Spatial Semantic Segmentation of Sound Scenes (S5)