Depth-based visual capture of sound in virtual reality is driven by perceived distance

Visual information typically dominates auditory cues in spatial localization, particularly in depth, leading to the proximity image effect (PIE) in which displaced sounds are perceived closer to plausible visual objects than they actually are (visual capture). It is unclear whether fundamental properties of audiovisual integration generalize to virtual environments where spatial perception is systematically distorted. Across two experiments, we tested whether real-world sound sources are visually captured by virtual objects in immersive virtual reality (VR). Participants completed a perceptual-matching task in which they indicated when a real-world sound source aligned with the depth of a virtual object. In Experiment 1, visual capture reliably occurred; however, the effect was opposite to real-world findings, with stronger capture happening when the sound originated in front of the visual object rather than behind it. Experiment 2 replicated this pattern and revealed significant visual distance underestimation measured with a blind-walking task. When judgments of auditory-visual co-location were interpreted relative to the participants' perceived visual distance rather than the actual physical distance, the expected real-world symmetry reemerged. These results demonstrate that core principles of multisensory integration are preserved in VR, but VR-specific biases, such as the visual underestimation of distance, may alter the direction of the established effects.