A Statistical Approach for Modeling the Expressiveness of Symbolic Musical Text

The application of artificial intelligence in the musical field is bringing about a rapid and radical transformation, with extraordinary potential to change the way music is created, analyzed and interpreted. This article intends to make an innovative contribution in the definition of a computational model capable of modeling the musical expressiveness of a score considered at its symbolic level (musical notation). To achieve this goal, the study used functional harmony to segment the musical score and information theory for the definition of musical dynamics: the organization of the executive intensity of the written notes respecting the stylistic and functional aspects of the musical composition. The results are encouraging because on the one hand they are very close to the executive interpretation of humans and on the other hand they show that the proposed model is “idiom-independent”, therefore capable of functioning regardless of compositional styles, and can be useful in a variety of music applications such as adaptive automatic music performance and score analysis.