Local and International Achievements in Speech and Language Processing.

Tanya Gilbakken and Arild Noven
GRI-Nordisk Språkteknologi AS
Rykkesvegen 94, 5700 Voss, Norge
arne.gilbakken@gri.no, arild.noven@gri.no


The demonstration consists of two parts:

1. Mr. Arild Noven of GRI - Nordisk Språkteknologi AS will demonstrate a combined system of speech recognition, translation and speech synthesis.

As input connected words of spoken Norwegian are recorded and recognized. This speech recognizer has been developed by Mr. Noven as a part of his work at the Section of Linguistic Studies, University of Bergen.

The words are then concatenated into an alfanumeric string, which is given as an input sentence to a Machine Translations System, which translates the sentence into English. This Machine Translation System has been developed by Prof.Helge Dyvik, Section of Linguistic Studies, University of Bergen.

The output, which is an English sentence, is given as input to a Text-to-Speech Synthesizer (TTS), which converts the alfanumeric string to audible speech. The TTS has been developed by the Belgian company Lernout & Houspie Speech Products (for more information see http://www.lhs.com/speechtech/teledevtools/tts.asp).

This is the first opportunity offered in Norway to see and hear a combined system of speech recognition, translation and speech synthesis. Written presentations of each of the parts will be available to conference participants.

2. Mrs. Tanya Gilbakken of Nordisk Språkteknologi AS will demonstrate Voice Xpress Plus, which is a commercial dictation system from Lernout & Houspie Speech Products, Belgium. (http://www.lhs.com/dictation/consumer/vxpressplus/) Voice Xpress Plus, available for USD 100, accepts American English, spoken continuously at normal speaking rate as input, and converts the spoken input to a written text. This text can be post-edited by spoken editing commands.

This is the first opportunity offered in Norway to see Voice Xpress Plus.