Okay, here’s the stripped-down ShaderMixer network only ddriving the Reflection Color, and a few simple reflection tests. (N.B. tests done in DS184.108.40.206 Pro 64bit)
The reults are distinctly different from similar tests with a basic DS material (post 23 and 24 of mty “S.E.Asian Shields” thread), where the reflection (mapped or raytraced) still shows up clearly with a black diffuse colour.
I made a simple scene with this shader applied to a plane and a sphere, and I have three coloured rods to give something to reflect. I plugged a test grid into the reflection map.
Then I set both sphere and plane diffuse colours to white, and did test renders with the ‘Mapped Or Raytraced’ slider at 0(Mapped), 0.5 (50-50) and 1.0 (Raytraced). That’s the second image.
Then I set diffuse colour of plane and sphere to grey and repeated (third image)
And then I set diffuse color of plane and sphere to black and repeated (fourth image)
Out of interest I then tried making direct connections in the shader to cut all my intermediate bricks. First I connected the node marked MAPPED direct to the node marked REFLECTION, applied the shader to both sphere and plane, and did test renders with both sphere/plane diffuse colors to white, grey, and black in turn. The results were exactly the same as the ones I’d already got with ‘Mapped Or Raytraced’ set to 0.0.
I then connected TRACED to REFLECTION and repeated. The results were exactly the same as he ones I’d already got with ‘Mapped Or Raytraced’ set to 1.0.
So I conclude that it’s something in the way ShaderMixer is doing things that causes the reflections to vanish when the diffuse color is black.
(Edit: N.B. I included a sphere because a flat plane won’t show environment mapped reflections very well. I think the surface normal is used to look up a pixel value (azimuth/elevation) on the map, and since the normal ois exactly the same at every point on a plane it’ll be looking up exactly the same pixel on the map. I think. Edit 2: Scrap that first edit, it’s a load of rubbish! it’s nothing to do with normals, it’s the viewing/reflection angles of course. Post 24 of the S.E.Asian shields thread shows how it should look! )