JAIST Repository >
b. 情報科学研究科・情報科学系 >
b11. 会議発表論文・発表資料等 >
b11-1. 会議発表論文・発表資料 >

このアイテムの引用には次の識別子を使用してください: http://hdl.handle.net/10119/18196

タイトル: Study on method to control fundamental frequency contour related to a position on Valence-Activation space
著者: Hamada, Yasuhiro
Elbarougy, Reda
Xue, Yuawn
Akagi, Masato
キーワード: Speech-to-speech translation
emotion
fundamental frequency contour
Fujisaki model
valence and activation space
発行日: 2015-12
出版者: WESPAC
誌名: Proceedings, 12th Western Pacific Acoustics Conference 2015
開始ページ: 519
終了ページ: 522
DOI: 10.3850/978-981-09-7961-4_P12000176
抄録: Speech-to-speech translation (S2ST) system is important for human-machine interface. In our previous study, we have proposed a speech conversion system from neutral to emotional ones by considering emotion space spanned by the Valence and Activation axes (V-A space). To build relationships between V-A space and acoustic features, Adapted Network Fuzzy Inference System (ANFIS) was applied. Neutral speech was converted to an emotional speech to control the values of acoustic features that were related to a position of V-A space. However, the proposed conversion system has some problems to control the acoustic features of the neutral speech. In this paper we propose a new method to control fundamental frequency (F_0) contour. In order to control the F_0 contour, Fujisaki model was used. F_0 contour was modified by controlling the parameters of Fujisaki model. The results showed the F_0 was able to control using Fujisaki model.
Rights: Copyright (C) 2016 WESPAC 2015. This material is posted here with permission of WESPAC (Western Pacific Acoustics Conference). Yasuhiro Hamada, Reda Elbarougy, Yuawn Xue, Masato Akagi, Proceedings of 12th Western Pacific Acoustics Conference 2015
URI: http://hdl.handle.net/10119/18196
資料タイプ: publisher
出現コレクション:b11-1. 会議発表論文・発表資料 (Conference Papers)

このアイテムのファイル:

ファイル 記述 サイズ形式
WESPAC2015.pdf120KbAdobe PDF見る/開く

当システムに保管されているアイテムはすべて著作権により保護されています。

 


お問い合わせ先 : 北陸先端科学技術大学院大学 研究推進課図書館情報係