Footprint reduction of Concatenative Text-To-Speech synthesizers using polynomial temporal decomposition | IEEE Conference Publication | IEEE Xplore