Private native ground context calibration

Twenty-four effective private context frames compare the unmodified production322source selection with the existing nonperiodic material candidate at4hex and1.25hex ground spans. Both runs use normal Region/Hex controls,canonical natural sites,1440/390viewports and disposable local QA worlds. Only private browser responses change the renderer;artist files,production source contracts,geography and saves are preserved. The three original ground inputs remain native1254x1254.

WOG-A-003576: Preferred private nonperiodic ground calibration at1.25hex span,compared with the unchanged4hex span at the natural grasslands WOG-A-003301 site. Four Region/Hex desktop/phone views have observed ground span ratio0.3125,identical foreground placement/source-only pixels and two exact native Hex subject replays. Direct visual review finds more proportionate grain while preserving the original native source;full material/biome/contact,all-view visual quality,performance,lifecycle and final-composite qualification remain open. The initial versioned-URL no-op is excluded. No production activation,source admission or artwork approval.

WOG-A-003578: Preferred private nonperiodic ground calibration at1.25hex span,compared with the unchanged4hex span at the natural wetlands WOG-A-003315 site. Four Region/Hex desktop/phone views have observed ground span ratio0.3125,identical foreground placement/source-only pixels and two exact native Hex subject replays. Direct visual review finds more proportionate grain while preserving the original native source;full material/biome/contact,all-view visual quality,performance,lifecycle and final-composite qualification remain open. The initial versioned-URL no-op is excluded. No production activation,source admission or artwork approval.

WOG-A-003579: Preferred private nonperiodic ground calibration at1.25hex span,compared with the unchanged4hex span at the natural fields WOG-A-003307 site. Four Region/Hex desktop/phone views have observed ground span ratio0.3125,identical foreground placement/source-only pixels and two exact native Hex subject replays. Direct visual review finds more proportionate grain while preserving the original native source;full material/biome/contact,all-view visual quality,performance,lifecycle and final-composite qualification remain open. The initial versioned-URL no-op is excluded. No production activation,source admission or artwork approval.

All six four-pane control/candidate sheets were directly viewed. Smaller grass and soil grain removes the oversized stems around the cottage;smaller wetland roots expose the strata silhouette more clearly;the finer meadow looks more proportionate to the courtyard keep. The native subject detail is unchanged. The wetland material/state fit,the keep neighbor/farm contact and other biomes/ground/water sources still require individual review. No whole-asset approval is granted.

The first requested fine-span attempt was ineffective because versioned URLs did not match unversioned strings. Its12frames are diagnostic no-ops,not additional successful calibration coverage. All three witnesses are byte-identical to controls;automatic approval review blocked their removal,so15MBof redundant PNGs and compact failed-run evidence remain. No alternate deletion was attempted.

The corrected run normalizes URL pathnames and requires the actual native-ground span ratio0.3125,identical foreground transform and identical source-only RGBA before retaining each frame. All12pass,including six exact native Hex subject replays. These checks do not prove final composed ground/subject pixel attribution. Seven scoped CPU material tests pass;readback-free browser performance,context loss/CPU fallback and legitimate state lifecycle are still pending. The nine earlier same-viewport composite differences remain unexplained.