This paper reports on present work, in which a recurrent neural network is trained to estimate velum height during continuous speech. Parallel acoustic-articulatory data comprising more than 400 read TIMIT sentences is obtained using electromagnetic articulography (EMA). This data is processed and used as training data for a range of neural network sizes. The network demonstrating the highest accuracy is identified. This performance is then evaluated in detail by analysing the networks output for each phonetic segment contained in 50 hand-labelled utterances set aside for testing purposes.
Cite as: Richmond, K. (1999) Estimating velum height from acoustics during continuous speech. Proc. 6th European Conference on Speech Communication and Technology (Eurospeech 1999), 149-152, doi: 10.21437/Eurospeech.1999-41
@inproceedings{richmond99_eurospeech,
author={Korin Richmond},
title={{Estimating velum height from acoustics during continuous speech}},
year=1999,
booktitle={Proc. 6th European Conference on Speech Communication and Technology (Eurospeech 1999)},
pages={149--152},
doi={10.21437/Eurospeech.1999-41}
}