Lychnos (Feb 2023)

Audiovisual Media and Automatic Speech Recognition: A Methodological Reflection

  • Emil Stjernholm

Abstract

Read online

This methodological reflection focuses on the use of automatic speech recognition in my ongoing research project Televising Information: The Audiovisual Communication of Government Agencies, which has been conducted in collaboration with the National Library of Sweden’s KB-Lab. The article maps the pitfalls and affordances of ASR in the study of audiovisual media and digital history more broadly. Three key topics are emphasized: first, the collaboration between scholars and audiovisual archives and the potential of automatic speech recognition to transform the way scholars navigate and engage with audiovisual media history; second, what knowledge can be gained from employing automatic speech recognition in the exploration of film and television history; and lastly, thoughts on trajectories for future research. Even though the National Library of Sweden’s audiovisual collection is frequently used in humanities and social science scholarship, no previous research project has utilized automatic speech recognition algorithms to analyze material from their collection. As such, this reflection aims to contribute toward methodological development.

Keywords