Deterministic annealing EM algorithm for developing TTS system in Gujarati

Author

Shah, N.J. ; Patil, Hemant A. ; Madhavi, Maulik C. ; Sailor, Hardik B. ; Patel, Tanvina B.

Author_Institution

Dhirubhai Ambani Inst. of Inf. & Commun. Technol. (DA-IICT), Gandhinagar, India

fYear

2014

fDate

12-14 Sept. 2014

Firstpage

526

Lastpage

530

Abstract

The generalized statistical framework of Hidden Markov Model (HMM) has been successfully applied from the field of speech recognition to speech synthesis. In this work, we have applied HMM-based Speech Synthesis System (HTS) method to Gujarati language. Adaption and evaluation of HTS for Gujarati language has been done here. Evaluation of HTS system built using Gujarati data is done in terms of naturalness and speech intelligibility. Apart from this, a conventional EM algorithm-based HTS and recently proposed Deterministic Annealing EM algorithm-based HTS has been applied to Gujarati (a low resourced language) and it´s relative comparison has been done. It has been found that HTS in Gujarati has very high intelligibility. It was verified from AB-test that 70.5 % times DAEM-based HTS has preferred over EM-based HTS developed for Gujarati language.

Keywords

expectation-maximisation algorithm; hidden Markov models; natural language processing; speech intelligibility; speech synthesis; AB-test; DAEM-based HTS; Gujarati data; Gujarati language; HMM-based speech synthesis system; HTS method; TTS system; deterministic annealing EM algorithm; expectation maximization; generalized statistical framework; hidden Markov model; low resourced language; naturalness; speech intelligibility; speech recognition; text-to-speech; Annealing; Approximation algorithms; Databases; Hidden Markov models; High-temperature superconductors; Speech; Speech synthesis; Deterministic Annealing Expectation-Maximization (DAEM); Expectation-Maximization; HMM;

fLanguage

English

Publisher

ieee

Conference_Titel

Chinese Spoken Language Processing (ISCSLP), 2014 9th International Symposium on

Conference_Location

Singapore

Type

conf

DOI

10.1109/ISCSLP.2014.6936666

Filename

6936666