JAIST Repository >
School of Information Science >
Grants-in-aid for Scientific Research Papers >
FY 2016 >

Please use this identifier to cite or link to this item: http://hdl.handle.net/10119/14328

Title: ヒト発話シミュレータによるStory Teller Systemの構築
Other Titles: Construction of story teller system by human voice production simulator
Authors: 赤木, 正人
Authors(alternative): Akagi, Masato
Keywords: 音声情報処理
Issue Date: 1-Jun-2017
Abstract: 表現豊かな音声合成システムの実現を目指して,本を読み聞かせるシステムの構築を行う。言うなれば,役で声質を演じわけるコンピュータを実現するプロジェクトである。HMM音声合成などの既存の方法では,学習対象とする大規模音声データベースに依存した音声が合成されるため,様々な役を演じる個性ある音声の合成には,発話スタイルそれぞれに大規模データベースが必要となる。これを打破するために,ヒトの音声生成機構を忠実に反映したヒト発話モデル(発話シミュレータ)を考案する。様々な声質を使い分けられるヒトの音声発話方法を解明し,モデルに組み込むことで,個性ある声の合成への適用を試みる。:Aiming to realize an expressive speech synthesis system, we build a story teller system. In other words, it is a project to realize a computer that plays various roles with their appropriate voice qualities. In existing methods such as HMM speech synthesis, since speech dependent on large-scale speech database to be learned is synthesized, large-scale databases are required to synthesize speech with various speech styles and individualities, in playing various roles. In order to overcome this, we devise a model for human speech production (voice production simulator) that faithfully reflects human speech production mechanism. Elucidating the human speech production method which can selectively use various voice qualities, we try to apply it to the synthesis of individual voice by implementing it into the model.
Description: 基盤研究(A)(一般)
Language: jpn
URI: http://hdl.handle.net/10119/14328
Appears in Collections:2016年度 (FY 2016)

Files in This Item:

File Description SizeFormat
25240026seika.pdf234KbAdobe PDFView/Open

All items in DSpace are protected by copyright, with all rights reserved.


Contact : Library Information Section, Japan Advanced Institute of Science and Technology