledge Learning and Collections Developer in the Data and Digital Preservation department at BFI National Archive. I’m usually caught up in the how questions. One great example of the value of why we should preserve is NTTW’s own collection of conference videos stretching back to their first event in 2016. Watching them today offers insight into the evolution of open standards like FFV1 (a lossless video codec used for video archiving and preservation) and tools like RAWcooked. These videos also reveal the No Time To Wait conference effect – a gradual upskilling of archivists supported and educated by this community.

It has been six years since I was first invited to join the panel How we FFV1, How we Matroska at the NTTW3, held at BFI Southbank. As an early career Digital Archivist, I’d started my own small investigations into transcoding uncompressed video files to the FFV1 codec, using open-source project FFMPEG. I was bursting with questions about how I was supposed to interpret these early experiments and I was lucky enough to have found the community with all the answers. Each year since, I’ve returned to learn as much as I can, while also trying to share my own knowledge and workflows as my skills have grown, thanks to the No Time To Wait conference effect.

For NTTW9, I hosted a workshop on extracting metadata with Python and gave a talk on recent improvements to our film-scanning preservation workflow. The workshop began with a short presentation on how our department adds technical metadata to digital media records in the Collection Information Database. This showed why automating metadata extraction with Python is useful. I shared examples of metadata and Python code from my GitHub repository and wrote a short instructional blog to support learning. When the venue Wi-Fi failed five minutes into the session, attendees couldn’t take part, so the blog became a valuable guide they could revisit later.

The next day, I presented our recent updates to film preservation workflows. We use RAWcooked to compress DPX and TIFF film scans into FFV1 Matroska, a tool we’ve relied on since 2019. Over the years, I expanded the original Linux shell scripts developed by Stephen McConnachie, Head of Data and Digital Preservation at the BFI. This gradually created a complex knot of code and folders. At the same time, improvements in the BFI National Archive’s scanning technology, such as the move 4K film scans, pushed this ageing code to its limits.

My presentation reflected upon these issues and explored my plans for improvement. In April 2025, we launched a new workflow built exclusively from Python code now runs inside an orchestrator called Dagster. This new workflow allows for wonderful transparency. We now have clear reporting of successes and failures, storage-savings summaries (169 TB saved since April), and tools that allow colleagues to interact directly with the workflows effecting change to the encoded files they have been working on.

From the many excellent presentations and workshops across the three days of NTTW9, my favourite was led by Jérôme Martinez, Media Area developer of RAWcooked. He discussed recent FFMPEG developments to dramatically improve coding and decoding times by development of ffv1-vulkan, a compute-based codec for GPU assisted encoding. His initial tests have found 3x speed improvements over current CPU coding and decoding. We look forward to testing this at the BFI National Archive as soon as we can!

The National Library of Ireland (NLI) were warm and welcoming hosts. They organised thoughtful tours including the Irish Film Institute, the NLI reading rooms. We visited the Bank of Ireland Cultural and Heritage Centre, where we saw the beautiful Seamus Heaney: Listen Now Again exhibition, as well as the NLI’s own Life and Works of William Butler Yeats, which features an installation of Sinéad O’Connor reading No Second Troy.

In truth, I don’t always find it comfortable standing up and talking at a conference, but communities like NTTW are hugely supportive, so I very much encourage archiving colleagues to try sharing knowledge where possible and help others learn from our experience. Presenting and sharing ideas helps to cultivate valuable friendships and conversations.

Thank you to the BFI National Archive for enabling attendance at this and other conferences, supporting open-source workflow development, and encouraging knowledge sharing.

If you would like to catch up with NTTW9 or any presentations from the last nine years, visit Media Area’s YouTube website. https://www.youtube.com/@MediaAreaNet/playlists

– Joanna White, Knowledge & Collections Developer

The Inside the Archive blog is supported by the BFI Screen Heritage Fund, awarding National Lottery funding.