Phonetic to acoustic mapping using recurrent neural networks
- 1 January 1991
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 753-756 vol.2
- https://doi.org/10.1109/icassp.1991.150450
Abstract
The application of artificial neural networks for phonetic-to-acoustic mapping is described. The specific task considered is that of mapping consonant-vowel-consonant (CVC) syllables to the corresponding formant values at different speech tempos. The performances of two different networks, the Elman recurrent network and a single hidden layer feedforward network, are compared. The results indicate that the recurrent network is able to generalize from the training set and produce valid formant contours for new CVC syllables that are not a part of the training set. It is shown that by choosing the proper input representation, the feedforward network is also capable of learning this mapping.Keywords
This publication has 4 references indexed in Scilit:
- Complete gradient optimization of a recurrent network applied to /b/,/d/,/g/ discriminationThe Journal of the Acoustical Society of America, 1990
- Generalization of back-propagation to recurrent neural networksPhysical Review Letters, 1987
- A methodology for modeling vowel formant contours in CVC contextThe Journal of the Acoustical Society of America, 1987
- Formant-Frequency Trajectories in Selected CVC-Syllable NucleiThe Journal of the Acoustical Society of America, 1970