Accent Lab
Six families, one field

FINDING 1Every family hits the same ceiling. None of them clears the bright field.

Studio’s field sits at L = 0.9210. The highest flare ceiling in the whole set is Instrument Cyan’s L ≤ 0.1151 — a hair above amber’s own 0.1135, and still eight times too dark to matter. On a bright field no saturated accent can be the armed core at 3:1, for the same physics reason amber cannot. Law C is not an amber workaround. It is what a bright field does to any hue. That is worth knowing before the owner picks: the burn treatment is not a compromise forced by the incumbent colour, it is the bright field’s native armed grammar, and it stays whatever he chooses.

FINDING 3“Isn’t a bright field also the current template look?” — no, and the number says which side of neutral you are on

This is the obvious objection to FINDING 2 and it deserves a measured answer, because R4 names two field tells, not one. The first is permanent dark mode (R4-accent-trends.md:53). The second is the 2026 backlash-default (R4-accent-trends.md:54): a cream/beige ground — #faf8f5, #f5f1e8 — under forest-green or sage, which became a cliché precisely because every piece of anti-slop advice converged on the same escape route. Leaving dark mode is not automatically leaving the template; it can be a lane change into the other template.

Studio is not that ground, and the separation is not a matter of taste — it is the sign of b* in CIELAB. Cream is warm: b* is positive. Studio is a cool near-white: b* is negative on all three of its grounds. They sit on opposite sides of the neutral axis.

Same lightness band — L* 92.7–97.7 for every row — and opposite chroma sign. The gap is small in absolute terms (2.75 b* units to the cream, 5.89 to the beige) but it is the difference between “warm paper” and “cold paper”, which is the whole read of the surface. And the accent seals it: the backlash-default pairs its cream with sage or forest green. Of the six families in §03, the one that would land MEtronome inside that cliché is Jewel Pine — and §07 already excludes it, for an unrelated reason (it sits 6° from success, §05). Two independent disqualifications is not a coincidence worth leaning on, so take this as the second one and not the first: a family can hold every ratio it needs and still fail the originality test. Pine holds all of them, on both fields. It is still the one hue on the page that a 2026 viewer has been primed to read as the tasteful default.

[unverified] The b* figures are computed here from sRGB under D65 with the standard CIE matrix; the two cliché hexes are R4’s own, quoted from the source, not sampled from a live product. The claim that cool near-white reads as un-templated to a human eye is a design judgement resting on the sign, not a measured perceptual result.

FINDING 4The inverse of amber — and it points at the owner’s own preference

Amber’s whole problem, per D1 §2 and R4 §2, is that it is too bright to be armed on a light field. The three families with the most headroom on Studio — Oxblood 12.72, Indigo 12.99, Signal Red 11.53 — have the least on graphite, for the mirrored reason: they are too dark to glow. They are bright-field-only accents. That is not a defect to engineer around; it is a constraint that happens to point the same way the owner already went. If he wants a genuinely non-amber accent, Studio is not a preference sitting next to it — Studio is the field where it is possible. The counterpart is equally firm: the accent dimension must be able to say “illegal on this field” in the type system, not in a comment (§06).

THE TESTWould this colour survive being screen-printed on a hardware faceplate?

A screen accent costs nothing — you type six hex digits and the compiler agrees. A faceplate colour costs a paint run, a minimum order, and a decade of the thing sitting on a desk in daylight looking like whatever it looked like. That asymmetry is exactly what R4 §4.3 identified as the fix: a chosen colour tied to a specific reference, not a statistically safe default. So the test is a proxy for deliberateness, and it is falsifiable — either a real manufacturer has committed this colour to physical goods in this category, or nobody has.

It rules out the two things that make generated colour look generated. A hue that only exists because it glows well on black fails it (no faceplate glows). A hue picked because it is inoffensive fails it too (nobody tools up for inoffensive). What passes is a colour somebody paid to be sure about.

1 · The burn treatment is settled, whatever he picks. No family clears the bright field’s armed ceiling (§02). Burn core + flare bloom is the bright field’s grammar, not amber’s workaround. And the field itself already clears both of R4’s field tells: it is not dark, and its b* is negative on all three grounds, putting it on the opposite side of neutral from the cream/beige backlash-default (§02, FINDING 3).

2 · Three families are Studio-only. Oxblood, Indigo and Signal Red cannot glow on a dark field without losing the character they were chosen for (§04). If the owner wants one of those, the theme menu owes the user a bright field, and the code owes the compiler a way to say so.

3 · Two families are already spoken for inside the app. Jewel Pine sits 6° from success; Signal Red sits 2° from destructive (§05). Both are shippable only by re-huing a semantic token — a bigger change than an accent swap, and one that touches screens outside this arc.

4 · That leaves Instrument Cyan and Oxblood as the two clean answers — which is exactly where R4 §6 landed from evidence alone, now with the numbers and the collisions agreeing. Cyan is the only family that clears every gate on both fields. Oxblood is the strongest character and the best faceplate case, at the price of being Studio-only and 15° from destructive.

5 · Cognac still does not answer the brief. It is 4° of hue from the incumbent, has no sourced precedent, and its Console window is 0.05 wide. It is on this page as the honest control for “maybe he meant saturation, not hue” — and if that is what he meant, the answer is not Cognac, it is Ember with the glow discipline of §02 applied.