Each work is a pair. The left panel is a photograph as it was made. The right panel is the same photograph carrying one intervention derived from its own measurements. Side by side, the difference often lives at the threshold of vision, and it is stated exactly, down to fractions of a single tonal step. The numbers are the work, and anyone may measure them again. You are offered almost nothing to see, and everything to check. · · · The water has been re-lit to the brightness the leaves' angles call for, so the pond's two readings of itself now nearly agree. This photograph contains two measurements coupled by one model. Take the lily pads as circles lying in one plane. Seen obliquely they compress, and their foreshortening reads out the local viewing angle, 69.6° from vertical at the top of the frame to 51.7° at the foot. The pads are a protractor. The gaps between them are a specular surface whose brightness, in the simplest model of a flat water surface under one sky, should follow Fresnel's equation, reflectance climbing toward grazing incidence, and it nearly does, falling from 125.0 in the top band to 74.2 at the bottom. Fitting the Fresnel curve at the angles the leaves report recovers a fitted sky term of 484, above the file's ceiling, and a baseline of 57 that the viewing angle does not touch. Against that curve the water's measured profile sat at RMS 11.87 levels, while assuming the water is simply one brightness gives 17.93. Before anything was touched, the Fresnel curve already accounted for a third of the error that flatness leaves behind. The geometry was then allowed to correct the photometry, and the open water was re-lit to what the leaves require. The RMS from Fresnel falls from 11.87 to 3.01 levels, a 75% reduction. The pads change by exactly 0.00 levels. The correction is carried by everything the green mask does not claim, the two floating blossoms included, not because petals obey Fresnel but because the rule was the mask, and the rule was kept. The median change across the frame is 0, and the largest excursions sit in the widest patches of open water, where there was the most to correct. Presentation. The mount is RGB (50, 57, 72), the pond at the model's baseline, the sky's angular term subtracted. The Fresnel fit recovers two constants, a sky too bright for the file to contain and a baseline of 57 the angles never touch, the component the fit assigns to the water rather than to the sky's grazing flare. Every square inch of the photograph carries some of that angular reflection, so this baseline is a view no part of the picture quite holds. The board supplies it. Both panels are placed at 1:1, since the source is large enough that nothing in this work is resampled. There is no rule. To check. 3,355,871 pixels move, none by more than 59 of 255. Download this work's original file, 10_fresnel.png, not a screenshot or a marketplace preview, attach it to an AI that can run code, and paste this: Run this exact line on the attached file, changing only the file path if needed. Show me its exact output, and say whether that output is exactly: peak 59 changed 3355871 python3 -c "import numpy, PIL.Image as I; c=numpy.asarray(I.open('10_fresnel.png').convert('RGB')).astype(int); a=c[906:3927,906:3927]; b=c[906:3927,4380:7401]; d=abs(b-a); print('peak', d.max(), 'changed', (d.max(2)>0).sum())"