Methods for the automatic learning of musical features applied to the composition of tonal music

Resumen

The main goal of the present Doctoral Thesis is to design, implement and test a system that resembles a human approach to the composition of music. Instead of tackling the complete composition process as a whole, we decided to take small steps to the creation of comprehensive algorithms. A novel probabilistic model for the characterisation of music learned from music samples is designed. This model makes use of automatically extracted music parameters namely rhythm, intonation (pitch contours) and harmony, to characterise music. Then a novel autonomous music composition that generates new melodies using the model developed is implemented. The methods proposed in the current Doctoral Thesis take into consideration different aspects related to the traditional way in which music is composed by humans such as harmony evolution and structure repetitions and apply them together with the probabilistic reutilisation of rhythm patterns and pitch contours learned beforehand to compose music pieces.