Neural Network Synthesizer
Artificial Neural Network
Additive Sound Synthesis
Sound Perception
ANN
Machine Learning
Perceptual Parameters
Perceptually-informed Sound Synthesis
Abstract
This work introduces a connectionist model that maps perceptual controllers to synthesis parameters to allow for an intuitive and powerful musical control of audio synthesis. After training on an instrument sound database, the system is able to generate control parameters for additive synthesis from pitch and loudness information only.
References
Le Groux, Sylvain. 2002. “A Neural Network Principal Component Synthesizer for Expressive Control of Musical Sounds.” DEA Research Report, Berkeley, CA USA: University of California at Berkeley/Paris VI University. http://www.atiam.ircam.fr/Archives/Stages0102/SylvainLegroux.pdf.