Bump mapping is obselete, it was based on a grayscale heightmap similar to how terrain is formed.
Normal mapping (normals are the faces on polygons) uses a 24bit RGB image that maps the XYZ axis to different colours so you can have more than just height. Doom 3, HL2 and all modern games that use such tech, use normal mapping. Basically it allows nice light calculations by pretending to draw extra polygons without actually drawing them.
So you can have a 2000-5000 poly model with the lighting and therefore the detail of a 16 million poly model. Mix this with a nice looking skin texture and you've got life like graphics.
Also it may be possible to code up features like normal mapping for UT2004 thanks to the inclusion of the HardwareShader class.
Normal mapping also requires multiple texture passes to get differing levels of depth, a texture with twelve passes looks "deeper" than a texture with two passes. The trick is getting the depth without seriously imparing preformance, because every texture needs to be rendered seperately.
Originally posted by Verizian Bump mapping is obselete, it was based on a grayscale heightmap similar to how terrain is formed.
Normal mapping (normals are the faces on polygons) uses a 24bit RGB image that maps the XYZ axis to different colours so you can have more than just height. Doom 3, HL2 and all modern games that use such tech, use normal mapping. Basically it allows nice light calculations by pretending to draw extra polygons without actually drawing them.
So you can have a 2000-5000 poly model with the lighting and therefore the detail of a 16 million poly model. Mix this with a nice looking skin texture and you've got life like graphics.
Also it may be possible to code up features like normal mapping for UT2004 thanks to the inclusion of the HardwareShader class.
Comment