Discussion

Despite the first and second reported presence ratings for each condition having been taken within minutes of each other, there was significant variation (.81 correlation). This is indicative of the volatility of reported presence.

The over-all reported presence difference in favor of the 48$^{\circ}$condition, and the lack of difference between the two narrower conditions, is consistent with a common belief in the VE community that one approaches a ``presence threshold'' at about 60$^{\circ}$ FOV.

However, about a third of the participants did not rate the widest FOV condition as having higher presence than the other two conditions. This may reflect a foreground occlusion effect (see Section 3.3.4): at the narrowest FOV, the moving scene may have appeared more distant and consequently a more salient rest frame. It would be interesting to apply the same protocol to the study of 3 wider FOV's, such as 60$^{\circ}$, 80$^{\circ}$ and 100$^{\circ}$, where the boundary is less noticeable and a foreground occlusion effect is less likely to occur.

Two pilot studies were conducted which examined a visual-inertial nulling measure for the same 3 FOV conditions. These are described in Section B.2. A primary purpose of the research described in this chapter was to measure a main effect for treatment with the nulling measure in agreement with prediction and reported presence. Because of the ambiguity in response to the three FOV conditions described above, it was decided that a different set of conditions would provide a better opportunity to establish the nulling measure. For this reason, the pilot studies described in Section B.2 were not pursued. Instead, a meaningful/random manipulation was examined, as described below.

Human Interface Technology Lab