1 paper · 1 filter
Michael Ungersböck, Florian Grötschla, Luca A. Lanzendörfer +3
Generative models have made significant progress in synthesizing high-fidelity audio from short textual descriptions. However, editing existing audio using natural language has rem…