A speech interface to the PENGASP system

Abstract

The increased presence and accessibility of online speech recognition services has encouraged an investigation of the effectiveness of using such a service to allow users to speak a textual specification in controlled natural language instead of typing it. Google’s Web Speech API provides an accessible and portable speech recognition service that integrates well with web-based interfaces. Using Google’s Web Speech API, we present the design and implementation of a speech-based interface for the PENGASP system. We do this in order to examine the usefulness of speech-based input for controlled natural language processing and to explore potential synergies between speech-based and text-based input.