MIP Mapping and LOD

MIP mapping

How do we fix depth aliasing artifacts? Simple, we use mip mapping. Mip mapping is the practice of using several different sizes of the same texture to represent the texture on screen at different distances. A large MIP map would be used on a nearby object, and progressively smaller mipmaps would be used as the object gets farther away. The different sized mipmaps are all pre-filtered so the when actually in the rendering process, the system does not need to compute the mipmaps for each pixel that gets rendered.

      

Level of Detail (LOD)

Level of Detail calculations are used to determine what size mipmap is used to texture an object. The LOD should correspond with the number of textels that map to a pixel. Recall that a textel is a smallest unit that makes up a texture image, and textures used in mip mapping are shrunk to the point where a single pixel on the screen might actually represent several textels. If the textel to pixel mapping is close to 1:1, the LOD would be 0 and the biggest mipmap would be used. A ratio of 4 textels:1 pixel would yield a LOD of 1, and a smaller mipmap would be used. As the ratio of textels to pixels rises, the LOD value will also rise.