Design and Evaluation of a Text-to-Speech System for Azerbaijani Turkish Language and Database Generation

Author

Damadi, Mir Saeed ; Azami, Bahram Zahir

Author_Institution

Electr. & Comput. Eng. Dept., Univ. of Kurdistan, Sanandaj, Iran

fYear

2009

fDate

20-25 July 2009

Firstpage

111

Lastpage

116

Abstract

This paper addresses the design of a text-to-speech system for Azerbaijani Turkish language and also explains the database preparation for the system. Various speech synthesis methods are explained and some important linguistic features of Azerbaijani Turkish language are investigated. According to this features among the variety of synthesis methods, corpus based concatenative speech synthesis is selected for this language. A database is generated for the system, subsequent to the selected synthesis method. The basic units of this database are diphones. For synthesized speech generation according to input text, appropriate units are selected from database and concatenated together. As far as we know, the output speech that is produced with this system is the first Azerbaijani Turkish synthetic speech ever created.

Keywords

audio databases; natural language processing; speech processing; speech synthesis; Azerbaijani Turkish language; Azerbaijani Turkish synthetic speech; concatenative speech synthesis; database generation; database preparation; diphones; linguistic features; text-to-speech system; Computer architecture; Concatenated codes; Data engineering; Design engineering; Information systems; Natural languages; Spatial databases; Speech synthesis; Synthesizers; Telecommunication computing; concatenative; corpus; database; diphone; speech synthesis;

fLanguage

English

Publisher

ieee

Conference_Titel

Digital Telecommunications, 2009. ICDT '09. Fourth International Conference on

Conference_Location

Colmar

Print_ISBN

978-0-7695-3695-8

Type

conf

DOI

10.1109/ICDT.2009.28

Filename

5205226