Generative Audio Synthesis with a Parametric Model

TL;DR

This paper introduces a parametric approach to generative audio synthesis, aiming to enhance control over sound generation by training models on parametric audio representations.

Contribution

It proposes a novel parametric model for audio synthesis that improves flexibility and control compared to traditional methods.

Findings

Achieves more precise sound control

Demonstrates effective generative capabilities

Outperforms baseline models in quality

Abstract

Use a parametric representation of audio to train a generative model in the interest of obtaining more flexible control over the generated sound.