Analysis of extracted pitch contours across speakers for intonation modelling in TTS synthesis

Author

Gerazov, Branislav ; Ivanovski, Zoran

Author_Institution

Fac. of Electr. Eng. & Inf. Technol., Ss. Cyril and Methodius Univ., Skopje, Macedonia

fYear

2012

fDate

2-4 May 2012

Firstpage

1

Lastpage

4

Abstract

The paper presents the results of an analysis of extracted pitch contours of spoken Macedonian across 7 native speakers and 3 discourse contexts. The 125 analyzed intonation phrases (IPs) were taken from a speech corpus recorded for this purpose. Pitch contours were extracted for each of the phrases, and then normalized to the speaker´s mode. This allowed group analysis of the contours. In total 8 groups were created based on speaker sex and 4 different discourse functions. Each group was then statistically analyzed and average normalized pitch as well as upper and lower bound vectors were calculated. The algorithms designed for this purpose have been described in detail. The calculated vectors were used as the basis for building linearsegment models used for automatic intonation generation in a text-to-speech (TTS) synthesis system.

Keywords

feature extraction; speech synthesis; Macedonian; TTS synthesis system; extracted pitch contours; intonation modelling; intonation phrases; lower bound vectors; speaker mode; speech corpus; text-to-speech; upper bound vectors; Algorithm design and analysis; Analytical models; Context; Digital filters; Materials; Speech; Vectors; Pitch contours; average pitch; discourse function; intonation pattern; pitch range; text-to-speech;

fLanguage

English

Publisher

ieee

Conference_Titel

Communications Control and Signal Processing (ISCCSP), 2012 5th International Symposium on

Conference_Location

Rome

Print_ISBN

978-1-4673-0274-6

Type

conf

DOI

10.1109/ISCCSP.2012.6217785

Filename

6217785