The current generation of VoIP services offer only single channel (mono) party-mix of voices. Thus when added to a network game engine, mono VoIP service can not really contribute to the gamers’ sense of virtual world immersion. As a future development, the Immersive VoIP service delivers to each gamer an Auditory Scene, mixing the live voices of surrounding gamers which are all directionally placed and distance attenuated according to the appropriate virtual world positions. In previous work, we have proposed the Wireless Immersive Communication Environment (WICE) which is special type of Immersive VoIP service tailored to the scarce resources of wireless gaming clients. Also in previous work, we have already addressed the processing scalability of WICE with the conjecture of distance-governed relaxation of acceptable angular errors which allows multiple gamers to share the voice localised along the same direction. In this work, we verify the conjecture of acceptable angular errors in a series of subjective listening tests. An important test finding is the apparent user sensitivity to angular shifts in voice localisation when such auditory movements correspond to little or no visual movements. This finding suggests that the re-establishment of Auditory Scenes across time can not be memoryless in the face of gamer mobility in the virtual world. A mechanism has thus been developed to address this issue and analytic results are obtained on the impact of virtual world mobility on the required execution frequency of such mechanism.