Codex is an Inria exploratory research project on the processing of written musical content. It is dealing with the creation organization, retrieval, and analysis of digital music scores and score collections.

Our approach focuses on original tree-structured representations of music scores. Our models aim at reflecting, in a natural way, the hierarchical organization of melodic and harmonic content in Common Western music notation. Besides score models, we are studying language theoretical tools for handling large sets of scores and notation styles, similarity metrics based on combinations of string and tree edit distances, and search and retrieval techniques for music score collections.

Applications include Automated Music Transcription, Crowd-sourced correction of Optical Music Recognition, Digital Humanities (Computational Musicology), and the management of digital music score databases for Cultural heritage preservation and study.

Our multidisciplinary activities take place in the Vertigo team on Complex Data at Cedric lab, CNAM, Paris, in collaboration with computer scientists of Sakai Lab at Nagoya University, musicologists of IReMus (UMR 8223 CNRS-SU), specialists of education sciences at Paris Cité University and librarians e.g. at the Bibliothèque Nationale de France (BnF).