A virtual classroom can elicit teachers’ speech characteristics: evidence from acoustic measurements during in vivo and in virtuo lessons, compared to a free speech control situation

To achieve pedagogic goals and deal with environmental constraints such as noise when lecturing, teachers adapt their speech production in terms of frequency, intensity, and temporal aspects. The mastery of appropriate vocal skills is key to teachers’ speech intelligibility, health, and educational effectiveness. This project tests the relevance of virtual reality (VR) for training teachers’ vocal skills by simulating a lesson in a realistic VR environment characterized by adjustable constraints such as background noise and fidgety children. The VR environment depicts an elementary school classroom with 16 pupils aged 9 to 12 years old animated with typical childlike actions. To validate this virtual classroom in terms of speech characteristics, we conducted acoustic analyses on the speech productions of 30 female teachers in three conditions: (1) giving a free speech while facing the experimenter (control), (2) teaching in their usual classroom (in vivo), and (3) teaching the same lesson in a virtual classroom (in virtuo). The background noise in the VR setting was adjusted for each talker so it was similar to the level measured in vivo. Repeated measures ANOVAs showed that teachers significantly increased their voice frequency, intensity, and intonation and made longer pauses while speaking in vivo and in virtuo, compared to the control condition (p