Studying melodic embodiment: computational approaches

Many people think of melodies as having contours. The association between musical melodies, and the visual representation of a contour—an arch, a rainbow, a zigzag line, a circle, and so on—appears to be united in our minds, and represents an essential quality of melodic identity. Ehrenfels, one of the founders of Gestalt theory, uses melodies as a starting point to illustrate how Gestalt perception might function at all, making the understanding of melodic contour to be a quintessential gestalt phenomenon. Lessons in songwriting often teach students how to diversify their thinking by creating contrasts with contours. In this talk, I investigate why we think of melodies as contours, and how this differs from symbolic or score-based representations of melodies. One of the key methods i use in my work is to analyze how people move to melodies, to try and understand how melodic contour is represented. I will present my work with experiments on the analysis of Hindustani classical music, modelling melodic responses with motion-capture in various musical melodies, and analyzing how facial expressions of singers can alter interval perception amongst listeners.