In 1997 Mexican poet and politician Octavio Paz moved to the Casa Alvarado, a house in Mexico City’s verdant district of Coyoacán. Paz, a Nobel-prize laureate, thus moved to a neighbourhood that included equally famous former residents such as Diego Rivera, Frida Kahlo and Leon Trotsky. However, Paz would not enjoy the magnificent residence for too long: he would die a few months later, in April 1998.

When Paz died his house passed on to a foundation, and it eventually became the site of the Mexican National Sound Archives (the Fonoteca Nacional), which hosted the annual conference of the International Association for Sound and Audiovisual Archives (IASA) in late September 2022.

More audio now available

Professor Perla Olivia Rodriguez Resendiz welcomed the crowd of professional sound archivists by highlighting the symbiotic relationship between the access needed by scholars, the preservation efforts carried out by professionals, and the eventual benefits to society: “[we] research to save, [and] save to disseminate”. To this end, there were several exciting projects that provided new access, from hometown ones to a couple led by the good folks at the American Archive of Public Broadcasting ([1] [2]), as well as an inspiring presentation of one “archive-centered” European broadcaster’s workflow.

Digitizing

The process of providing access to historical audio includes digitization, and institutions all over the globe are busy doing just that. At the conference, CUNY’s Dave Rice moderated a session where former NYPR Archives intern Veronica Smith and NEDCC’s Bryce Roe presented new ways to digitize 1990s Digital Audio Tapes (DATs) and analog discs; in another session, Karl Fleck described NEDCC’s protocol to document tape speed changes so that spring peeper frogs do not end up sounding like a horror movie soundtrack. Across the pond, Italy is busy adapting its audio file formats to that country’s new law on requirements for digital file formats; thankfully, the recommendations echo those in the US by the National Archives and the Library of Congress (see an international comparison here).

Progress in these fronts is not uniform. While the digitization process is by now generally well understood and well established, and ever more sophisticated guidelines continue to appear (including when using external vendors), sometimes elements such as basic vocabularies still need normalization. For example, IASA’s Organising Knowledge Committee may soon be working on a comprehensive nomenclature for audio and video carriers (the most complete list we are aware of is somewhat buried in Annex 3 of EBU’s Tech 3279 document); these vocabularies may someday be published using the CIDOC-CRM model or, more generally, as RDF triples.

Archivists continue to debate what to do with original carriers once digitization is complete, and we noticed presentations where tapes are supposedly destroyed after digitization. At the NYPR Archives we feel strongly about the long-term preservation of physical formats, while acknowledging it may be an impractical option for some; on this front, some excellent tools to help with that irreversible decision are now available.

Machine learning and AI

Not surprisingly, several presentations at the IASA conference included Machine learning (ML) topics. Everyone seemed to agree that (1) ML can be very helpful when describing and classifying vast amounts of material; (2) information professionals need to be better trained in ML; (3) it is still generally good to insert humans in the process; and (4) ML is here to stay (Georgina Sanabria Medina of the Universidad Nacional Autónoma de México summarized it well: “This is like the industrial revolution. No one makes their own shoes anymore”).  

A few talks explored the use of ML to index digital sound archives for educational use, to manage preservation, and to describe AV materials, while WGBH also mentioned its plans to investigate improvements to Kaldi and other potential new advancements in open-source speech-to-text tools. Some talks also included references to higher-level explorations of AI tools from organizations such as the Netherlands AV Institute or the Aeolian Network (which led us to discover this excellent introductory video from Bohyun Kim). While we are all excited by all these possibilities, everyone seems aware of some of the ethical issues associated with ML –issues that reflect general societal inequalities and that, presumably, should be solvable.

Sobering perspectives

While it is very exciting indeed to consider new technologies that aid us in managing AV archives, we should not forget that humans continue to be the source and ultimate destination of all our activity. On this front, we met colleagues concerned about supervisors who are showing excessive focus on monetizing, general disinterest in archives, or who implement constrictive policies; as well as colleagues concerned with such mundane troubles as keeping their server rooms cold enough in the face of an impending energy crisis in Europe. After a global pandemic, one felt a sense of appreciation for our jobs combined with a reinforced perspective of focusing on long-term sustainability in the face of unexpected challenges.

It’s all about the people

It had been three years since IASA attendees had met in person, and there was much joy in seeing our colleagues from WGBH, the Biblioteca de Catalunya, Drexel University, the Library of Congress, Disney Music, Indiana University, and many other places. It was equally exciting to get to know like-minded colleagues from Cornell University, Portugal’s National Sound Archive, the Basque Music Archive, the National Swiss Archives, the J.H. Kwabena Nketia Archives, and many others, as well as meeting our wonderful hosts from the Fonoteca in Mexico City.

Some of our colleagues have retired but continue to attend the annual conference or help with the organization’s administration. Others are contemplating retirement or job changes, and some are being honored by their colleagues. All in all, you would be hard-pressed to find a more dedicated group of professionals.

It was also great to attend sessions focused on people. Michael Marlatt presented a terrific workshop on “Strategies for a more inclusive archive” that went way beyond just including resources aimed at compliance (although those were also included [1] [2]) and emphasized that “universal design” and inclusive policies [3] [4] are good for everyone. We all still have lots to learn, but I feel that there are good teachers [5] [6] [7] [8] out there.

People power also manifested itself in other ways, from Torin Andersen’s quiet gathering of tapes during a pandemic to crowdsourced transcription repairs to efforts to translate metadata schemas to other languages. And the open source community was present in such projects as Preforma’s FFV1 software.

People can also disagree, of course. At the meeting for the Diversity, Equity, and Inclusion Committee there was a long debate over whether the committee’s upcoming statement on discrimination should include the word “race”. We learned from our European colleagues that academic papers and policies in Europe tend to avoid the term, since it is “a construct”. “There is only the human race”, a member emphasized.

On a personal note, the most thrilling moment for me was listening to a recording of a 1950s program written by my great-uncle, who had fled the Spanish Civil War and settled in Mexico City. Just hearing his name uttered by the station’s announcer almost brought tears to my eyes –a powerful reminder of the emotional impact that the audio materials we preserve can have.

There is a tree in Casa Alvarado believed to be almost 500 years old. It stands not far from the Fonoteca’s laboratories, where a dedicated group of professionals is hard at work every day to preserve the country’s audio legacy. Few of us know what will happen to these materials 100 years from now, let alone 500; but the IASA conference provided reassurance that we are contributing, singly and collectively, our grain of sand to the cause of audio preservation.