Annotating conversational speech for corpus-based dialogue speech synthesizer — A first step

Author

Mori, Hisamichi ; Hitomi, Tadaaki

Author_Institution

Grad. Sch. of Eng., Utsunomiya Univ., Utsunomiya, Japan

fYear

2012

fDate

9-12 Dec. 2012

Firstpage

135

Lastpage

140

Abstract

This paper describes an HMM-based speech synthesis that allows dimensional description of emotion as inputs. A spontaneous dialogue speech corpus that was designed for studying paralinguistic phenomena in expressive social interactions was used to train the models, utilizing its emotional state description as additional contextual factors. In the perceptual experiment, a very high correlation was observed (R ≃ 0.8) between given pleasantness/arousal values and averaged subjective evaluations, which means that the synthesized utterances could successfully convey specified paralinguistic information.

Keywords

audio databases; hidden Markov models; linguistics; speech synthesis; HMM-based speech synthesis; arousal value; conversational speech annotation; corpus-based dialogue speech synthesizer; emotion dimensional description; emotional state description; expressive social interaction; hidden Markov model; paralinguistic information; paralinguistic phenomenon; pleasantness value; spontaneous dialogue speech corpus; subjective evaluation; Context; Correlation; Databases; Hidden Markov models; Pragmatics; Speech; Speech synthesis; HMM-based speech synthesis; UU Database; spontaneous speech;

fLanguage

English

Publisher

ieee

Conference_Titel

Speech Database and Assessments (Oriental COCOSDA), 2012 International Conference on

Conference_Location

Macau

Print_ISBN

978-1-4673-2811-1

Electronic_ISBN

978-1-4673-2812-8

Type

conf

DOI

10.1109/ICSDA.2012.6422461

Filename

6422461