Speechalator: two-way speech-to-speech translation on a consumer PDA

Alex Waibel $^{\rm 1\&4}$, Ahmed Badran$^{\rm 1}$, Alan W Black $^{\rm 1\&2}$, Robert Frederking$^{\rm 1}$, Donna Gates$^{\rm 1}$,
Alon Lavie$^{\rm 1}$, Lori Levin$^{\rm 1}$, Kevin Lenzo$^{\rm 2}$, Laura Mayfield Tomokiyo,$^{\rm 2}$
Jürgen Reichert$^{\rm 4}$, Tanja Schultz$^{\rm 1}$, Dorcas Wallace$^{\rm 1}$, Monika Woszczyna$^{\rm 3}$, Jing Zhang$^{\rm 4}$
$^1$ Language Technologies Institute, Carnegie Mellon University, Pittsburgh, PA
$^2$ Cepstral, LLC, $^3$ Multimodal Technologies Inc, $^4$ Mobile Technologies Inc.
speechalator@speechinfo.org

Abstract:

This paper describes a working two-way speech-to-speech translation system that runs in near real-time on a consumer handheld computer. It can translate from English to Arabic and Arabic to English in the domain of medical interviews.

We describe the general architecture and frameworks within which we developed each of the components: HMM-based recognition, interlingua translation (both rule and statistically based), and unit selection synthesis.