・名古屋工業大学学術機関リポジトリは、名古屋工業大学内で生産された学術情報を電子的に収集・保存・発信するシステムです。 ・論文の著作権は、著者または出版社が保持しています。著作権法で定める権利制限規定を超える利用については、著作権者に許諾を得てください。 ・著者版フラグに「author」と記載された論文は、著者原稿となります。実際の出版社版とは、レイアウト、字句校正レベルの異同がある場合もあります。 ・Nagoya Institute of Technology Repository Sytem is built to collect, archive and offer electronically the academic information produced by Nagoya Institute of Technology. ・The copyright and related rights of the article are held by authors or publishers. The copyright owners' consents must be required to use it over the curtailment of copyrights. ・Textversion "Author " means the article is author's version. Author version may have some difference in layouts and wordings form publisher version.
Analysis of Stream-Dependent Tying Structure for HMM-based Speech Synthesis
利用統計を見る
File / Name
License
本文_fulltext
c2008 IEEE. Personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution to servers or lists, or to reuse any copyrighted component of this work in other works must be obtained from the IEEE.
In conventional HMM-based speech synthesis framework,spectral features are modeled in one stream, andstream-dependent tree-based clustering was then appliedfor tying the model parameters. In this paper, we investigateseveral different stream-dependent tying structuresfor spectral features by splitting the feature vectorinto several streams. One splitting approach is to spliteach feature dimension into each stream. Another oneis to split the static and dynamic features into differentstreams. Although splitting spectral features into differentstreams would ignore the correlation of context dependencybetween them, the number of model parameterscan be optimized for each stream after stream-dependentclustering. From the experimental results, both splittingapproaches can improve the quality of synthesizedspeech. However, the quality of synthesized speech becameworse when we combined these two splitting approaches.
September 15-18, 2008Tokyo, Japan
Analysis of Stream-Dependent Tying Structure for HMM-based Speech Synthesis
本文_fulltext
(376.17KB)
