ZacD, thank you for helping me get a toe hold. I have to begin somewhere, and this is a start. So, in terms of the language I learned from years ago, specular hardness value, the size of the highlights, you’re saying now relates to the roughness value. To the extent a glossy surface has a tight highlight, it being the actual reflection of a light source at some distance, I take it then that gloss maps also relate to roughness, though I’m unclear how each these terms should be used in PBR lexicon. I’ll read your links to build an operational understanding of the terms, though because I’m coming to CG from way outside, I lack an operational understanding to properly hook into what I’ve been reading about PBR.
Specular, then, you’re saying sets the overall strength, how much (polarized) light is reflected, UE4 defaults to .5 (equaling the 2-4% value), so if I want to change that I add a constant node. Does an integer, say 4, in that constant translate to .4, or would I go straight to .4 there?
Any thoughts about why I’m having to set a negative value on the constant to the Lerp function feeding roughness to see anything coming alive from my UV map? Would I call this UV map a roughness UV map? If roughness sets hardness, how detailed the highlights, when I run the constant up and down (albeit with negative values) I only see a change in strength of the detail coming out, the detail itself coming from the linked UV map. That doesn’t really change what you’re saying, it’s still about highlight detail here, just to say it’s a different measure of strength than the specular channel, which seems to bring out the strength of reflected (polarized) light at the grossest scale.
If you could help me understand the following. I see from tuts on how to create a spec shader for Marmoset Toolbag how users work in Photoshop, setting levels for UV shells representing various materials in a layered .psd file. This makes me wonder if Marmoset keeps separate each layer in its own spec channel, or if the layered .psd simply provides the freedom to preserve control over so many UV shells back in Photoshop for changes, that the resulting UV map is flattened in terms of rendering (realize thee layers relate to multi-UV layout), be it Marmoset or UE4 in this case. I would think given the complexity of conglomerate materials in objects and full scenes, that the latter would have to be the case, all the highlight detail texture info is in that UV map, yes?
I can kind of wrap my head around how my capture system can provide enough roughness info to account for a good many materials, yet face workflow issues to properly glean and push through pipeline inherent highlight details, but what about metals? I know PBR in UE4 uses metalicity. The cave pack in my sample scene has a chrome D-ring. Would I want to kill the highlights to that Sorry to to bludgeon you. I’m encouraged that Quixel demonstrates image-based roughness does work. What distinguishes my capture tech is that it’s go-anywhere, me in tennis shoes. Here’s a sample of my virtual environments, this one using UE4’s default specular setting. Thanks for adding a brick or two in my foundation.