This overarching effort of our lab aims to examine the nature and development of tonality and tonal harmony as well as counterpoint and musical forms across a broad timespan comprising a wide variety of historical styles (between ca. 1500 and today). At present, one major drawback affecting digital musicology is the paucity of large corpora annotated with features relevant in both music theory and cognition (to some extent similar, for instance, to the Penn Treebank in linguistics). To this end, we are in the process of compiling a large corpus of music with expert annotations of tonal harmony, counterpoint, and musical form, using newly developed annotation standards. The symbolic corpora originating from this project can be used to train, evaluate, and improve computational models. The annotation project is complemented by the analysis of Jazz and Pop recordings from the Montreux Jazz Archive.
See those annotated scores which have already been published on GitHub.
Here you find documentation and guidelines regarding the annotation standard used in our corpora.
This short presentation demonstrates how we extract and evaluate annotation labels from our MuseScore files.