See presentationsIn classical Western music, when students learn how to write nice music, they are taught to respect a large set of rules: no parallel fifths, opposite movements between the extremal voices, cadence at the end of the piece, etc. Translated into a computer-science vocabulary, writing nice music is stated as a constraint problem, where the notes are the variables, the tempered scales are the domains, and the rules are the constraints (and, happily, this problem has lots of solutions). In practice, this problem, called automatic harmonisation, has been investigated since the very beginning of CP, back in the 1980s. Since then, CP has been applied to a great variety of musical problems, and used to produce score drafts, to organise sounds, to model musical properties, etc. In this tutorial, we will present an introduction to musical constraints, starting with automatic harmonisation, and then exploring examples in contemporary music and in sound processing. We will try and focus on the works where CP has been used in an unconventional way.
