This paper defends the claim that directly sensed objects are those delivered by sub-personal processes.  It is shown how this thesis runs counter to perceptual atomism, the view that wholes are always sensed indirectly, through their parts.  For in both vision and audition, some directly sensed objects are parts of others.  In other words, some direct sensa  are composite. For example, while the direct objects of audition are all composed of sounds, these direct objects are not all sounds.    There is, in both vision and audition, a great variety of such composite objects – in vision, material objects, shadows, patches of light, vapours, volumes, and so forth; in audition, melodies, harmonies, sequences of phonemes, individual voices, meaning-carrying sounds, and so on.  This diversity of sensory objects has an important application to aesthetics. Perceivers do not naturally or easily attend simultaneously to objects that overlap in space and time.  Yet, aesthetic appreciation depends on such an allocation of attention to overlapping objects.