Automatic evaluation of synthesized speech

Author

Martincic-Ipsic, Sanda ; Pobar, Miran ; Ipsic, I.

Author_Institution

Dept. of Inf., Univ. of Rijeka, Rijeka, Croatia

fYear

2009

fDate

22-25 June 2009

Firstpage

305

Lastpage

310

Abstract

This paper presents the development of a Croatian text-to-speech (TTS) system capable of synthesizing speech from arbitrary text. For speech generation pitch synchronous overlap and add (PSOLA) and hidden Markov models (HMM) based methods were used. The generated Croatian speech was evaluated subjectively by the mean opinion score (MOS) of the 40 evaluators and objectively by the automatic speech recognition (ASR) system. In the paper we propose an evaluation approach which combines objective and subjective evaluation results.

Keywords

hidden Markov models; speech synthesis; Croatian text-to-speech system; HMM; MOS; PSOLA; TTS system; automatic speech recognition system; hidden Markov model; mean opinion score; pitch synchronous overlap-and-add; speech synthesis; Application software; Automatic speech recognition; Dictionaries; Frequency estimation; Hidden Markov models; Informatics; Natural languages; Speech analysis; Speech synthesis; Text analysis; ASR; Hidden Markov Models; MOS evaluation; PSOLA; diphone speech synthesis; speech synthesis;

fLanguage

English

Publisher

ieee

Conference_Titel

Information Technology Interfaces, 2009. ITI '09. Proceedings of the ITI 2009 31st International Conference on

Conference_Location

Dubrovnik

ISSN

1330-1012

Print_ISBN

978-953-7138-15-8

Type

conf

DOI

10.1109/ITI.2009.5196099

Filename

5196099