Statistical analysis of old Serbian music
Milan J. Merkle, Miomir M. Mijić · The Journal of the Acoustical Society of America · 1990
Being a link between mathematical scrutiny and a fuzziness of the real world, statistics is often used to help investigate relations between objects or phenomena that are intuitively similar (or dissimilar). In these problems, statistical methods set objective criteria of similarity, thus reducing the human proneness to errors. This work is an attempt to make a contribution to an objective analysis of music. It is not quite clear what are the main objective characteristics of a musical work, and much less is known about how to translate them to a mathematical writing such that when two songs sound similar, one gets two numbers that are close to each other. This is a problem of coding. In this work, the problem of coding is investigated, and the power of various statistical methods in comparison of songs is demonstrated. This is done using samples of an old Serbian chant from the 15th century. This kind of music is used because of its relative simplicity.