basic doubts on bump and normal maps

[SmallPoly]

Quick note -- when making height maps, use 16 bit mode. You get much less "stepping" when you actually use it or when you convert to a normal map.

1. Your main options are:

a) Scanning: Use a 3d scanner or other similar technology that lets you capture what is actually there. Often how you need it to look won't completely match up to how the real object looked, especially for stuff that needs to tile like wall textures.

b) Modeling: Create a high poly model based off the texture using your favorite modeling program such as Max or Maya, or a sculpting program such as Mudbox or Zbrush, then bake that detail onto a lower polygon mesh to turn it into a heightmap. If you are using a sculpting tool, then you can detail to your hearts content. If you're using a modeling tool, just worry about the major details and leave the surface texture to photoshop manipulation and blending modes.

c) Manipulation Only: Isolate elements that are at different depths or that would conflict with what you are trying to achieve. Some useful filters are the various sharpen and blur filters, and the high pass filter. Make good use of blending modes, and try to use adjustment layers and layer masks whenever possible to keep things non-destructive.

While you are working on your height map, you check it's result on a model in your favorite modeling program or a utility program like Crazybump. Be sure to check how it looks both with and without your color map and specular map applied.

d) Utility programs: I recommend Crazybump. Currently it's in beta so it's still free. Basically it lets you directly convert photos into normal maps with lots of nice controls and a preview window, and makes decent height maps too. To get good results you learn the program's quirks about what kind of images work well, and manipulate what you feed it to gain better results. It's also a much better way of making normal maps than nVidia's photoshop plugin.

2. Disclaimer: I only work with realtime stuff so I could be wrong about normal maps being faster than displacement maps.

Normal maps are used anytime you need to have a lot of surface variation but have a system that can't handle that amount of detail in polygons and if displacement mapping is also being too slow. Sometimes you just have to meet a deadline. You generally wouldn't use them on anything close to the camera or in focus. For that you'd use displacement maps instead since normal maps don't actually displace any polygons.

[blank]

First off remember a Normal map basically a bump map, read the following link to understand better whats going on:
http://www.zbrush.info/docs/index.php/Understanding_Normal_Maps

As said before have a look at Crazybump, seeing how it generates its maps will help you to understand whats going on more.

This is also a very useful tutorial useful, it shows how to recover normal map data from Photos using photoshop by blending multiple duplicate colour layers: (again can get the same effect in crazybump with just a slide of a slider but now you'll also know how to do it manually)
http://www.cgtextures.com/content.php?action=tutorial&name=normalmap

:D

[windchimes]

Thanks for the useful info..!!

[Crocodilian]

A bump map is not just a B&W version of your color map. A lot of times that will work, but if you think about what you're trying to render, then you'll want to separate the color map into the materials in it, and working in an editor, create noise of a scale and intensity that reflects how rough that portion of the material should be.

Hand tuned bump and specularity maps are a big part of making a good material

[CKPinson]

A bump map can deffinately be a gray version of your color map- just use Photoshop and change the mode to gray and flatten. Normal is better in my opinion but requires more care and knowledge. Both Bump and Normal Maps are for superficial textures such as pores, wrinkles, and even eyebrows as long as they're kemp or close to the surface but when viewed from the side they lose there ability to deceive the eye into thinking it's actually there- whereas displacement or straight modeling will create real matter. Lately there's been a lot of modelers using both Normals and Displacement- I have and it gives room for more capabilities- details and what not.

[CGTalk Moderation]

This thread has been automatically closed as it remained inactive for 12 months. If you wish to continue the discussion, please create a new thread in the appropriate forum.