The study of auditory spatial cognition is usually concerned with the perception of sound emitting objects. Sound reflections at surfaces, such as the walls of a room, may then be considered disturbing: Reflections are rarely targets which need to be localised, and the perception of echoes is at least partially suppressed by the auditory system. Conversely, we are interested in the abilities of human listeners to determine the spatial location of a particular reflective wall of a virtual room based on sound reflections of self-produced sounds, much like some blind humans habitually do when they echolocate. For this, we compare localisation performance when a listener controls a virtual sound source attached to a part of their body (head or hand); when they generate sounds with their tongue; and when a wall of a room acts as a sound source itself, depending on the pointing direction of the listener's hand. To facilitate our experiments, we have extended the real-time, low-latency "liveRAZR" implementation of the perceptually plausible room-acoustical simulation (Wendt et al., J Audio Eng Soc 62, 2014) to support the online processing of room geometries based on quadrilateral meshes and the acoustic coupling of multiple rooms that are connected by doors.