Peripheral Stimulation and its Effect on Perceived Spatial Scale in Virtual Environments J. Adam Jones ∗ Institute for Creative Technologies University of Southern California J. Edward Swan II † Dept. of Computer Science & Engineering Mississippi State University Mark Bolas ‡ Institute for Creative Technologies University of Southern California ABSTRACT The following series of experiments explore the effect of static pe- ripheral stimulation on the perception of distance and spatial scale in a typical head-mounted virtual environment. It was found that applying constant white light in an observer’s far periphery enabled the observer to more accurately judge distances using blind walk- ing. An effect of similar magnitude was also found when observers estimated the size of a virtual space using a visual scale task. The presence of the effect across multiple psychophysical tasks pro- vided confidence that a perceptual change was, in fact, being in- voked by the addition of the peripheral stimulation. These results were also compared to observer performance in a very large field of view virtual environment and in the real world. The subsequent findings raise the possibility that distance judgments in virtual en- vironments might be considerably more similar to those in the real world than previous work has suggested. Index Terms: I.3.7 [Computer Graphics]: Three-Dimensional Graphics and Realism—Virtual Reality; I.4.8 [Scene Analysis]: Depth Cues H.5.1 [Information Systems]: Multimedia Information Systems—Artifical, Augmented, and Virtual Realities; H.1.2 [In- formation Systems]: User/Machine Systems—Human Factors 1 I NTRODUCTION Users of virtual environments, when judging positions along the ground plane, have consistently indicated that the apparent dis- tance of objects appear closer than their actual geometric posi- tion [5, 13–15, 19, 25, 30, 32, 33, 37, 38, 40, 44–47]. This effect is confusing as great lengths are often taken to ensure that the reti- nal projection of the virtual environment closely matches that of an equivalent real world environment [2, 28, 31, 42]. This is even more interesting when comparing these results to the significantly larger body of distance judgment literature that exists for real world environments. These studies quite uniformly demonstrate veridi- cal performance across a large range of distances when full visual cues are available [19, 25, 29, 41, 43]. However, it is frequently the case that not all visual signals available in the real world can be reasonably reproduced in the virtual world. It is an open question as to which of these signals are necessary to enable accurate spatial perception in virtual environments. Jones et al. [14,16] provide ev- idence indicating that visual stimulation in the far periphery may be an important source of information when performing spatial judg- ments. This was an intriguing result as many virtual environments seldom present visual stimulation outside of the near periphery, see Figure 1. However, their work was not conclusive as to the exact manner by which this stimulation provided benefit. This document details a series of five experiments intended to expand upon the ∗ e-mail: jadamj@acm.org † e-mail: swan@acm.org ‡ e-mail: bolas@ict.usc.edu Figure 1: Regions within the human binocular field of view. Adapted from Taylor [39], Boring et al. [3], and Strasburger et al. [36]. work described by Jones et al. [14, 16] and to further inform the nature of the benefit provided by peripheral stimulation. 2 RELATED WORK It has long been established that vision is a strong guide to ac- tion, especially with regard to our movements within an environ- ment [9–11, 41]. As such, visually guided movements, including walking, jumping, pointing, or reaching, are generally well cali- brated with regard to our spatial understanding of the world around us [8, 10, 18, 29, 32, 41]. This connection between movement and vision has led to a considerable body of research devoted to how we judge distances under varying viewing restrictions. Cutting and Vishton [6,7] offer an informative introduction to the basic vi- sual cues indicative of distance and their effectiveness across three ranges of distances: personal, action, and vista space. An often used alternative nomenclature for these spaces is respectively near, medium, and far field. These labels are effectively equivalent, with near field distances being those within arm’s length, medium field being distances extending to roughly 30m, and the far field repre- senting all distances beyond 30m. There is no lack of prior work in the field of distance judgments as visually guided movements are an important part of our every- day lives. In real world viewing conditions, people are quite accu- rate in judging positions across a large range of distances using a variety of methods [19, 24, 29, 32, 33, 37, 41]. Observers have been shown to be able to accurately judge distances within arm’s length