Telepresent walking allows visits to remote places such as museums, exhibitions, architecture, or industrial sites with a high degree of realism. While walking freely around in the user environment, the user sees the remote environment through the “eyes” of a remote mobile teleoperator. For that purpose, the user's motion is tracked and transferred to the teleoperator. Without additional processing of the motion data, the size of the remote environment to be explored is limited to the size of the user environment. This paper proposes an extension of telepresent walking to arbitrarily large remote or virtual spaces based on compressing wide-area motion into the available user space. Motion compression is a novel approach and does not make use of scaling or walking-in-place metaphors. Rather, motion compression introduces some deviation of curvature between user motion and teleoperator motion. An optimization approach is used to find the user path of minimum curvature deviation with respect to a given predicted teleoperator path that fits inside the boundaries of the user environment. Turning angles and travel distan ces are mapped with a 1:1 ratio to provide the desired impression of realistic selflocomotion in the teleoperator's environment. The effects of the curvature deviation on inconsistent perception of locomotion are studied in two experiments.