Refactor clearcoat extension to look similar to Appendix B.

extensions/2.0/Khronos/KHR_materials_clearcoat/README.md

@@ -61,28 +61,100 @@ All implementations should use the same calculations for the BRDF inputs. Implem
 |**clearcoatRoughnessFactor**      | `number`                                                                        | The clearcoat layer roughness.         | No, default: `0.0`   |
 |**clearcoatRoughnessTexture**     | [`textureInfo`](/specification/2.0/README.md#reference-textureInfo)             | The clearcoat layer roughness texture. | No                   |
 |**clearcoatNormalTexture**        | [`normalTextureInfo`](/specification/2.0/README.md#reference-normaltextureinfo) | The clearcoat normal map texture.      | No                   |
-
-The clearcoat formula `f_clearcoat` is computed using the specular term from the glTF 2.0 Metallic-Roughness material, defined in [Appendix B](/specification/2.0/README.md#appendix-b-brdf-implementation).  Use specular F0 equal `0.04`, base color black `0.0, 0.0, 0.0`, metallic value `0.0`, and the clearcoat roughness value defined in this extension as follows:
+
+If `clearcoatFactor` (in the extension) is zero, the whole clearcoat layer is disabled.
+
+The values for clearcoat layer intensity and clearcoat roughness can be defined using factors, textures, or both. If the `clearcoatTexture` or `clearcoatRoughnessTexture` is not given, respective texture components are assumed to have a value of 1.0. All clearcoat textures contain RGB components in linear space. If both factors and textures are present, the factor value acts as a linear multiplier for the corresponding texture values.
 
 ```
+clearcoat = clearcoatFactor * clearcoatTexture.r
 clearcoatRoughness = clearcoatRoughnessFactor * clearcoatRoughnessTexture.g
 ```
 
-The following abstract code describes how the base and clearcoat layers should be blended together:
-
+If `clearcoatNormalTexture` is not given, no normal mapping is applied to the clear coat layer, even if normal mapping is applied to the base material.  Otherwise, `clearcoatNormalTexture` may be a reference to the same normal map used by the base material, or any other compatible normal map.
+
+The clearcoat effect is modeled via a microfacet BRDF. The BRDF is layered on top of the glTF 2.0 Metallic-Roughness material, including emission and all extensions, using a new `fresnel_coat` function:
+
+```
+# from glTF 2.0 Metallic-Roughness material (core)
+material = mix(dielectric_brdf, metal_brdf, metallic)
+
+# clearcoat
+clearcoat_brdf = specular_brdf(
+  normal = clearcoatNormal,
+  α = clearcoatRoughness^2)
+
+# layering
+coated_material =
+  fresnel_coat(
+    normal = clearcoatNormal,
+    ior = 1.5,
+    weight = clearcoat,
+    base = material,
+    layer = clearcoat_brdf)
 ```
-clearcoatBlendFactor = clearcoatTexture.r * clearcoatFactor
-clearcoatFresnel = fresnel(0.04, NdotV)
 
-color = (f_emissive + f_diffuse + f_specular) * (1.0 - clearcoatBlendFactor * clearcoatFresnel) +
-        f_clearcoat * clearcoatBlendFactor
+The `fresnel_coat` function adds a BRDF as a layer on top of another BSDF according to `weight` and a Fresnel term. The layer is weighted with `weight*fresnel(ior)`. The base is weighted with `1-(weight*fresnel(ior))`. `normal` is a float3 vector that affects the top layer, but not the base.
+
+### Implementation
+
+*This section is non-normative.*
+
+Implementations of the BRDF or the layering operator can vary based on device performance and resource constraints.
+
+#### Clearcoat BRDF
+
+The specular BRDF for the clearcoat layer `clearcoat_brdf` is computed using the specular term from the glTF 2.0 Metallic-Roughness material, defined in [Appendix B](/specification/2.0/README.md#appendix-b-brdf-implementation). Roughness and normal are taken from `clearcoatNormal` and `clearcoatRoughness`.
+
+#### Layering
+
+The `fresnel_coat` function is computed using the Schlick Fresnel term from the glTF 2.0 Metallic-Roughness material, defined in [Appendix B](/specification/2.0/README.md#appendix-b-brdf-implementation).
+
+```
+function fresnel_coat(normal, ior, weight, base, layer) {
+  f0 = ((1-ior)/(1+ior))^2
+  fr = f0 + (1 - f0)*(1 - abs(NdotV))^5   // N = normal
+  return mix(base, layer, weight * fr)
+}
 ```
 
-If `clearcoatFactor` (in the extension) is zero, the whole clearcoat layer is disabled.
+Applying the functions we arrive at the coated material
 
-The values for clearcoat layer intensity and clearcoat roughness can be defined using factors, textures, or both. If the `clearcoatTexture` or `clearcoatRoughnessTexture` is not given, respective texture components are assumed to have a value of 1.0. All clearcoat textures contain RGB components in linear space. If both factors and textures are present, the factor value acts as a linear multiplier for the corresponding texture values.
+```
+coated_material = mix(material, clearcoat_brdf(clearcoatRughness^2), clearcoat * (0.04 + (1 - 0.04) * (1 - NdotV)^5))
+```
 
-If `clearcoatNormalTexture` is not given, no normal mapping is applied to the clear coat layer, even if normal mapping is applied to the base material.  Otherwise, `clearcoatNormalTexture` may be a reference to the same normal map used by the base material, or any other compatible normal map.
+and finally, substituting and simplifying, using some symbols from [Appendix B](/specification/2.0/README.md#appendix-b-brdf-implementation) and `Nc` for the clearcoat normal:
+
+```
+clearcoatFresnel = 0.04 + (1 - 0.04) * (1 - abs(VdotNc))^5
+clearcoatAlpha = clearcoatRoughness^2
+
+f_clearcoat = clearcoatFresnel * D(clearcoatAlpha) * G / (4 * abs(VdotNc) * abs(LdotNc))
+
+coated_material = (f_diffuse + f_specular) * (1 - clearcoat * clearcoatFresnel) +
+                  f_clearcoat * clearcoat
+```
+
+#### Emission
+
+The clearcoat layer is on top of emission in the layering stack. Consequently, the emission is darkened by the Fresnel term.
+
+```
+coated_emission = emission * (0.04 + (1 - 0.04) * (1 - NdotV)^5)
+```
+
+#### Discussion
+
+In order to make the material energy conserving with a simple layering function, we compute the microfacet Fresnel term with `NdotV` instead of `VdotH`. That means that we ignore the orientation of the microsurface. As the clearcoat roughness is usually very low the microfacets orientation is very close to the normal direction, and `NdotV ≈ NdotL`.
+
+The simple layering function ignores many effects that occur between clearcoat and base layer. For example:
+* The clearcoat layer is assumed to be infinitely thin. There is no refraction.
+* The index of refraction of clearcoat and base layer do not influence each other. The Fresnel terms are computed independently.
+* There is no scattering between layers.
+* There is no diffraction.
+
+More sophisticated layering techniques that improve the accuracy of the renderings are described in [Appendix B](/specification/2.0/README.md#appendix-b-brdf-implementation).
 
 ## Schema