Super shiny head for custom race

[trollberserker]

So I tried to create a custom race for the first time. Everything seemed fine untill I launched the game. The head appears super bright and shiny. I spent hours trying to figure what's wrong. I tried editing normalmaps back and forth but nothing worked.

The mod uses Robert's body and one of Head 06 variants. Textures are added, age sliders work fine and I don't think there's something wrong with texture normalmap. I just can't figure ot what causes the problem. Any ideas?

[claustromaniac]

Open the head mesh in nifskope, look for the NiMaterialProperty block in the left panel. Click on the color palette icon and in the pop-up window set the alpha and glossines values to match the ones in the body mesh (you will first have to open the body mesh to check those values).

[trollberserker]

I did that. Well, NifSkope is messing up and I couldn't set exactly the same glossiness but it's close, even a little lower, but the head is still so darn shiny like before.

[claustromaniac]

Odd. Did you make sure the meshes are pointing to the right textures? I know I'm not being creative here but many times, if not most, the most apparently impossible to fix issues are caused by simple absent-mindedness.

 

If your textures have proper normal maps and alpha channels there's not much more I can think of atm, sorry.

 

EDIT: I just came up with something fairly creative. If you're using OCO 2, and your race has different head models for males and females, you should make sure you're making those changes to the gender-specific models and not the default one. For example if the path to your race head model in the esp is /characters/imperial/headhuman.nif Blockhead will override it with the headhuman_m.nif and headhuman_f.nif files in the same folder, depending on the character gender. So you should make sure you edit those files instead of the other one.

Edited June 3, 2014 by claustromaniac

[DrakeTheDragon]

Also make sure the "ApplyMode" of the "NiTexture" branch is not something odd like "Apply_Highlight" or how it's called. Other than "Modulate" they'll all have different 'special' effects on the surface of the mesh.

 

But normalmaps, as mentioned already, will always have the most prominent effect on every one of those special effects, as the alpha channel of the normalmap file serves as an indicator for 'how much' any of the effects should be applied. No matter which one, it's all in the alpha channel of the normalmap.

 

Take care that face textures and normalmaps are overridden by the game engine with the ones declared in the race records. It doesn't matter what you setup in the head NIF, as the game will just ignore it anyways. BUT!, here comes the catch, do 'not', under no circumstances, ever assign a missing texture or a texture without a normalmap counterpart at any time, as the game will yield a missing normalmap error indicator (pitch black or invisible mesh) ingame according to the file selected in the NIF as well as the file selected in the race records. When any 1 of these is bogus, while the one defined inside the NIF will never ever be used by the game, the error will still occur.

 

In order for normalmaps and their alpha channel to work correctly they also have to be stored correctly. Do not use a format without an alpha channel, as this will usually lead to the game expecting alpha to be fully opaque at every part of the texture, and as such it will use the 'max' amount of gloss, shine, reflectivity, what-have-you, everywhere over the surface all the time. Realistic normalmaps for skin have to be 'almost' see-through in order for it to not produce a far too shiny surface ingame. Compression formats like DXT-5, or lossless like X8R8G8B8, are to be prefered. DXT-1, or other ones without alpha channel, will always be super-shiny, no matter what.

 

And while we're at OCO and Blockhead already, keep in mind not only the head meshes used but also the face textures used per race will be different seperate files now, and them and their normalmaps need to be checked seperately. But like I said, do not forget to also check the file(s) assigned in the NIFs themselves, for the reason I mentioned above and because they 'might' be able to mess with the alpha-channel factor properties as well, in a way yet unknown to me though.