Summer forest: native diagnosis and material revision

The isolated diagnostic player in QA/20260916-000225-campaign completed Alps01 and captured three actual driven poses. Each pose has original, Albedo, DiffuseLighting and SpecularLighting PNGs. Camera, bike and frozen-time invariants passed; original debug settings, time and camera drivers were restored. This development artifact is separate from the playable Windows build. Lighting debug modes suppress atmosphere and sky; their PNGs are not additive radiometric measurements.

Confirmed finding

The gray forest ground is already present in Albedo. Specular contribution is not the primary cause. Tools/Art/diagnose_alpine_albedo.py independently casts approximate camera rays against the source DEM to inspect original photograph pixels. It does not alter images. Exposed ground pixel (1800,260) samples source RGB (58,60,55), versus native Albedo (58,59,53). Pixel (1540,340) samples (59,63,62), versus native (56,60,59). Small differences include filtering, import compression and the approximate ray/mesh match. Tree-covered pixels cannot be compared this way.

These samples support correct geographic texture sampling. There is no demonstrated UV or white-fallback shader error. The orthophoto itself contains predominantly gray-brown woodland at these locations. Its exact capture season is unknown; a leafless/dry appearance is an interpretation, not metadata.

Bounded art revision

Only summer woodland layer 3 receives the authored SetColor inputs (0.40, 0.60, 0.30), chosen to bring the woodland appearance closer to the summer tree crowns. These are material-colour inputs, not established direct shader-linear multipliers. This is a deliberate art adjustment, not a repaired light calculation. The source photo bytes, geology layer 2, winter layers, DEM, route geometry and scene lighting remain unchanged. The emissive tree-billboard image is a visual reference, not a measured physical ground albedo.

Native review of the ordinary Windows build is required before claiming improvement. This correction alone does not establish photorealism or AAA quality.

Colour-space audit correction

The earlier statement that this is a linear multiplier producing approximately 54% luminance is superseded. That CPU estimate incorrectly treated the SetColor inputs as direct shader-linear constants. In the pinned HDRP 17.3.0 LayeredLit.shader, _BaseColor0..3 are ordinary Color properties. The project uses Linear colour space. Unity documents conversion of material sRGB colours when supplied to the shader; storage behaviour described by SetColor is a separate step. See Unity 6.3 shader colour properties and SetColor.

Under that documented contract, the expected shader-linear multiplier is approximately (0.132868, 0.318547, 0.073239). This is an API-derived expectation, not a measurement of this Player's GPU constants or rendered luminance. No replacement luminance percentage is claimed. Existing native observations remain evidence of the actual appearance. The authored values have not been converted with .gamma, and this audit changes no material value or source texture.