The Applied Intelligence Research Centre has a diverse team.
The goal of this project is the control of high level parameters for a sound synthesis model. Sound synthesis used to be a classic signal processing problem but these methods lead to samples with “robotic” effects. Therefore, during the last years, many approaches examined the use of machine learning in the field increasing the computational cost without the sufficient results. This project investigates the marrying of deep generative models with classic signal processing techniques for the synthesis of more natural sounds.