An Examination of the Impact of Synthetic Speech on Unattended Recall in a Dichotic Listening Task

Abstract

Synthetic speech, which is generated by a computer, is widely used in both everyday situations (e.g. GPS devices; weather alerts) and the military (e.g. aviation). Synthetic speech is not identical to spoken speech, as it has a different pacing and varying pronunciations. Participants engaged in a Dichotic Listening Task in which they actively repeated information that was being presented in one ear, while ignoring their other (unattended) ear. The task was adapted and used both synthetic and spoken speech in the attended and unattended ears. It was found that when the unattended information was Spoken participants were more likely to hear an alert word (“fire”) and familiar character names when engaging in a difficult task. When unattended information was in the form of Synthetic Speech less of these low-threshold words were reported by the participants. This research has important implications for the development and use of synthetic speech in high workload situations, and in alerts.