Fish Tank VR – Paper Summary
The paper discusses a study which showed that stereoscopic vision is more important than eye-coupled perspective for visually guided hand movements.
Depth judgment is critical while interacting with objects. Research in depth perception focuses more on depth cues and coupling the perspective is important for a VR system. A comparison between stereoscopic depth perception and motion parallax shows that stereo is a very strong cue for judging relative depth of nearby objects, but poor cue for large dept differences. Motion parallax on the other hand increased the sense of realism of virtual environment.
Previous studies showed that stereoscopic depth greatly improve performance of 3 dof pick-and-place tasks. But these studies did not include head coupled perspective. In those studies the subjects were not required to change their viewing position. The study discussed in the paper is similar with the difference that subject had to make substantial changes in viewing position. The goal of the study was to examine the effects of correct vs incorrect perspective and stereoscopic depth for visually guided reaching.
The VR experiment was a variation to Fitts’s experiment, where the subject has to tap between tops of 3D virtual cylinders. The diameters of cylinder tops and distances between them were varied. the independent variables were ‘Stereo vs no Stereo’, ‘Head tracked vs fixed perspective’ and index of difficulty based on distance between targets and size of targets. The results showed that average time for task increased by about 11% without head tracking, the average time was 33% longer with stereo disabled, the advantage of stereo as well as head tracking increased with increase in task difficulty and the average time decreased over course of trails.
The results contrasted with previous experiments that showed that motion parallax was more important depth cue. However the study failed to show the expected evidence for adaptation during the series of taps.
Comments are closed.