Music can be described best by music. However, current research in the design of user interfaces for the exploration of music collections has mainly focused on visualization aspects ignoring possible benefits from spatialized music playback. We describe our first development steps towards two novel user-interface designs: The Sonic Radar arranges a fixed number of prototypes resulting from a content-based clustering process in a circle around the user’s standpoint. To derive an auralization of the scene, we introduce the concept of an aural focus of perception that adapts well-known principles from the visual domain. The Sonic SOM is based on Kohonen’s SelfOrganizing Map. It helps the user in understanding the structure of his music collection by positioning titles on a two-dimensional grid according to their high-dimensional similarity. We show how our auralization concept can be adapted to extend this visualization technique and thereby support multimodal navigation.