What is it about?
Allophones are various phonetic realizations of a language. It is difficult to classify them using acoustical representation, only. Therefore, we added data from the face motion capture system, where lip movements are clearly visible. We discuss results of results of allophone recognition using such a bi-modal approach.
Featured Image
Why is it important?
Speech recognition on the allophonic level is important for automatic translation of speech to IPA (International Phonetic Alphabet)
Perspectives
We still work at automatic speech translation to IPA. The audio-video corpus was prepared accessible at the web address: www.modality-corpus.org
Andrzej Czyzewski
Gdańsk University of Technology
Read the Original
This page is a summary of: Bimodal classification of English allophones employing acoustic speech signal and facial motion capture, The Journal of the Acoustical Society of America, September 2018, Acoustical Society of America (ASA),
DOI: 10.1121/1.5067951.
You can read the full text:
Contributors
The following have contributed to this page
