This lab is the front door. The same rule the canon uses for the ear — physics first, clinic second — governs the eye. So this bench doesn't belong to one college. It belongs to the seam between two: it begins under the prism at Stephens Science and walks the student all the way to a chair in the B.J. Medical Center.
The canon already handles audiology this exact way: "the ear is a wave problem before it's a clinical one." Sound goes to Stephens first, then comes back clinical at VII. Vision follows the same law. You cannot screen an eye you don't understand as an optical instrument — and you can't understand the instrument without the three-cone collapse this lab makes you watch happen.
Came up through laser physics before he ever held a scalpel — which is why he holds the joint IX ⇄ VII appointment vision already requires. He teaches PHYS 244 at Stephens and the vision-screening cross-list at B.J. Medical, the same students walking both doors. "I spent ten years learning how light bends. Then thirty learning how to bend it back. Same equation — the cornea just doesn't care that you passed the physics exam."
In honor of Dr. Ming Wang of Nashville — laser physicist turned eye surgeon, who has restored sight in 55 countries free of charge. Wing Ming is a fictional OPA character; the dedication is real.
A browser perception lab built on one question: what shape is the space of all the colors you can see? The answer turns out to need three disciplines at once.
Physics hands the eye a continuous spectrum — effectively infinite-dimensional. Anatomy collapses it onto three numbers, because the retina carries exactly three cone types (L, M, S). Geometry then asks how those three-number colors are spaced apart from one another — and the answer, settled only recently, is stranger than a century of theory assumed.
Color space is three-dimensional not because color is, but because you have three cones. And the distances inside that space don't obey ordinary curved-space (Riemannian) rules: large color differences are perceived as less than the sum of the small differences that compose them. That single fact — "diminishing returns" — breaks the model Riemann, Helmholtz, and Schrödinger built.
The cone sensitivity curves are smooth approximations of the real Stockman–Sharpe cone fundamentals, drawn for teaching, not photometric work. The metamer match, the diminishing-returns gap, and the Bezold–Brücke shift are illustrative of the real phenomena; the numbers are tuned for legibility, not lab-grade ΔE values. The geometry visual uses an HSL bicone as a stand-in for the true perceptual solid. Use the cited papers for real coordinates.
OPA Physics Suite · companion to Light & Optics (which handles the physics of the light before it reaches the eye). Where Light & Optics ends at the retina, this lab begins. Cross-listed Stephens Science (IX) ⇄ B.J. Medical (VII) as the optical front door to the clinical vision track — same routing the canon gives audiology: physics first, clinic second. Where the trees grow toward each other; the roots already touch underground.
Three cone types means any incoming spectrum is reduced to three responses (L, M, S) — a 3-vector. Two different spectra with the same 3-vector are metamers: indistinguishable to the eye. Color space is the space of these 3-vectors.
If the whole is reliably perceived as smaller than the sum of its parts, no Riemannian metric can fit the data — additivity along shortest paths is a defining property of such spaces. This is the core finding of Bujack et al. (2022).
An isoluminant hue circle around gray measures a circumference larger than its straight-line geometry predicts — hue differences seem to carry extra weight. It's often cited as evidence of weird color geometry, but the 2022 authors note this particular effect can still be modeled inside a Riemannian space. So it's a curiosity, not the proof. Diminishing returns is the proof.
The follow-on work defined the neutral (gray) axis directly from the color metric — the piece Schrödinger's hue/saturation/lightness definitions depended on but never specified — and corrected the Bezold–Brücke hue shift by replacing straight-line distance with the shortest geodesic path through the (non-Riemannian) space.
Bujack, R., Teti, E., Miller, J., Caffrey, E., & Turton, T. L. "The non-Riemannian nature of perceptual color space." PNAS 119(18), 2022. DOI: 10.1073/pnas.2119753119.
Bujack, R., et al. "The Geometry of Color in the Light of a Non-Riemannian Space." Computer Graphics Forum, 2025 — the neutral-axis definition and the completion of Schrödinger's framework. Both from Los Alamos National Laboratory.