System and method for spatial interaction using automatically positioned cameras

Methods and systems are described for generating a virtual geometry based on position information for at least a first user in a first physical location, a second user in a second physical location, and at least a third user in a third location, determining a virtual line-of-sight in the virtual geometry between the first user and the second user, determining a real-world line-of-sight associated with the virtual line-of-sight by mapping the virtual line-of-sight into the first physical location, automatically moving at least one moveable camera of a plurality of moveable cameras in the first location to a position and orientation along the determined real-world line-of-sight, and displaying a rendered version of the second user along the determined real-world line-of-sight.