Two experiments examined spatial knowledge access after imagined perspective switches in everyday environments. Blindfolded volunteers had to point to target objects in a well-known surrounding while imagining being repositioned into different spatial perspectives defined as self-rotations of the observer in the horizontal plane (0 degrees , 45 degrees , and 135 degrees ). Testing was either conducted in the space the target locations had been learned in (actual room testing), or while participants were away from this room, but should imagine being situated there (remote room testing). Experiment 1, in which perspective switches were tested on a trial-to-trial basis, revealed increases in pointing latency and error as a function of the amount of angular disparity between real and imagined perspective under actual and remote room conditions. Experiment 2, in which knowledge access was tested in blocks of multiple trials per perspective, showed increases in pointing latency with angular disparity for actual space testing, and to a much lesser degree for remote space testing. Implications of both findings for theoretical accounts of knowledge access after imaginal perspective switches are discussed, and recommendations for using different testing methods and settings are given.