Prosodic phrase segmentation by pitch pattern clustering

Author

Shimodaira, Hiroshi ; Nakai, Mitsuru

Author_Institution

JAIST, Nomi, Japan

Volume

ii

fYear

1994

fDate

19-22 Apr 1994

Abstract

This paper proposes a novel method for detecting the optimal sequence of prosodic phrases from continuous speech based on data-driven approach. The pitch pattern of input speech is divided into prosodic segments which minimized the overall distortion with pitch pattern templates of accent phrases by using the One Pass search algorithm. The pitch pattern templates are designed by clustering a large number of training samples of accent phrases. On the ATR continuous speech database uttered by 10 speakers, the rate of correct segmentation was 91.7% maximum for the same sex data of training and testing, 88.6% for the opposite sex

Keywords

linguistics; search problems; speech processing; speech recognition; ATR continuous speech database; accent phrases; continuous speech recognition; correct segmentation rate; data-driven approach; input speech; one pass search algorithm; optimal sequence detection; pitch pattern clustering; pitch pattern templates; prosodic phrase segmentation; testing; training samples; Clustering algorithms; Databases; Gratings; Humans; Natural languages; Pattern clustering; Speech processing; Speech recognition; Stress; Testing;

fLanguage

English

Publisher

ieee

Conference_Titel

Acoustics, Speech, and Signal Processing, 1994. ICASSP-94., 1994 IEEE International Conference on

Conference_Location

Adelaide, SA

ISSN

1520-6149

Print_ISBN

0-7803-1775-0

Type

conf

DOI

10.1109/ICASSP.1994.389688

Filename

389688