Composición e Interpretación Concurrente de Melodías Musicales en Tiempo Real Sintetizadas en Base a Emociones Humanas utilizando Técnicas de Inteligencia Artificial

Abstract

This work explores the application of Artificial Intelligence regarding music melodies and the techniques that supports the generation of them. A model that represents the musical knowledge is explained based on a revision for the state of the art and the criteria from experimental music’s practitioners; also, a description for a system that supports this model, which processes the input from a musical keyboard for composing and performing real-time concurrent melodies. This system is guided by the emotional intention from the performer such way it generates emotions on the audience compared to what a human artist would do. The architecture of this purposed model and the system prototype were subject to an aesthetic test, Turing test, emotional perception test for listeners, and computational performance test. These tests are analyzed in order to evaluate the effectivity for the selected approach, in which two components were implemented: Markov Chains, because of its use for sequence of elements, and Fuzzy Logic, due to its use in uncertainty situations, for this case, human emotions.