An experimental speech to graphics system
Andrew Golightly, Tony C. Smith · 2002
Ever improving speech technology continues to revolutionise the way we interact with computers. This paper describes a speech-driven graphics system that allows the user to construct and manipulate 3-dimensional (3D) graphical images using only their voice, averting the need to learn a graphics programming language or the point-and-click options of a conventional graphics software interface. The system combines an inexpensive Java-based speech-to-text package with open-source Java packages for constructive solid geometry and text-to-speech generation to create a completely hands-off graphics application. These components are integrated with context-free input/output grammars modeled from observations about the language used when a person unfamiliar with computer graphics software directs an experienced user in the creation of 3D images. The result is a natural, conversation-style interface that allows anyone to make effective use of 3D-graphics packages regardless of their technical expertise.