|
Brooks continued researching advanced visualization technology, in particular, immersive virtual environments, work also funded in major parts by ONR and NIH. The science this year was focused on measuring the effects of various system parameters on virtual environment effectiveness. Students not supported by the ITR but under Brooks's direction:
(1) Studied the effects of various approximate lighting models on subjects' sensations of presence and their task performance. A major finding is that subjects performed as well with OpenGL lighting as with the more costly and more realistic global illumination. This was quite a surprise. This work by Paul Zimmons yielded a Ph.D. dissertation.
(2) Studied the effects of various 3-D sound models on subjects' sound-source-location performance. For an anechoic environment, subjects performed as well with the DirectX sound as with full 3-D sound, also surprising. Studies with echoic environments, which might very well change things radically, will be done next year.
(3) Studied the degree to which subjects walking and stopping at barriers in a virtual world have the same trajectories as in normal real-world walking and blindered real-world walking that simulates the field of view in virtual worlds. As hoped, subjects who really walk about in a virtual environment do it with essentially the same velocity and direction profiles as in the real world, but somewhat slower.
(4) Studied data-driven Gaussian colored dots, static and moving, as a way of visualizing multiple functions of x,y defined on the same terrain. The object of such visualization is to make each layer separately understandable while enabling correlations among layers to also be evident to the viewer. Alexandra Bokinsky showed in her Ph.D. dissertation that up to nine static layers can be effectively studied, discriminated, and correlated by her technique. Moving Gaussian dots work even better.