Moving ahead with peephole pointing: modelling object selection with head-worn display field of view limitations

Head-worn displays (HWDs) are now becoming widely available, which will allow researchers to explore sophisticated interface designs that support rich user productivity features. In a large virtual workspace, the limited available field of view (FoV) may cause objects to be located outside of the available viewing area, requiring users to first locate an item using head motion before making a selection. However, FoV varies widely across different devices, with an unknown impact on interface usability. We present a user study to test two-step selection models previously proposed for 'peephole pointing' in large virtual workspaces on mobile devices. Using a CAVE environment to simulate the FoV restriction of stereoscopic HWDs, we compare two different input methods, direct pointing, and raycasting in a selection task with varying FoV width. We find a very strong fit in this context, comparable to the prediction accuracy in the original studies, and much more accurate than the traditional Fitts' law model. We detect an advantage of direct pointing over raycasting, particularly with small targets. Moreover, we find that this advantage of direct pointing diminishes with decreasing FoV.