It is recognised that some phonemes are more useful for speaker recognition than others. For example, it is commonly accepted that unvoiced phonemes are less useful than voiced ones. This paper overviews published work investigating the speaker-discriminating properties of phonemes. Following this, a preliminary experiment is described, in which the performances of nasals and unvoiced fricatives are compared, using an on-line transputer-based speaker recognition system.
Cite as: Eatock, J.P., Mason, J.S.D. (1992) Phoneme performance in speaker recognition. Proc. 2nd International Conference on Spoken Language Processing (ICSLP 1992), 1411-1414, doi: 10.21437/ICSLP.1992-377
@inproceedings{eatock92_icslp,
author={J. P. Eatock and J. S. D. Mason},
title={{Phoneme performance in speaker recognition}},
year=1992,
booktitle={Proc. 2nd International Conference on Spoken Language Processing (ICSLP 1992)},
pages={1411--1414},
doi={10.21437/ICSLP.1992-377}
}