EUROSPEECH '93
Third European Conference on Speech Communication and Technology

Berlin, Germany
September 22-25, 1993

      

The Benefits of Tiered Segmentation for the Recognition of Phonetic Properties

Mark Huckvale

Dept. Phonetics & Linguistics, University College London, London, UK

The modelling and recognition of the phonetic composition of speech using a number of independent levels or tiers is contrasted with conventional linear phone models both theoretically and experimentally.

Keywords: Phonetic recognition, speech recognition, non-linear phonology.

Full Paper

Bibliographic reference.  Huckvale, Mark (1993): "The benefits of tiered segmentation for the recognition of phonetic properties", In EUROSPEECH'93, 1473-1476.