Development of Hindi speech synthesizer for Metro Rail Information system

Author

Balyan, Archana ; Agrawal, S.S. ; Dev, Amita

Author_Institution

Dept. of Electron. & Commun. Eng., Maharaja Surajmal Inst. of Eng., New Delhi, India

fYear

2011

fDate

16-18 Sept. 2011

Firstpage

2166

Lastpage

2170

Abstract

Building a natural conversational interface is a major challenge faced by the researchers. Speech synthesis is a technology used for automatic information retrieval. In this paper, we described the development of unit selection for Hindi language. The aim of this study is to develop speech synthesizer considering phonemes as the basic unit in Hindi language, as applicable to Metro Rail Passenger Information System (MRPIS).The phoneme has been chosen as the basic unit as larger domain synthesizer cannot be created using syllable as the basic unit. We describe the build process and address the issue of speech segmentation using Hidden Markov Model (HMM) based techniques. We report the comparison of automatically segmented labels (speech units) using base line model of HMM with the manually segmented labels in the context of Hindi Speech.

Keywords

hidden Markov models; natural languages; railways; speech synthesis; Hindi language; Hindi speech synthesizer development; automatic information retrieval; hidden Markov model based techniques; metro rail passenger information system; natural conversational interface; phonemes; speech segmentation; Buildings; Databases; Hidden Markov models; Mel frequency cepstral coefficient; Speech; Speech synthesis; Synthesizers; Automatic Phonetic Segmentation; Concatenation; Hidden Markov Model; Mel frequency Cepstral Co-efficient; Unit selection synthesis;

fLanguage

English

Publisher

ieee

Conference_Titel

Electrical and Control Engineering (ICECE), 2011 International Conference on

Conference_Location

Yichang

Print_ISBN

978-1-4244-8162-0

Type

conf

DOI

10.1109/ICECENG.2011.6057268

Filename

6057268