Comparison of speech perception in virtual and real-life acoustic environments

Recent years have seen growing interest in using virtual acoustic scenes in laboratory environments to emulate real-life complex listening environments. Questions remain, however, regarding the degree to which virtual environments reflect typical complex listening environments in daily life. In this study, participants with and without hearing loss completed speech perception testing in eight virtual acoustic scenes. Then, participants were sent home for a week with a hearing aid research platform, the Portable Hearing Lab (PHL), and instructed to seek out situations that represented their typical complex listening environments. The PHL was used to record the environment, and ecological momentary assessment was used to characterize the environment, including the location, number of noise sources, orientationof speech and noise, and access to visual cues. Acoustic, non-acoustic, and speech perception differences between virtual and real-life complex environments were compared. Although most of the complex environments in daily life were encountered at home, the virtual scenes were fairly representative in terms of the sound levels, noise sources, noise locations, and signal-to-noise ratios. Speech perception scaled similarly between the virtual environments and real life as a function of sound level and acoustic complexity. The results generally support the ecological validity of virtual acoustic scenes.