A strategy to solve data scarcity problems in corpus based intonation modelling

Author

Cardeñoso, Valentín ; Escudero, David

Author_Institution

Departamento de Informatica, Valladolid Univ., Spain

Volume

1

fYear

2004

fDate

17-21 May 2004

Abstract

Data scarcity in corpus-based intonation modelling for text-to-speech (TTS) applications is addressed. Multiple model dictionaries are proposed to predict patterns not found in the training corpus. A grouping strategy is proposed to improve models of classes without a high enough number of training samples. An experimental study of this strategy shows that better pitch profiles can be predicted in this way.

Keywords

dictionaries; linguistics; speech synthesis; statistical analysis; statistical distributions; Bezier functions; acoustic intonation parameters; class grouping; corpus based intonation modelling; data scarcity problems; intonation units; linguistic analysis; linguistic prosodic features; multiple dictionaries; pitch contours; reduced prediction errors; statistical distributions; stress groups; text to speech synthesis; Bibliographies; Classification algorithms; Contracts; Dictionaries; Knowledge based systems; Neural networks; Predictive models; Speech synthesis; Testing; Training data;

fLanguage

English

Publisher

ieee

Conference_Titel

Acoustics, Speech, and Signal Processing, 2004. Proceedings. (ICASSP '04). IEEE International Conference on

ISSN

1520-6149

Print_ISBN

0-7803-8484-9

Type

conf

DOI

10.1109/ICASSP.2004.1326073

Filename

1326073