Participants 2022
AIBeatz
AI Song Contest 2022 / participantsTEAM / AIBeatz
SONG / AIBeatz Urban Music Reel
TEAM MEMBERS / Pablo Saen de Jubera, Jorge García Abad, Javier puyuelo, Carlos Hernández Oliván, Adrián Ligüerre José Ramón Beltrán Blázquez, Daniel Murcia, Carlos Oyarzábal, Carlos de Miguel Arribas
About the TEAM
Aibeatz team has a wide skillset in machine learning engineering, Audio signal processing and music production. Thanks to our close collaboration with the Telecomunication engineering department from the University of Zaragoza, we were able to build our online synthesiser and our entire automatic composition technology. Furthermore, we had the priviledge to count with more than 20 students from the Master in Music production and sound design at Xpresate Advanced music school, what enabled us to build a proprietary urban music dataset with houndreds of thousads of self produced music files and over 1500 self-made instruments.
About the SONG
We decided to create a reel with a variaty of beats generated with our AI beat making platform. This reeal is formed by compositions in the most heard urban music genres: Trap, Drill, Boombap, Latin and Afrotrap music.
About the HUMAN-AI PROCESS
Our system is based on a mix of two neural networks that create short music segments and algorithmic music. One is a RNN-based network and the other one is a Transformer-based model. We use both models and we choose the segments of each models that we liked the most.
After the short music segments of a few bars are generated, we selected one of them and we imposed a high-level structure to generate the full song automatically in one click. Our musical team supervised the generated MIDI files and then the song was produced and the effects were applied in the AIBeatz’s beat making platform.