Presentation 2012-11-08
Modeling of Local Variance of Spectral Features and Its Application to Parameter Generation in HMM-based Speech Synthesis
Takashi NOSE, Vataya CHUNWIJITRA, Takao KOBAYASHI,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) In this paper, we describe a technique for modeling local variance (LV) of speech features and propose a novel parameter generation algorithm using the LV model for HMM-based speech synthesis. In the proposed technique, We define the LV as a feature that represents the local variation around each frame of the spectral features and model them using context-dependent phone HMMs. To appropriately model the dynamic characteristics of LVs, we take into account the dynamic features of LVs as well as the static one. In the parameter generation process, a spectral parameter sequence is estimated so as to maximize a target function where conventional HMMs and LV models are combined. By using the LV models, the proposed technique can impose a more precise variance restriction in the parameter generation than the conventional technique where the global variance (GV) model is used. Through objective and subjective evaluations, we examine the effectiveness of the proposed technique.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) HMM-based speech synthesis / parameter generation / global variance (GV) / local variance (LV)
Paper # SP2012-79
Date of Issue

Conference Information
Committee SP
Conference Date 2012/11/1(1days)
Place (in Japanese) (See Japanese page)
Place (in English)
Topics (in Japanese) (See Japanese page)
Topics (in English)
Chair
Vice Chair
Secretary
Assistant

Paper Information
Registration To Speech (SP)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Modeling of Local Variance of Spectral Features and Its Application to Parameter Generation in HMM-based Speech Synthesis
Sub Title (in English)
Keyword(1) HMM-based speech synthesis
Keyword(2) parameter generation
Keyword(3) global variance (GV)
Keyword(4) local variance (LV)
1st Author's Name Takashi NOSE
1st Author's Affiliation Interdisciplinary Graduate School of Science and Engineering, Tokyo Institute of Technology()
2nd Author's Name Vataya CHUNWIJITRA
2nd Author's Affiliation Interdisciplinary Graduate School of Science and Engineering, Tokyo Institute of Technology
3rd Author's Name Takao KOBAYASHI
3rd Author's Affiliation Interdisciplinary Graduate School of Science and Engineering, Tokyo Institute of Technology
Date 2012-11-08
Paper # SP2012-79
Volume (vol) vol.112
Number (no) 281
Page pp.pp.-
#Pages 6
Date of Issue