Presentation 2011-01-28
Automatic Prediction of Accent Phrase Boundaries using Linguistic and F0 Information
Asami YAMAMOTO, Kook CHO, Yoichi YAMASHITA,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) Speech synthesis techniques require speech data which is annotated with labels concerning prosodic information. This paper describes a method of automatic labeling of prosodic information focusing on accent phrase boundaries. A probabilistic model using both linguistic and prosodic information predicts the boundaries under the condition that the contents of speech and phoneme labels are given. We use CRF and multidimensional normal distribution for the linguistic probability model and the F0 probability model, respectively. We try to improve accuracy of the accent phrase boundary prediction using the cumulative mora count from the preceeding accent phrase boundary. The cumulative mora count is calculated by making hypotheses of the accent phrase boundaries. Evaluation experiments show that the cumulative mora count improves accuracy of accent phrase boundary prediction for read speech of ATR503 sentences.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) accent phrase boundary / CRF / F0 pattern / cumulative mora count
Paper # SP2010-109
Date of Issue

Conference Information
Committee SP
Conference Date 2011/1/20(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) Automatic Prediction of Accent Phrase Boundaries using Linguistic and F0 Information
Sub Title (in English)
Keyword(1) accent phrase boundary
Keyword(2) CRF
Keyword(3) F0 pattern
Keyword(4) cumulative mora count
1st Author's Name Asami YAMAMOTO
1st Author's Affiliation Graduate School of Science and Engineering, Ritsumeikan University()
2nd Author's Name Kook CHO
2nd Author's Affiliation College of Information Science and Engineering, Ritsumeikan University
3rd Author's Name Yoichi YAMASHITA
3rd Author's Affiliation College of Information Science and Engineering, Ritsumeikan University
Date 2011-01-28
Paper # SP2010-109
Volume (vol) vol.110
Number (no) 401
Page pp.pp.-
#Pages 6
Date of Issue