"Deep Voice 3: 2000-Speaker Neural Text-to-Speech."

Wei Ping et al. (2017)
a service of Schloss Dagstuhl - Leibniz Center for Informatics