Abstract
Although simulated self-motion through virtual environments has been widely used to investigate perceptual odometry, the characteristics of the virtual environments used, and the reported results, have varied greatly. To investigate how the characteristics of a virtual environment affect the perception of travel distance, we systematically vary the characteristics of the environment through which observers are moved in order to explore the effect of (1) the structure of an environment including the presence and texture of a ground surface, (2) the naturalism and scale of an environment, (3) colour, and (4) the density of a starfield. In all four experiments, participants were visually moved forwards through a virtual environment and perceived travel distance was estimated by either (1) stopping at the location of a previously seen target (the Move-To-Target Task) or (2) adjusting the position of a target to indicate a previously travelled distance (the Adjust-Target Task). Data were analyzed in terms of gain (perceived travel distance/actual travel distance). Results show no significant differences that depended on the structure of an environment or on the presence or absence of a ground surface (Experiment 1), or on the naturalism of the environment (Experiment 2), or on whether the environment was in colour or in black and white (Experiment 3). However, there was a small effect of the texture of the ground surface and of the scale of the environment. In Experiment 4, we show that there may be a very low ceiling effect in the density of a starfield needed to accurately estimate travel distance. Together these experiments have implications for the design of real and virtual environments where perceived motion is important and will enable us to further predict our perception of moving through an environment.