Effect of Three Variables on Synthetic Speech Intelligibility in Noisy Environments.

摘要

Military Command and Control (C2) requires easy access to information needed for the commander's situation assessment and direction of troops. Providing this information via synthetic speech is a viable alternative, but additional information is required before speech systems can be implemented for C2 functions. An experiment was conducted to study several factors which may affect the intelligibility of synthetic speech. The factors examined were: (1) speech rate; (2) synthetic speech messages presented at lower, the same, and higher frequencies than background noise frequency; (3) voice richness; and (4) interactions between speech rate, voice fundamental frequency, and voice richness. Response latency and recognition accuracy were measured. Results clearly indicate that increasing speech rate leads to an increase latency and a decrease in recognition accuracy, at least for the novice user. No effect of voice fundamental frequency or richness was demonstrated.