next up previous
Next: Bibliography Up: Speechalator: two-way speech-to-speech translation Previous: System


The current demonstration is designed for the medical interview domain, with the doctor speaking English and the patient speaking Arabic. At this point in the project no formal evaluation has taken place. However, informally, in office-like acoustic environments, accuracy within domain is well over 80%.

Speechalator Arabic Input
Arabic input Screen
Speechalator snapshot

Alan W Black 2003-06-12