Three experiments are reported which aimed at analyzing the learnability of discrete speech as well as the joint influences of discrete speech, modality and location of input commands and speech velocity on performance (composing standardized business letters; reading short texts). Results reveal that discrete speech can be learned quickly, and that a considerable amount of additional tasks can be managed while speaking discretely. Some consequences of these results are discussed which may be relevant for the design of interfaces to speech-to-text systems. Keywords: Discrete speech; Human Factors; interface design; speech recognition; speech-to-text technology.
Bibliographic reference. Day, P. / Grünupp, A. / Muthig, K.-P. (1991): "A human factors study of speech-to-text technology: consequences of discrete speech", In EUROSPEECH-1991, 61-64.