Let the photographer look at the pixels, and keep looking

Noise reduction and capture sharpening are judgements about individual
pixels, and at a fitted view several of the file's pixels are averaged into
each one on screen. The frame therefore looks cleaner and softer than it is,
the photographer corrects for a softness the display invented, and
over-sharpening is the documented result. Nothing in the develop view
reached 1:1 at all: the wheel and the pinch zoom by ratios, the double tap
dropped straight to fit, and the only readout was a percentage nobody was
aiming at.

So the double tap now does what FR-UI-4 always said it did — toggle fit and
1:1 — and the zoom readout, which used to be a dead "Fit" button on a fitted
photograph, becomes the way in when there is nothing to clear. Z does the
same from the keyboard, and the back gesture goes out through the same
toggle so putting the magnifier down really puts it down. 1:1 is computed
from the file's own resolution against the viewport rather than fixed at
some multiple, because that is the only version of it that answers the
question the two detail controls are asking.

The point being inspected and whether the magnifier is up are held beside
the session rather than in it, on the argument focus peaking already makes:
a session is one photograph and this is a way of looking at a folder of
them. Checking the same eye across forty portraits is the reason to reach
1:1 in the first place, and a magnification that reset with the session
would make that forty zooms and forty pans instead of forty keystrokes. It
stays a viewing state throughout — the view is kept out of `is_active`,
`output_size`, the sidecar and the export, and an export still suspends it —
so none of this reaches the file.

This commit is contained in:

dtourolle

2026-09-06 19:01:50 +02:00

parent 43652bd613

commit 9b674a88d8

4 changed files with 429 additions and 60 deletions

docs/traceability.md

+51 -51

View File

File diff suppressed because one or more lines are too long