# L4RD industrial lighting audit · r273

The packaged L4RD start remains dark beneath the overpass. A same-camera
Chrome capture compared the existing city-light gain (`0.7`) with `1.5`:

- Road crop `(90,390)-(490,630)` mean RGB changed from `(40.0,41.1,44.1)` to
  `(47.9,49.1,55.2)`.
- Elevated road/ramp crop `(620,60)-(790,220)` changed from
  `(64.7,63.5,60.3)` to `(76.6,75.5,73.6)`.
- Both frames report two local fixtures plus one global light.

At the stopped car position `(465.25,-1357.36,14.54)`, the nearest authored
fixtures are `EL_StreetlightAtoE_032` at `(436.87,-1342.35,33.10)`, 37.08 units
away, and `EL_StreetlightC_129` at `(416.95,-1364.64,23.25)`, 49.62 units away.
Both have radius 30 and inner radius 10. The recovered type-2 radial weight is
zero outside the radius, so neither contributes at that position. The global
gain increase therefore raises distant lit geometry without putting a local
streetlight over this road. This is why `CITY_LIGHT_GAIN` was left unchanged.

The captures establish a local coverage gap in the current light response; they
do not establish that the authored record placement is wrong or what the retail
scene should look like. The shader already uses the recovered source radius,
packed color, and radial weight. A stronger correction needs evidence for the
retail light packet/profile or other city illumination source, rather than a
global gain multiplier.

- [Gain 0.7 capture](../webgl-city/review/images/407-r273-l4rd-light-gain-07.png)
- [Gain 1.5 capture](../webgl-city/review/images/408-r273-l4rd-light-gain-15.png)
- [Viewer lighting implementation](../webgl-city/viewer.js)