Development of augmented speech production techniques based on combination of statistical approaches and speech production modeling approaches

2016 Fiscal Year Final Research Report

• Project/Area Number: 26280060
• Research Category: Grant-in-Aid for Scientific Research (B)
• Allocation Type: Partial Multi-year Fund
• Section: 一般
• Research Field: Perceptual information processing
• Research Institution: Nagoya University (2015-2016); Nara Institute of Science and Technology (2014)
• Principal Investigator: Toda Tomoki 名古屋大学, 情報基盤センター, 教授 (90403328)
• Co-Investigator(Kenkyū-buntansha): 亀岡 弘和 日本電信電話株式会社NTTコミュニケーション科学基礎研究所, メディア情報研究部 メディア認識研究グループ, 主任研究員(特別研究員) (20466402) ; 中村 哲 奈良先端科学技術大学院大学, 情報科学研究科, 教授 (30263429) ; 猿渡 洋 東京大学, 情報理工学(系)研究科, 教授 (30324974) ; サクリアニ サクティ 奈良先端科学技術大学院大学, 情報科学研究科, 助教 (00395005) ; Neubig Graham 奈良先端科学技術大学院大学, 情報科学研究科, 助教 (70633428) ; 川波 弘道 奈良先端科学技術大学院大学, 情報科学研究科, 助教 (80335489)
• Project Period (FY): 2014-04-01 – 2017-03-31
• Keywords: 音声変換 / 音声合成 / 信号処理 / 統計処理 / 機能拡張

Outline of Final Research Achievements

In this research, we developed fundamental techniques for augmented speech production and its applications to break down existing barriers caused by physical constraints in our speech production. High-quality speech conversion methods to be effectively used in our physical speech production mechanism were successfully developed by combining a statistical approach capable of generating high-quality converted speech and a speech production modeling approach capable of intuitively controlling converted speech by manipulating movements of speech organs. Moreover, we developed various applications of the augmented speech production techniques, such as a speaking aid technique towards restoration of lost voices, a foreign speech generation technique while keeping speaker identity, and a telecommunication technique using body-conducted speech.

Free Research Field

音メディア情報処理