3A: mAchine learning Algorithm Applied to emotions in melodies

• Cláudio Gomes Federal University of Amapá
• Josue da Silva Federal University of Amapá
• Marco Leal Federal University of Amapá
• Thiago Nascimento Federal University of Amapá

DOI: https://doi.org/10.5753/sbcm.2019.10450

Resumo

At every moment, innumerable emotions can indicate and provide questions about daily attitudes. These emotions can interfere or stimulate different goals. Whether in school, home or social life, the environment increases the itinerant part of the process of attitudes. The musician is also passive of these emotions and incorporates them into his compositions for various reasons. Thus, the musical composition has innumerable sources, for example, academic formation, experiences, influences and perceptions of the musical scene. In this way, this work develops the mAchine learning Algorithm Applied to emotions in melodies (3A). The 3A recognizes the musician’s melodies in real time to generate accompaniment melody. As input, The 3A used MIDI data from a synthesizer to generate accompanying MIDI output or sound file by the programming language Chuck. Initially in this work, it is using the Gregorian modes for each intention of composition. In case, the musician changes the mode or tone, the 3A has an adaptation to continuing the musical sequence. Currently, The 3A uses artificial neural networks to predict and adapt melodies. It started from mathematical series for the formation of melodies that present interesting results for both mathematicians and musicians.

Palavras-chave: Artificial Intelligence, A-Life and Evolutionary Music Systems, Music, Emotion and Communication, Real-time Interactive Systems

Referências

Jose D Fernández and Francisco Vico. Ai methods in algorithmic composition: A comprehensive survey. Journal ofArtificial Intelligence Research, 48:513–582, 2013.

International MIDI Association et al. Midi musical instrument digital interface specification 1.0. Los Angeles, 1983.

John Strawn and Curtis Roads. Foundations of computermusic. MIT Press, 1985.

Manuel Alfonseca, Manuel Cebrián Ramos, and Alfonso Ortega. A fitness function for computer-generated music using genetic algorithms. WSEAS Transactions on InformationScience and Applications, 2006.

Teuvo Kohonen. Self-organization and associative memory,volume 8. Springer Science & Business Media, 2012.

Ge Wang, Perry R Cook, and Spencer Salazar. Chuck: Astrongly timed computer music language. Computer MusicJournal, 39(4):10–29, 2015.

Raspberry Pi Foundation. Raspberry pi. http://www.raspberrypi.org, may 2019. Accessed: 2019-05-05.