STiLL - Speech Technology in Language Learning

May 25-27, 1998
Marholmen, Sweden

Language-Specific Pronunciation Training Using the HearSay System

Jonathan Dalby (1), Diane Kewley-Port (2), Roy Sillings (1)

(1) Communication Disorders Technology, Inc.
(2) Indiana University Department of Speech and Hearing Sciences, USA

This paper describes a multimedia system for explicit second language pronunciation training using automatic speech recognition technology. The system has been provisionally named HearSay. The term 'explicit' is used in the description because the system includes training curricula designed for specific native/target language pairs. In the design of these curricula, phonological error analyses were conducted to discover the inventories of segmental errors typically made by speakers of a given native language when speaking a particular target language. The methods used in compiling these inventories, and for evaluating candidate automatic speech recognition technologies for the system will be described. Assumptions about effective speech intelligibility training that underlie the major design features of the system will be discussed.

Full Paper

Bibliographic reference.  Dalby, Jonathan / Kewley-Port, Diane / Sillings, Roy (1998): "Language-specific pronunciation training using the hearsay system", In STiLL-1998, 25-28.