[davidtorcivia]

What this changes

Gamma 2.2 and 2.4 compressed HDR highlights at 100% brightness because the headroom blend started from a clamped SDR-white value. At 200-nit SDR white, the old math mapped a 1,000-nit input to approximately 364 nits.

Preserve the headroom target's full signal range and fade only the correction at SDR white. Gamma alone now passes through every HDR LUT sample above SDR white; brightness alone follows the requested dimming curve. Apply the same correction to the measured-response LUT path. Update the comparison renderer, whitepaper equation, release notes, and version to 1.9.1.

Checklist

• dotnet test src/Gloam.sln -c Release passes: 1,351 tests.
• dotnet build src/Gloam.sln -c Release -p:GloamStrictBuild=true is warning-free.
• No existing test was weakened, skipped, or deleted to make this pass.
• Changes under src/Gloam.Core have tests.

Calibration math

• Pinning tests were added before the behavioral change and committed separately. All 12 new cases failed before the fix and passed afterward.
• The whitepaper HTML and equation source are updated; Markdown was regenerated.
• The golden fixtures pass without baseline changes.

Regression coverage includes gamma 2.2/2.4, SDR white levels from 80 to 1,000 nits, measured-response LUTs, linear and perceptual dimming, shadow-only controls, and exact identity entries after 256-entry/16-bit hardware-ramp conversion. Existing night-mode boost and continuity tests pass.

What was measured? Generated LUT signals and the hardware-ramp conversion were checked numerically. No new physical display/colorimeter measurements were performed. The 1,000-to-364-nit example describes the computed old transform, not a measured panel output.