Simultaneous Modeling Of Spectrum, Pitch And Duration In HMM-Based Speech Synthesis

Simultaneous Modeling Of Spectrum, Pitch And Duration In HMM-Based Speech Synthesis (1999)

Cached

Download Links

[crow.ee.washington.edu]

by Takayoshi Yoshimura , Keiichi Tokuda , Takao Kobayashi , Takashi Masuko , Tadashi Kitamura

Citations:

64 - 21 self

Versions

Version History

Metadata Version 4

User correction supplied by mph

Datum	Value	Source
TITLE	Simultaneous Modeling Of Spectrum, Pitch And Duration In HMM-Based Speech Synthesis	user correction - Legacy Corrections
AUTHOR NAME	Takayoshi Yoshimura	user correction
AUTHOR AFFIL	ÝÝTokyo Institute of Technology	user correction
AUTHOR ADDR	Nagatsuta, Midori-ku, Yokohama, 226-8502 Japan	user correction
AUTHOR NAME	Keiichi Tokuda	user correction
AUTHOR AFFIL	Tokyo Institute of Technology	user correction
AUTHOR ADDR	Nagatsuta, Midori-ku, Yokohama, 226-8502 Japan	user correction
AUTHOR NAME	Takao Kobayashi	user correction
AUTHOR AFFIL	ÝÝTokyo Institute of Technology	user correction
AUTHOR ADDR	Nagatsuta, Midori-ku, Yokohama, 226-8502 Japan	user correction
AUTHOR NAME	Takashi Masuko	user correction
AUTHOR AFFIL	ÝÝTokyo Institute of Technology	user correction
AUTHOR ADDR	Nagatsuta, Midori-ku, Yokohama, 226-8502 Japan	user correction
AUTHOR NAME	Tadashi Kitamura	user correction
AUTHOR AFFIL	ÝÝTokyo Institute of Technology	user correction
AUTHOR ADDR	Nagatsuta, Midori-ku, Yokohama, 226-8502 Japan	user correction
ABSTRACT	In this paper, we describe an HMM-based speech synthesis system in which spectrum, pitch and state duration are modeled simultaneously in a unified framework of HMM. In the system, pitch and state duration are modeled by multi-space probability distribution HMMs and multi -dimensional Gaussian distributions, respectively. The distributions for spectral parameter, pitch parameter and the state duration are clustered independently by using a decision-tree based context clustering technique. Synthetic speech is generated by using an speech parameter generation algorithm from HMM and a mel-cepstrum based vocoding technique. Through informal listening tests, we have confirmed that the proposed system successfully synthesizes natural-sounding speech which resembles the speaker in the training database.	user correction
YEAR	1999	user correction - Legacy Corrections
CITATIONS	16 found	ParsCit 1.0