FULL COVARIANCE STATE DURATION MODELING FOR HMM-BASED SPEECH SYNTHESIS: Nagoya Institute of Technology Repository System

	Nagoya Institute of Technology Repository System
	HOME Japanese library university Feedback

	detail-search

list

	authors
	type
	date

	Journal Title

	Top Downloads

community-list

	ALL LIST
	Omohi College [288]
	Shikumi College [293]
	Tsukuri College [280]
	Nagare College [285]
	Department of Materials Science and Technology [115]
	Department of Engineering physics, Electronics and Mechanics [25]
	Department of Computer Science and Engineering [45]
	Department of Architecture, Civil Engineering and Industrial Manageent Engineering [32]
	Department of Environmental Technology and Urban Planning [31]
	Department of Techno-Business Administration [0]
	Department of Systems Engineering [37]
	Department of Electrical and Computer Engineering [54]
	Ｄｅｐａｒｔｍｅｎｔ of Architectural and Civil Engineering [28]
	Department of Frontier Materials [12]
	Department of Scientific and Engineering Simulation [6]
	Others [0]
	Bulletin of Nagoya Institute of Technology [592]
	Annual Report of the Ceramics Research Laboratory [115]
	Technical Report at Research Center for Nano-Device and System [64]
	Journal of Engineering Ethics [34]
	New Directions [93]

Title	:	FULL COVARIANCE STATE DURATION MODELING FOR HMM-BASED SPEECH SYNTHESIS
Authors	:	Lu, Heng Wu, Yi Jian Dai, Li Rong Wang, Ren hua Tokuda, Keiichi
Authors alternative	:	徳田, 恵一
Issue Date	:	24-Apr-2009
Abstract	:	This paper proposes a state duration modeling method using full covariance matrix for HMM-based speech synthesis. In this method, a full covariance matrix instead of the conventional diagonal covariance matrix is adopted in the multi-dimensional Gaussian distribution to model the state duration of each context-dependent phoneme. At synthesis stage, the state durations are predicted using the clustered context-dependent distributions with full covariance matrices. Experimental results show that the synthesized speech using full-covariance state duration models is more natural than the conventional method when we change the speaking rate of synthesized speech.
Type Local	:	会議発表論文
ISBN	:	9781424423545
DOI	:	10.1109/ICASSP.2009.4960513
Publisher	:	Institute of Electrical and Electronics Engineers
URI	:	http://repo.lib.nitech.ac.jp/handle/123456789/2205
Citation	:	ICASSP 2009. IEEE International Conference on Acoustics, Speech and Signal Processing, 2009. p.4033 -4036
Appears in Collections	:	Academic Paper - Nagare College

Files in This Item:

File	Description	Size	Format
ICASSP2009_lu_heng.pdf		202Kb	Adobe PDF	View/Open