Bridging the Gap Between Musicological Knowledge and Performance Practice with Audio MIR

Abstract

In an oral tradition, such as is the case with the art music of India, musicology research can reap rich benefits by exploiting available computational methods on the audio recordings of great artists. Apart from providing a better understanding of the common practices employed in performance, deeper insights into the structural and theoretical aspects of the genre can be obtained using techniques from music information retrieval (MIR). In this paper, we take specific examples of musicological questions that can be answered by the analysis of the physical sound signals corresponding to concert recordings. We present the signal processing techniques and the resulting computational representations that allow us to compare two performances objectively on chosen attributes such as the melody. Available musicological knowledge is used to fine-tune the parameters of the representation. The computational melody representation, together with an appropriate distance measure, provides a model that can be fruitfully applied to study large datasets for new insights into performance practice.