The tools for spatial composition typically model just a small subset of the spatial audio cues known to researchers. As composers explore this medium it has become evident that the nature of spatial sound perception is complex. Yet interfaces for spatial composition are often simplistic and the end results can be disappointing. This paper presents an interface that is designed to liberate the composer from thinking of spatialised sound as points in space. Instead, visual images are used to define sound in terms of shape, size and location. Images can be sequenced into video, thereby creating rich and complex temporal soundscapes. The interface offers both the ability to craft soundscapes and also compose their evolution in time.