Differences in Three Vection Indices (Latency, Duration, and Magnitude) Induced by "Camera-Moving" and "Object-Moving" in a Virtual Computer Graphics World, Despite Similarity in the Retinal Images.

To create a self-motion (vection) situation in three-dimensional computer graphics (CG), there are mainly two ways: moving a camera toward an object ("camera moving") or by moving the object and its surrounding environment toward the camera ("object moving"). As both methods vary considerably in the amount of computer calculations involved in generating CG, knowing how each method affects self-motion perception should be important to CG-creators and psychologists. Here, we simulated self-motion in a virtual three-dimensional CG-world, without stereoscopic disparity, which correctly reflected the lighting and glare. Self-motion was induced by "camera moving" or by "object moving," which in the present experiments was done by moving a tunnel surrounding the camera toward the camera. This produced two retinal images that were virtually identical in Experiment 1 and very similar in Experiments 2 and 3. The stimuli were presented on a large plasma display to 15 naive participants and induced substantial vection. Three experiments comparing vection strength between the two methods found weak but significant differences. The results suggest that when creating CG visual experiences, "camera-moving" induces stronger vection.

Material surface properties modulate vection strength.

Realistic appearance and complexity in the visual field are known to affect the strength of vection (visually induced self-motion perception). Although surface properties of materials are, therefore, expected to be visual features that influence vection, to date, the results have been mixed. Here, we used computer graphics to simulate self-motion through rendered 3D tunnels constructed from nine different materials (bark, ceramic, fabric, fur, glass, leather, metal, stone, and wood). There are three ways in which the new stimuli are changed from those found in previous studies: (1) as they move, their appearances interactively change with the 3D structures of the simulated world, as do all the lighting effects and 3D geometric appearances, (2) they are colored, (3) and their components covered a large portion of the visual field. The entire inner surface of each tunnel was composed from one of the nine materials, and optic flow was evoked when an observer virtually moved through the tunnel. Bark, fabric, leather, stone, and wood effectively induced strong vection, whereas, ceramic, glass, fur, and metal did not. Regression analyses suggested that low-level image features such as the lighting and amplitude of spatial frequency were the main factors that modulated vection strength. Additionally, subjective impressions of the nine surface materials showed that the perceived depth, smoothness, and rigidity were related to the perceived vection strength. Overall, our results indicate that surface properties of materials do indeed modulate vection strength.