Voice-Responsive Virtual Reality Training Environment for Occupational and Non-Professional Voice Users

Abstract

We present a virtual reality environment (VRE) dedicated to improving public speaking and vocal skills. To design our VR system, we first recorded 529 data samples of fifteen students delivering their oral presentations. Next, we used speech processing to analyze voice parameters like pitch, timbre and speech rate. Moreover, six dedicated annotators assessed the stress levels within each speech sample. These procedures helped us select a subset of parameters for real-time animation of virtual characters and environment in response to the speaker's voice. This interactive feedback aims to train users in voice modulation, enhancing their speaking abilities and confidence, as well as self-consciousness of their voice quality to support proper occupational voice usage. Finally, we report on the experimental assessment of our system with five users.