Text composition by voice: design issues and implementations

Publisher Website
Google Scholar
Abstract
It is now technically possible to use speech as an interfacing medium between a disabled person and an assistive aid. As yet, the advantages and disadvantages of this new interfacing technique have not been seriously assessed, In this paper, we describe the design and implementation of two different text-composition aids featuring speech input. One system is based on direct selection of letters and certain common words; the other, on encoded selection of letter-sequences, or graphemes. By considering two rather different systems, the implications of choice of message element and means of indication are made clear. In a relatively informal experiment using one of the authors as the speaker, the direct-selection system allowed text composition (with all errors corrected) at a rate of 39 characters per minute. The corresponding figure for the encoded-selection system was 28 characters per minute. A great deal more evaluative work is required before a useful, practical aid can emerge, and the scope of this essential work is outlined.

This publication has 3 references indexed in Scilit: