StandardMaterial::emissive units not documented

[JMS55]

The emissive property does not document what radiometric/photometric unit it's in.

[JMS55]

glTF emissive texture spec https://registry.khronos.org/glTF/specs/2.0/glTF-2.0.html#additional-textures

[NthTensor]

The gltf spec seems to indicate the emissive texture represents quantities of Luminance. It indicates this indirectly by giving the units as Nits (or Candela per Square Meter, or Lumens per Steradian per Square Meter).

[fintelia]

Frustratingly, the linked spec actually gives two mutually conflicting definitions of the scale. It first says the unit is nits:

For implementations where a physical light unit is needed, the units for the multiplicative product of the emissive texture and factor are candela per square meter (cd / m2), sometimes called nits.

But then it says that a value of 1.0 might be expected to result in a fully exposed pixel (despite 1 nit being hundreds of times dimmer than a computer monitor...)

[T]he exact conversion from physical light units to the brightness of rendered pixels requires knowledge of the camera’s exposure settings, which are left as an implementation detail unless otherwise defined by a glTF extension.

Many rendering engines simplify this calculation by assuming that an emissive factor of 1.0 results in a fully exposed pixel.

[JMS55]

I think we should go with the former.

[fintelia]

I'm now remembering that this topic was discussed before in #13133. I believe the resolution was to switch from the first option (use nits) to the second one (treat 1.0 as being bright)

[JMS55]

For bevy_solari (pathtracer), I will be going with the former (treating emissive as luminance/nits). I guess for users who want it to match raster, they'll have to use emissive_exposure_weight = 1.0.

We should document this more, though.