chaosgroup / vraymtl_glsl Goto Github PK

in function 'vec3 getGGXDir()' in vratmtl_glsl microNormalLocal was scaled before get reflect dir. Then you pass this scaled microNormalLocal to 'getGGXContribution()', and in this function the microNormalLocal was scaled again to calculate BRDF.

is it right to scale this vector twice?

Are threre plans to support or partially support these unsupported features in the future ?

Unsupported features:
Glossy Fresnel (supported only in sheen) because the calculations are a bit heavy for real-time rendering.
SSS and fog parameters
Normal/Bump Mapping - this is outside of the scope of the current implementation because the normal is usually computed separately before any BRDF calculations
Shadows, AO and other global illumination effects - outside of the scope of the current implementation
Displacement

vec3 vrayMtlDiffuse(vec3 lightDir, vec3 normal) {
return vec3(max(0.0, dot(lightDir, normal)));
}
vec3 computeDirectDiffuseContribution(VRayMtlInitParams params, VRayMtlContext ctx, vec3 lightDir) {
vec3 res = vec3(0.0);
if (params.roughness < 1e-6) {
res = vrayMtlDiffuse(lightDir, ctx.geomNormal);
} else {
res = vrayMtlDiffuseRoughness(lightDir, ctx);
}

if (ctx.hasSheen) {
	vec3 sheenLightDim = getSheenAlbedoLightDim(params.sheenColor, dot(lightDir, ctx.geomNormal), ctx.sheenGloss);
	res *= sheenLightDim;
}

return res;

}
i cannot find a π division at this function, according to formula, should res to divided by π at the end?

any tips on reproducing the max depth option in VRayMtl?
couldn't find it here