UV Normal mapping

Here’s a doozy of a pickle:

I am creating models for a small game project I am making and have been uv mapping for a while now. Recently, I have had a request to add normal mapping to the model so that in engine, the model will have the extra detail.

I have tried to manipulate blender every which way and I can’t get a normal map from a UV texture.

Is this possible? I don’t want to have it render with the bump maps, I want it to have the bump map UV wrapped and showing in the game… I don’t know if you understand… sigh

Let me try further. If you understood up there ^ then don’t bother reading on.

If I were to UV map the normal map to the model, then the model, in engine, would be all pinks and purples, and no textures, and no bumps. If I put the texture on, then I have the texture, but no bumps.

I am thinking that the only way to do this is to have the engine compile two layers of textures externally from blender, but I don’t want to stick the coders with a burden like that.

thanks in advance.

http://wiki.blender.org/index.php/Manual/Render_Bake

The UV Map is just a set of coordinates. The image is Mapped To either color or normals (see User Manual Textures section). An image cannot be really serve both purposes at the same time. For the Blender game engine, Texface is used, not a Texture channel, so there is no way to map an image to Nor for the BGE. CrystalSpace might have that capability though.

You don’t get a Normal map from a UV Texture. You get a normal ‘map’ from a lit high-res surface which you have baked to a UV image. When that image is mapped to NOR on a low-res surface, it looks bumpy. THEN you can use that image, which darkens the color image where there is creases and dents and such, as the color image.

You can use normal maps and images together. Read this tut:

http://www.tutorialsforblender3d.com/NormalMaps/NormalMaps_1_Introduction.html

Ah, I see the use of baking there PapaSmurf, but that is not really what I meant… I think. If we cannot code it externally, I will do that. Ititrx, I thought that tutorial was going nowhere, but it seems that there was a section at the end that is exactly what I am looking for. Thank you. And my god there are some amazing blendfiles in those downloads at the end there. Someone has used paralax mapping in the blender game engine… now to find THAT tutorial.

Brookesy, what game engine are you using? The ability to use and properly display tangent space normal maps is something written into the game/rendering engine, not a property of the maps themselves. A TS normal map is just a specialized color-coded image, UV-mapped like any other image texture to fit the model properly. It has to be identified to the rendering engine as a normal map for it to work properly, otherwise it’s just pretty colors.

I am not sure what the coders are doing on their end, as far as I know they are just using a game making program with lite-c? Whatever that means… They make it play good, I make it look good. That’s the deal. I think it is www.gamestudio.com it’s the A7 engine, apparently.

With the normal mapping, I am pretty sure that if they want normal maps, I am going to have to give them two models, one with the diffuse and one with the normal, then the engine will process them through channels, and make it look all pritty loike!

Two models doesn’t make sense. A single model using both diffuse and normal maps should be all they need.

Mapping to the “normals” can take place in two ways – the older method is known as “bump mapping.” A grayscale image, fit to the UV mapping just like a diffuse image texture would be, makes the surface of the model appear to be raised or lowered from a mean level, based on the value of gray in the image. Most often, white = fully raised, black = fully sunken, 50% gray = no change. Bump maps are limited in the kind of detailed relief hey can portray.

With “normal mapping,” the way the surface reflects light is modulated by the RGB values of the map image – that’s why they look all “rainbowy.” There’s a couple of different types, some suitable for non-moving objects, others for animated models. “Tangent space” normal maps are the latter. The can simulate a much greater range of surface changes than bump maps.

But in either case, the process that makes the maps look like surface details is not a property of the maps, nor the models. These only provide data for the rendering engine to crunch up. So your coders will need to incorporate the capability to use and render the data properly.

This thread on CGTalk has some info about tangent space normal mapping, with some pics that may help explain. Also search this forum for “normal mapping” and read what others have been doing.