This paper presents a schematic eye model designed for use by virtual environments researchers and practitioners. This model, based on a combination of several ophthalmic models, attempts to very closely approximate a user's optical centers and intraocular separation using as little as a single measurement of pupillary distance (PD). Typically, these parameters are loosely approximated based on the PD of the user while converged to some known distance. However, this may not be sufficient for users to accurately perform spatially sensitive tasks in the near field. We investigate this possibility by comparing the impact of several common PD-based models and our schematic eye model on users' ability to accurately match real and virtual targets in depth. This was done using a specially designed display and robotic positioning apparatus that allowed sub-millimeter measurement of target positions and user responses. We found that the schematic eye model resulted in significantly improved real to virtual matches with average accuracy, in some cases, well under 1mm. We also present a novel, low-cost method of accurately measuring PD using an off-the-shelf trial frame and pinhole filters. We validated this method by comparing its measurements against those taken using an ophthalmic autorefractor. Significant differences were not found between the two methods.
