Footprint reduction of concatenative text-to-speech synthesizers using polynomial temporal decompositionTamar ShohamDavid Malahet al.2010ISCCSP 2010