In this paper, I describe a possible creative model for approaching realtime 3D Multi-user Virtual Environments (MUVEs) as formal, audio-visual composition environments. The model is a result of my practice-based research, creating audio-visual art work in Second Life and other realtime 3D MUVEs. Some of the conventions and approaches of musical composition, sound art and visual art are considered and compared. These approaches are evaluated within the context of realtime 3D MUVEs, and compared to a post-convergent approach attempting to identify some qualities native to the realtime 3D MUVE. Keywords Multi-User Virtual Environments, Second Life, Audio-Visual Composition, Post-Convergent Space.