Computational Music Aesthetics: a survey and some thoughts

While computational aesthetic evaluation has been applied to images and visual output, it is not as widely employed for generative music systems. Computational aesthetic evaluation is not to be confounded with numerical evaluation of the system’s output; such a notion is in danger of offering a reduced and impoverished interpretation of the aesthetic experience, which is innately dialogical, between the creator or the user, the sociological context, and the creative process or product.

This paper reviews common computational aesthetic measures that have been used for musical applications, whilst arguing for a pragmatist perspective and a framework foregrounding the primacy of intentionality and agency in inducing aesthetic responses.