AI-Assisted Curation and Transcription of Norwegian Folk Music
Summary
This collaboration, established in 2017 between the National Library of Norway and RITMO, University of Oslo, develops computational tools to curate and analyse the Library’s Norwegian folk music collection. Initial work focused on automatically segmenting long tape recordings into individual tunes, with expert verification through a dedicated interface. The main focus is now automated transcription: detecting notes in audio recordings and representing the music as a score. The hybrid AI approach combines deep learning with rule-based models informed by musicology and music cognition, addressing the distinctive intonation and playing techniques of Norwegian fiddle traditions. External collaborators: Hans-Hinrich Thedens (National Library of Norway) and Joan Gatti (University of South-Eastern Norway).
Work Package
Primary WP: WP6
Duration
2017–ongoing
People
- Project leader: Olivier Lartillot
- Participants: Lars Monstad