Royston Vane Collection of Optical Accidents

Room 9Handling permitted · the screens are replaceable

The Moiré
Parlour

Acc. RVIX·093·B  ·  two acrylic gratings, cut from one block, re-registered on the light table  ·  bequeathed by the estate of Iris Fell, lens grinder (1896–1974)  ·  not an error, an addition

Fig. 9.3  —  the same marks, twice

State: stacked · comb spokes

Bench — the upper screen

Iris Fell estate

Comb — one block, two impressions

Grating — the discs stay identical

8.0 px
0.0 %
4.00°

Bench fittings

Instrument

60 fps
Relative angle
4.00°
Pitch, lower
8.0 px
Pitch, upper
8.0 px
Fringe, from θ / Δp
—
Separation
0 px
Overlay
complete
Marks, each disc
78

Drag the upper disc to push it. Drag the ticked ring outside it to turn it. Bring the lens over the case and the rings come apart into plain lines.

←↑→↓ push  [ ] turn  S split  L lens  C counter  R reset

Neither screen is doing this.

Wall text — Room 9 · please do not lean on the case

How the rings are made

Two identical combs. One sits still; the other is pushed, turned or detuned. Where their lines cross they darken, and between the crossings something wide and slow appears — a ring, a petal, a lattice of ghosts — at a scale no single line in either disc has any right to reach.

The case is a light table, so the stack multiplies light instead of adding it: the drawing on this page is composited with a multiply blend, which is what two transparencies actually do to a beam. A multiplier is a mixer, and a mixer makes sums and differences. Lay two gratings eight pixels apart at two degrees to each other and out of the arithmetic walks a fringe two hundred and thirty pixels wide — twenty-eight times coarser than either input, and owing its existence to both. What you are looking at is the difference term.

two parallel gratings, pitch p, angle θ  →  D = p / (2 sin(θ/2))
concentric gratings, pitch mismatch Δp  →  D = p¹ / Δp

The lower disc carries the same marks as the upper one. Every bright ring on this case is a claim about both of them at once, and belongs to neither. Release the screens to counter-rotate and each turns at about a hundredth of a degree per frame — nine minutes to the revolution — while the rosettes cross the case every three seconds. The artifact outruns its makers by two orders of magnitude, which is why the room feels animated when nothing in it is.

Note from the registrarNo pigment was added to produce these rings. If you find one you particularly like, the settings are on the case. If you cannot get it back, they are on this screen as well, and they are also nothing.

Where else it happens

On the scanner. Every halftone print — newspaper, comic, the printed painting in the next room — is already a comb: dots laid down at a screen angle, conventionally near 45°, to stay out of trouble. Lay the scanner’s own grid over the top and the two interfere. Conservators photograph such pages rather than scan them, and turn the original two or three degrees on the bed — pushing the beat so far apart in frequency that the optics swallow it.

On the sensor. A camera samples with a mosaic of photosites at a fixed pitch; shoot a shirt, a fence, a facade of louvres, and the repetitive subject approaches that pitch. The mosaic aliases and false bands swim across the fabric. Many cameras carry an optical low-pass filter — a slab of crystal that blurs the image a whisker before it is measured — and the honest remedies remain: step back, change the angle, stop down.

In the print shop. Four ink screens, four angles, spaced about 22.5° apart, because two screens at the same angle produce precisely this nuisance. The rosette you are looking at is the reason the angles exist.

In the fabric trade. The name comes from moirer, to watermark: watered silk, made by pressing two ribbed fabrics together so their weavings beat against one another and leave a standing wave in the sheen. These two discs are the same trick with better tolerances.

In the record. Lord Rayleigh described the bands in 1874, in a paper on what observers saw through ruled gratings at the eyepiece of a telescope; he had them by the yard, and noticed that the coarseness of the artifact was a plain arithmetic function of the coarseness of the two plates. Makers of optical toys sold pairs of discs under turnable bezels for exactly that reason. This room has one such bezel on loan, and it is not as good as the software.

Note from the registrarVisitors frequently report that the rings are “printed on top”. They are not. There are two discs in this case, and there are two discs in this case.

Curatorial instructions

Spokes. The counter-rotating pair is what the object was made for: press Counter-rotate and let go of everything else. Rings. The rosette is the centre of a ring set; push it off the other and the whole disc fills with giant concentric beats — the same arithmetic as Newton’s rings, without any Newton. Hex dots. Two lattices, small angle: straight fringe sets appear, then cross where the mismatch is pushed. This is the geometry the print shop spends its angles avoiding.

Rotation lock holds the upper screen at its present angle so you can study pure offset. Case drift leans the exhibit a few pixels over four minutes, so the interference never sits entirely still while nobody is touching it. Switch it off if it unsettles you; it is meant to.

Catalogue

The Moiré Parlour, RVIX·093·B. Two cast acrylic discs, engraved from a single block by Iris Fell, 1961; re-registered on the light table, 2019. 560 × 560 mm; the mark count stands on the instrument, and changes when you turn the coarseness. Acquired with RVIX·093·A (one disc, unstacked), which the gallery insists is not a different work and the insurance assessor insists is.

Comparative material: the Jesuit Ring (RVIX·011·C), a 60 mm brass bezel holding two counter-rotating engraved discs, the earliest moiré toy in the collection and slower than this one; Joseph Plateau’s rosette wheels, 1835; the Benham disc, which is a different and hotter argument; Line Systems and Striae, Room 4; and the same effect, unwanted, on the surface of the vitrine in Room 2.

Note from the registrarOn the magnifier: the lens adds no detail that is in the glass. It removes a layer. That is the whole of the unmasking, and viewers consistently refuse to believe it until they have done it themselves — which is why it is bolted to your hand.