Where to stand
The point you have to stand at
How a picture's correct viewing distance follows from its focal length and displayed width, and what follows from nobody standing there.
Wide angle is not distortion
Why wide-angle photographs stretch faces at the frame edge, why it is not a lens fault, and the viewing distance at which it vanishes entirely.
Anamorphosis is only a viewpoint
Anamorphic construction as an ordinary projection with an oblique centre, the homography that produces it, and why straight strokes staying straight is the test.
When the picture surface is not flat
Cylindrical and spherical picture surfaces against the flat plane, what each preserves, and why a panorama bends the horizon.
An anamorph at true size, on paper
The oblique anamorph built by projecting an upright word onto the sheet from the eye, checked by re-projection, and shipped as a printable sheet at true scale.
The cylindrical mirror unbends it
Catoptric anamorphosis computed by tracing the reflection in a cylinder, checked by running the light backwards from the paper mark to the eye, with the unevenness of the map measured as a ratio of local area scales.
Standing in the wrong place
What a perspective picture depicts when it is viewed from somewhere other than its station point — depth scaled by the ratio of the distances, transverse dimensions unchanged, and every image point exactly where it was.
A floor anamorph is three numbers
The map from an upright design to its floor marks is a homology whose axis is the ground line, whose centre is the eye's height plus its distance, and whose ratio is minus the distance over the height.
Where the anamorph still works
A wrong eye composes an anamorph with an elation or a homology on the ground line, and the departure is the step times the height over the eye's height.
The marks name the place, not the height
An anamorph's marks determine the reader's position on the floor exactly and the eye's height only in the product height times aspect.
The ceiling that is not a plane
The map from an intended picture to its painted marks is a collineation exactly while the receiving surface is flat, and the residual across that boundary is fifteen orders of magnitude.
A set cut for one eye
A forced-perspective set matches its intended row exactly from one eye, and the disparity two eyes read is a ratio of reciprocal depths rather than the depth compression the set was cut at.
The room the eye may stand in
The region of eye positions from which a pavement anamorph still reads, measured as a solid rather than a radius, with its shape, its long axis, its volume and how all three scale with the tolerance allowed.
An anamorph has one eye
What the second eye is told about a pavement anamorph, measured as a binocular disparity in arcminutes and set against the disparity the intended picture would have produced.
The design that outruns the floor
How much pavement an intended picture requires, why the growth is a pole at eye level rather than a slope, and what the stretch at the far edge does to a design that tries to include a horizon.
The anamorph that crosses a corner
A design cast across two planes at once — two homologies sharing an axis, continuous at the corner and not smooth, with the scale jump computed from the two grazing angles alone.
The height a flat floor cannot give
Why the eye's height is unrecoverable from marks on a plane, why it becomes recoverable on a floor with any curvature at all, and what that trade says about the price of a projective description.
A stair does not use all its faces
The design of an anamorph on a flight of steps, the faces no ray from the design eye ever reaches, and why every stair mural is painted on the risers.
What an eye can paint
The paintable subset of a faceted object as a function of the eye, generalised from a flight of steps to four other objects, with the design band, the per-face shares and the depth ratio measured on each.
Facing the reader is not being reachable
The difference between a face that turns toward the eye and a face the design reaches, measured as an area rather than as a count, with the height at which occlusion clears entirely.
The eye that reaches the most
Sweeping the eye's height over four faceted objects and finding that the quantity a designer maximises decides the answer, with one object where the optimum is genuinely interior and three where it is at the end of the range.
A design that lands in two rooms
An anamorph whose design lands on surfaces at two very different depths, the scale jump at the edge of the opening, and why the design band on such an object has to be sampled rather than bisected.
The stretch decides the band
The local stretch of an anamorph's map, why it rather than the geometry should decide how much design an object is given, and what a cap on it leaves of each of six surfaces.
One flight, two pictures
Whether one faceted object can carry an anamorph for each of two eyes, measured on a flight of stairs from its two ends, with the asymmetry that one reader sees both pictures and the other sees only theirs.
A fold names the height
Whether a single crease is enough to break the height ambiguity a flat floor leaves an anamorph recovery, measured against the angle of the fold and against the size of the error it is asked to detect.
A flight that has ends
The step ends the earlier flight did not model, the count of disjoint pictures a real staircase supports, and the correction to a rule that counted the wrong thing.