2 datasets found

Adversarially trained end-to-end korean singing voice synthesis system.

This paper presents a high quality singing synthesizer that is able to model a voice with limited available recordings. Based on the sequence-to-sequence singing model, we...