WaveNetによる言語情報を含まない
感情音声合成方式における話者性の検討

Study of speaker individuality in emotional speech synthesis using
WaveNet without linguistic information

Authors: 松本 剣斗, 原 直, 阿部 匡伸 (岡山大学)
Kento Matsumoto, Sunao Hara, Masanobu Abe (Okayama University)

In this page, there are generated speeches by our proposed method.
Please pay attention to a volume setting of your computer.

Generated Normal Generated Angry Generated Happy
Speaker F


Speaker U


Speaker T


If you have any comment and question, please send email to k_matsu★a.cs.okayama-u.ac.jp (★ to @)