Repertorium Project Unearths 4,000 “Lost” Chants

A medieval chant manuscript in a digitized negative for scanning

In one of the better uses of technology (and EU funding), the Repertorium project is an effort to utilize “a set of AI-based tools to automate the digitalisation and cataloguing of historical mediaeval and classical music manuscripts.” In their own description:

REPERTORIUM is an ambitious European project with the primary goal of preserving and disseminating Europe’s historical musical heritage. Five Spanish organisations are participating in this initiative: the Higher Polytechnic School of Linares at the University of Jaén, the Complutense Institute of Musical Sciences, Complutense University, the University of Alicante, and the Hispanic Association for the Study of Gregorian Chant. This project, involving a total of 13 institutions and companies from eight countries, is creating a system based on open-source Artificial Intelligence (AI) techniques to facilitate the digitisation and cataloguing of historical musical archives.

They are laying claim to some remarkable fruits of this endeavor already:

Thanks to the system developed by the REPERTORIUM project, around 4,000 pieces of Gregorian chant have been recovered, pieces that had not been transcribed, studied, catalogued, or sung for over 1,000 years. These pieces have been digitised and catalogued, making them accessible to researchers and the general public through digital libraries.

There is not let a place online where the rediscovered chants may be viewed or downloaded, but their availability is part of the overall project, as can be found on page 24 of their vision document:

The database of mediaeval works being developed will find its home in DIAMM, the Digital Image Archive of Medieval Music. Additionally, it will be indexed within the MMMO, also known as the Medieval Music Manuscripts Online database. DIAMM will inherit the metadata added via MMMO via API. These platforms will serve as the designated spaces for hosting and cataloguing the valuable collection of mediaeval musical works.

       The mediaeval data will be publicly accessible, allowing researchers and enthusiasts to explore and utilise it. The metadata associated with this data will include essential information linked to the Cantus ID. The Cantus ID will be the unique identifier for data access, enabling users to locate and retrieve the desired information effectively. Lastly, the extensive collection of around 2,000 hours of audio recordings, which captures the entire liturgical cycle of traditional Latin chants from the Abbaye Sainte-Madeleine (Task 2.3), will be allocated individual Cantus IDs as they are added to the Neumz app. This identification system guarantees that each recording is distinctly identified within the database.

On January 25, 2025, a concert was performed at the Cathedral of Salamanca utilizing at least some of these rediscovered chants:

To read more about the concert, visit this link.