A Deep Generative Model of Speech Complex Spectrograms

This paper proposes an approach to the joint modeling of the short-time Fourier transform magnitude and phase spectrograms with a deep generative model.

BibTex: