"Fast, low-artifact speech synthesis considering global variance."

Matt Shannon, William Byrne (2013)
a service of  Schloss Dagstuhl - Leibniz Center for Informatics