Development of speech analysis-synthesis system enabling flexible control of voice quality caused by peculiarity of vocal cords

2015 Fiscal Year Final Research Report

• Project/Area Number: 25870883
• Research Category: Grant-in-Aid for Young Scientists (B)
• Allocation Type: Multi-year Fund
• Research Field: Perceptual information processing; Entertainment and game informatics 1
• Research Institution: Meijo University
• Principal Investigator: BANNO Hideki 名城大学, 理工学部, 准教授 (20335003)
• Project Period (FY): 2013-04-01 – 2016-03-31
• Keywords: 高品質音声分析合成 / 声帯振動 / 歌唱音声 / モデル化 / 高速化 / GPGPU

Outline of Final Research Achievements

In order to convert pitch and quality of input speech and regenerate the converted speech, speech analysis-synthesis systems are widely used as a foundation technology. Although the state of the art of speech analysis-synthesis systems enables quite high-quality speech regeneration, it fails to regenerate high-quality speech for input speech which has peculiarity in vibration of vocal cords such as abnormal voice caused by laryngeal disease, shout singing voice, etc. In this research, we deal with such speech having peculiarity of vocal cords, and tackle it by improving a pitch-frequency extraction method and modeling the vibration of the vocal cords.

Free Research Field

音声信号処理