Changes in care of articulation (COA) affect both the spectral and durational characteristics of speech. This can have severe repercussions on both the success of speech recognition, and the quality of speech synthesis. Although auto-segmentation has proven useful for measuring the durational effects of COA, an automatic spectral measurement has proven more problematic. In this paper, we will explore the use of the acoustic log likelihoods generated by HMM autosegmentation as a measure of these changes in comparison with two phonetically motivated modeling systems based on vocalic F1/F2 values. When duration variation is controlled, the HMM output does not correlate with the human perception of vowel goodness, whereas, the phonetically motivated models do.
Cite as: Aylett, M.P. (2001) Modelling care of articulation with HMMs is dangerous. Proc. 7th European Conference on Speech Communication and Technology (Eurospeech 2001), 1491-1494, doi: 10.21437/Eurospeech.2001-369
@inproceedings{aylett01_eurospeech,
author={Matthew P. Aylett},
title={{Modelling care of articulation with HMMs is dangerous}},
year=2001,
booktitle={Proc. 7th European Conference on Speech Communication and Technology (Eurospeech 2001)},
pages={1491--1494},
doi={10.21437/Eurospeech.2001-369}
}