Design of Tibetan Continuous Speech Corpus Based on Triphone

Yong Hong Li; Peng Cuo Dawa

doi:10.4028/www.scientific.net/AMM.644-650.2245

Paper Titles

Study on the Determination of the Separated Layer Water Injection Based on the Grey Synthetic Degree
p.2227

Research of MPC Method on Preventing PIOs and Comparison with other Methods
p.2231

Design of Household Heat Metering Management System Based on Cloud Computing
p.2235

A New Delegation Provable Data Possession in Public Cloud Storage
p.2239

Design of Tibetan Continuous Speech Corpus Based on Triphone
p.2245

Study of the Logistics Transportation Vehicle Terminal Path Optimization and Algorithm Based on GIS
p.2249

Analysis on Scheduling Algorithms of Real-Time Hybrid Tasks
p.2253

Estimate of the Interfacial Length Scale in Boiling Heat Transfer Simulation Based on Mixture Model
p.2258

Adaptive Genetic Algorithm Verification Technology Driven by Functional Coverage
p.2264

HomeApplied Mechanics and MaterialsApplied Mechanics and Materials Vols. 644-650Design of Tibetan Continuous Speech Corpus Based...

Design of Tibetan Continuous Speech Corpus Based on Triphone

Abstract:

Large vocabulary continuous speech recognition system performance largely depends on the quality of speech corpus and how to select corpus is the key of corpus design. By taking Tibetan Amdo dialect in XiaHe as the research object, this paper builds continuous speech corpus based on triphone. At first, we collected text corpus with 1000 thousand Tibetan sentences and transformed them into IPA according to real pronunciation in XiaHe dialect, and then summarized the structure of triphone juncture, analyzed the combination types and frequency of triphone in corpus statistically with text-processing platform in detail. At last by comprehensively considering coverage rate and sparseness of triphone and class-triphone we designed the algorithm for extraction of corpus and realized automatic selection to corpus.

You might also be interested in these eBooks

View Preview

Info:

Periodical:

Applied Mechanics and Materials (Volumes 644-650)

Pages:

2245-2248

DOI:

https://doi.org/10.4028/www.scientific.net/AMM.644-650.2245

Citation:

Cite this paper

Online since:

September 2014

Authors:

Yong Hong Li*, Peng Cuo Dawa

Keywords:

Class-Triphone, Phone, Speech Corpus, Tibetan, Triphone

Export:

RIS, BibTeX

Price:

Permissions CCC:

Request Permissions

Permissions PLS:

Request Permissions

Сopyright:

Citation:

* - Corresponding Author

References

[1] R.H. Wang, J.F. Ni: The Design of Chinese Speech Corpus based on the juncture, Computer Applications and Software, Vol. 34 (1994) No. 1, pp.30-39.

Google Scholar

[2] J.F. Cao: Mandarin Environment Phonetic Change and the Structure of Diphone and Triphone, Applied linguistics, Vol. 45 (1996) No. 2, pp.58-63.

Google Scholar

[3] J.F. Cao: Mandarin Corpus Represents Set of Diphone and Triphone Structure System, Applied linguistics, Vol. 27 (1997) No. 1, pp.60-68.

Google Scholar

[4] Y.Q. ZU: The Text Design for Continuous Speech Database of Standard Chinese, Acta Acustica, Vol. 33 (1999) No. 3, pp.236-247.

Google Scholar

[5] H. Wu, B. Xu, T.Y. Huang: Automatic Corpus Selecting Algorithm Based on Triphone Models, Journal of software Vol. 47 (2000) No. 2, p.271.

Google Scholar

[6] Y.H. Li, J.P. Kong, H.Z. Yu: Rules for the Auto-Transformation of Tibetan Text to IPA, Journal of Tsinghua University (Science and Technology) Vol. 38 (2008) No. 4, pp.621-626.

Google Scholar