Purchase this article with an account.
Lindsay Houck, John Philbeck; Assessing Distance Estimation in Brief Viewing Durations in Virtual Reality. Journal of Vision 2020;20(11):1705. doi: https://doi.org/10.1167/jov.20.11.1705.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
At views <100 ms, distance underestimation is common. Underestimation is reduced if longer views precede the brief exposures, suggesting a role for memory, yet the mechanisms remain unclear. Virtual environments (VEs) could accelerate this research, due to their widespread accessibility, but unexplained differences between VE versus real world perception raise the possibility that the time course of distance perception in VE may also differ from real world environments. To assess this, Experiment 1 (n = 38) presented cone targets on the floor of a 10x10 m room in an Oculus VE. Participants verbally estimated cone distance (2.2-9.8m). Viewing duration (100 vs. 5000 ms) was manipulated in counterbalanced blocks. Following real-world results, we expected greater underestimation under 100 ms views, but only if preceding the 5000 ms block. Surprisingly, there was slightly LESS underestimation at 100 ms durations, irrespective of block order (p=.009). This suggests underestimation in VE may be partially due to gradual extraction of flatness cues. To reduce performance at 100 ms, Experiment 2 (n = 19), used a target sphere of constant angular size, removing relative angular size. Two blocks of 100 ms glimpses with targets 4.8-30 m distant were interspersed with a 10 sec view of the 10x40 m environment; we expected improvement following the longer glimpse. A third block of 5000 ms views established asymptotic performance. Results showed improvement in block 3 from blocks 1-2 (both p’s<.03), but not blocks 1-2. This suggests that memory of the VE does not facilitate performance in subsequent short duration trials. Underestimation in asymptotic performance was greater in experiment 2, suggesting that relative angular size is a particularly strong cue in VEs. Future work will determine whether the differing patterns here versus past real-world results are due to response mode differences, or instead to bona fide perceptual differences.
This PDF is available to Subscribers Only