P3-06: KDC: an open corpus for computational research of dastgāhi music
Nikzat, Babak*, Caro Repetto, Rafael
Subjects (starting with primary): Domain knowledge -> computational ethnomusicology ; Domain knowledge -> computational music theory and musicology ; Evaluation, datasets, and reproducibility -> novel datasets and use cases ; Musical features and properties -> melody and motives
Presented In-person, in Bengaluru: 4-minute short-format presentation
Iranian dastgāhi music is considered as the classical repertory of contemporary Iran. In the 19th century, the melodic modes that developed during its long history were grouped in categories, each of them known as dastgāh. The dastgāhi system presents unique features, that have been object of musicological study since its inception. However, computational methods for its research are still scarce, due in good part to the lack of open, well curated corpora. The aim of the KUG Dastgāhi Corpus (KDC) is to contribute to the development of computational corpus driven research for this tradition. KDC is created following the FAIR principles, and in close collaboration with performers and scholars, who contribute to it with annotations and qualitative evaluations. Besides presenting the first version of KDC, in this paper we explore the possibilities that Iranian dastgāhi music offers to computational research. In order to test the performance of state-of-the-art technologies applied to this music tradition, we present preliminary results for several analytical tasks, and discuss thei opportunities and limitations learnt in the process.