Users of IVEs typically underestimate distances during blind walking tasks, even though they are accurate at this task in the real world. The cause of this underestimation is still not known. Our previous work found an exception to this effect: When the virtual environment was a realistic, co-located replica of the concurrently occupied real environment, users did not significantly underestimate distances. However, when the replica was rendered in an NPR style, we found that users underestimated distances. In this study we explore whether the inaccuracy in distance estimation could be due to lack of size and distance cues in our NPR IVE, or if it could be due to a lack of presence. We ran blind walking trials in a new replica IVE that combined features of the previous two IVEs. Participants significantly underestimated distances in this environment.