Improvement of Naturalness for a HMM-based Korean TTS using the prosodic boundary information (운율경계정보를 이용한 HMM기반 한국어 TTS 자연성 향상 연구)

임기정; Jungchul Lee (이정철)

Improvement of Naturalness for a HMM-based Korean TTS using the prosodic boundary information

Journal of The Korea Society of Computer and Information
Abbr : JKSCI
2012, 17(9), pp.75-84
Publisher : The Korean Society Of Computer And Information
Research Area : Engineering > Computer Science

임기정 ¹, Jungchul Lee ¹

¹울산대학교

Accredited

ABSTRACT

HMM-based Text-to-Speech systems generally utilize context dependent tri-phone units from a large corpus speech DB to enhance the synthetic speech. To downsize a large corpus speech DB, acoustically similar tri-phone units are clustered based on the decision tree using context dependent information. Context dependent information includes phoneme sequence as well as prosodic information because the naturalness of synthetic speech highly depends on the prosody such as pause, intonation pattern, and segmental duration. However, if the prosodic information was complicated, many context dependent phonemes would have no examples in the training data, and clustering would provide a smoothed feature which will generate unnatural synthetic speech. In this paper, instead of complicate prosodic information we propose a simple three prosodic boundary types and decision tree questions that use rising tone, falling tone, and monotonic tone to improve naturalness. Experimental results show that our proposed method can improve naturalness of a HMM-based Korean TTS and get high MOS in the perception test.

KEYWORDS

HTS, HMM, tri-phone, decision tree-based clustering

Citation status

* References for papers published after 2023 are currently being built.

[book] K. Tokuda / 2004 / An HMM based approach to multilingual speech synthesis, In Text to speech synthesis: New paradigms and advances / Prentice Hall : 135~153

[confproc] A. W. Black / 2007 / Statistical parametric speech synthesis / Proc. ICASSP 2007 4 : 1229~1232

[journal] 이현창 / 2008 / A study of Implementing An Embedded System for Conversion from Text to Speech / 한국컴퓨터정보학회논문지 / 한국컴퓨터정보학회 13(3) : 77~84

[web] S. Young / The Hidden Markov Model Toolkit (HTK) / http://htk.eng.cam.ac.uk/

[web] K. Tokuda / The HMM based speech synthesis system (HTS) / http://hts.sp.nitech.ac.jp/

[web] A.W. Black / The festival speech synthesis system / http://www.festvox.org/festival/

[journal] 한민수 / 200611 / HMM-based Korean speech synthesis system for hand-held devices / IEEE TRANSACTIONS ON CONSUMER ELECTRONICS / IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC 52(4) : 1384~1390

[journal] 이정철 / 2010 / UA Tree-based Reduction of Speech DB in a Large Corpus-based Korean TTS / 한국컴퓨터정보학회논문지 / 한국컴퓨터정보학회 15(7) : 91~98

[confproc] S. Imai / 1983 / Cepstral analysis synthesison the melfrequency scale / Proc. ICASSP 1 : 93~96

[confproc] K. Tokuda / 1995 / An Algorithm for Speech Parameter Generation from Continuous Mixture HMMs with Dynamic Features / Proc. of EUROSPEECH 1 : 757~760

[confproc] J. Latorre / 2011 / Continuous F0 in the source-excitation generation for HMM-based TTS : Do we need voiced/unvoiced classification? / Proc. ICASSP : 4724~4727

[confproc] Q. Zhang / 2010 / Improved modeling for FO generation and V /U decision in HMM-based TTS / Proc. ICASSP : 4606~4609

[journal] K. Tokuda / 2002 / Multi-space probability distribution HMM(Invited paper) / IEICE Trans. Inf. & Syst. E85-D(3) : 455~464

[confproc] S. J. Young / 1994 / Tree-based state tying for high accuracy acousticmodeling / Proc. ARPA Human LanguageTechnology Workshop : 307~312

[journal] K. Shinoda / 2000 / MDL-based contextdependent subword modeling for speech recognition / J. Acoust. Soc. Jpn. (E) 21(2) : 79~86

[confproc] K. Shinoda / 1997 / Acoustic modeling based on the MDL criterion for speech recognition / Proc. Eurospeech 1 : 99~102

KJCKorea
Journal Central

Journal of The Korea Society of Computer and Information 2023 KCI Impact Factor : 0.65