Duration Study for the Bell Laboratories Mandarin Text-to-Speech System
We present in this chapter the methodology and results of a duration study designed for the Mandarin Chinese text-to-speech system of Bell Laboratories. A greedy algorithm is used to select text from on-line corpora to maximize the coverage of factors that are important to the study of duration. The duration model and some interesting results are discussed.
KeywordsGreedy Algorithm Bell Laboratory Closure Duration Vowel Duration Syllable Type
Unable to display preview. Download preview PDF.
- [AHK87]J. Allen, S. Hunnicut, and D. H. Klatt. From text to speech: The MITalk system. Cambridge University Press, Cambridge, UK, 1987.Google Scholar
- [Ber93]R. Berkovits. Utterance-final lengthening and the duration of final-stop closures. J. Phonetics21(4):479–489, 1993.Google Scholar
- [Fen85]L. Feng Beijinghua yuliu zhong sheng yun diao de shichang (Duration of consonants, vowels, and tones in Beijing Mandarin speech). In Beijinghua Yuyin Shiyanlu (Acoustics Experiments in Beijing Mandarin),Beijing University Press, Beijing, 131–195, 1985.Google Scholar
- [FGO93]R. M. French, A. Greenwood, and J. P. Olive. Speech Segmentation Criteria. Technical report, AT&T Bell Laboratories, 1993.Google Scholar
- [Kla75]D. H. Klatt. Vowel lengthening is syntactically determined in a connected discourse. J. Phonetics3:129–140, 1975.Google Scholar
- [LR73]D. Lindblom and K. Rapp. Some temporal regularities of spoken Swedish. Publication of the Institute of Linguistics, University of Stockholm,21:1–59, 1973.Google Scholar
- [Noo72]S. G. Nooteboom. Production and Perception of Vowel Duration. University of Utrecht, Utrecht, 1972.Google Scholar
- [OGC93]J. P. Olive, A. Greenwood, and J. Coleman. Acoustics of American English Speech: A Dynamic Approach. Springer-Verlag, New York, 1993.Google Scholar
- [Ren85]H. Ren. Linguistically conditioned duration rules in a timing model for Chinese. In UCLA Working Papers in Phonetics 62,I. Maddieson, ed. UCLA, Los Angeles, 1985.Google Scholar
- [SSGC94]R. W. Sproat, C. Shih, W. Gale, and N. Chang. A stochastic finite-state wordsegmentation algorithm for Chinese. In Proceedings of the 32nd Annual Meeting of the Association for Computational Linguistics,New Mexico State University, 66–73, 1994.Google Scholar
- [van92b]J. P. H. van Santen. Diagnostic perceptual experiments for text-to-speech system evaluation. In Proceedings of ICSLP,Barff, Alberta, Canada, 555–558, 1992.Google Scholar