Audio conferencing is carried out in stereo by centrally assigning a position in the stereo spectrum (either bipolar stereo or surround sound) and instructing each conference endpoint to transmit at the stereo position assigned to it. In this way signals from a number of endpoints can be simply added to one another in order to provide a composite broadcast signal. This method also facilitates multiple server conferences in which endpoints attached to additional conference servers can be instructed to change their virtual stereo position in the case of a conflict with the stereo position of an existing conferee.