Speech vocoding for laboratory phonology

详细信息    查看全文

• 作者：Milo&scaron ; Cerňak^a ; ^{milos.cernak@idiap.ch} ; ^{milos.cernak@ieee.org} ; &Scaron ; tefan Beňu&scaron ; ^b ; Alexandros Lazaridis^a
• 关键词：Phonological speech representation ; Parametric speech synthesis ; Laboratory phonology
• 刊名：Computer Speech & Language
• 出版年：2017
• 出版时间：March 2017
• 年：2017
• 卷：42
• 期：Complete
• 页码：100-121
• 全文大小：3665 K
• 卷排序：42

文摘

We propose a platform of a phonological speech vocoding for laboratory phonology. Extended Sound Pattern of English achieves slightly better results than others. We present phonological atoms – acoustic realisations of the phonological features. We propose a compositional model for generating arbitrary speech sounds. The parametric phonological TTS achieves 85% of intelligibility of HTS synthesis.