In this paper, we propose an efficient multi-view system that transmits two view sequences for stereoscopic display and audio according to user's request. The process for users' preferences is integrated into MPEG-21 DIA. MultiViewAdaptation descriptors defined in DIA is used to deliver users' preferences and device capabilities. Furthermore, we newly introduce multi-view descriptions that are represented in XML (extensible Markup Language) schema. Our proposed system also provides view-dependent audio according to user preferences. When a user changes a view from the center of the multi-view video to either the left or the right, audio signals are rendered to have directional information of viewer's movement prior to being played out. Integrating audio with multi-view delivers more realistic perceptions.