Presentation	2013-12-20 Application of HMM-Based Speech Synthesis Techniques to a Singing Style Synthesis Job Plugin Makoto TACHIBANA, Keijiro SAINO, Yuji HISAMINATO,
PDF Download Page	PDF download Page Link
Abstract(in Japanese)	(See Japanese page)
Abstract(in English)	Recent HMM-based speech synthesis systems have the capability to control speaker/style characteristics by statistically modeling prosodic features of speech. In this paper, we aim to introduce such flexibility into VOCALOID^, a singing voice synthesizer based on concatenative synthesis. An HMM is used to model the pitch trajectory, which is an important feature for expressing singing style. In order to synthesize appropriate results for pitches which do not have training data, we propose using pitch bend change as a feature for training the HMM. We also propose a segment-level unit modeling and multiple-stage clustering technique for the expression of long note sequences. We show the proposed technique implemented as a VOCALOID^3 Job Plugin.
Keyword(in Japanese)	(See Japanese page)
Keyword(in English)	HMM-based speech synthesis / singing voice synthesis / singing style / pitch bend / context clustering
Paper #	SP2013-94
Date of Issue

Conference Information
Committee	SP
Conference Date	2013/12/12(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)	Application of HMM-Based Speech Synthesis Techniques to a Singing Style Synthesis Job Plugin
Sub Title (in English)
Keyword(1)	HMM-based speech synthesis
Keyword(2)	singing voice synthesis
Keyword(3)	singing style
Keyword(4)	pitch bend
Keyword(5)	context clustering
1st Author's Name	Makoto TACHIBANA
1st Author's Affiliation	Speech Technology Group, Development Department 1, Research & Development Division, Yamaha Corporation()
2nd Author's Name	Keijiro SAINO
2nd Author's Affiliation	Speech Technology Group, Development Department 1, Research & Development Division, Yamaha Corporation
3rd Author's Name	Yuji HISAMINATO
3rd Author's Affiliation	Speech Technology Group, Development Department 1, Research & Development Division, Yamaha Corporation
Date	2013-12-20
Paper #	SP2013-94
Volume (vol)	vol.113
Number (no)	366
Page	pp.pp.-
#Pages	6
Date of Issue