Human-Machine Musical Composition in Real-Time

Objective: Explore musical composition strategies based on artificial intelligence technics for producing musical melodies generated by an autonomous agent that accompanies a harmonic piece given by a human musician in real-time (Musical Improvisation).

Prototype:

A system that considers hardware (computer and MIDI controller) and software (artificial intelligence system and audio synthesis engine) was implemented to acquire and use musical knowledge from several musicians to feed a knowledge base in which artificial intelligence strategies are applied in order to generate melodies together with a harmonic base from a human composer such way an improvisation rises from both agents. The knowledge base considers musical notes, musical durations, and specific emotions that a composer wants to transmit to his or her audience. .

Human-Machine Musical Piece

Results

Several musical pieces were produced by a musician who used the system as accompany and defined specific emotions and intensities for each of them. These pieces were delivered to people who were asked to define the emotions that they felt when they listened the audios, as well as the intensity of those emotions in order to verify that the emotional intention from the composer is congruent with the emotional perception from the audience. There was found congruency between the composer and the audience, also other additional emotions were reported.

Publications:
2015: Adquisición del Conocimiento en el Proceso de Composición Musical en Base a Técnicas de Inteligencia Artificial
Efraín Astudillo, Pedro Lucas, Enrique Peláez

2015: Human-Machine Musical Composition in Real-Time Based on Emotions Through a Fuzzy Logic Approach (Artículo de Conferencia)
Pedro Lucas, Efraín Astudillo, Enrique Peláez

2016: Human-Machine Musical Composition in Real-Time Based on Emotions Through a Fuzzy Logic Approach (Capítulo de Libro)
Pedro Lucas, Efraín Astudillo, Enrique Peláez