Generative Audio Synthesis with a Parametric Model
arXiv:1911.08335
Abstract
Use a parametric representation of audio to train a generative model in the interest of obtaining more flexible control over the generated sound.
ISMIR 2019 Late Breaking/Demo