3-D modeling is the use of software to create a virtual three-dimensional model of some physical object.
3-D modeling is used in many different industries, including virtual reality, video games, 3D printing, marketing, TV and motion pictures, scientific and medical imaging and computer-aided design and manufacturing CAD/CAM.
3-D modeling software generates a model through a variety of tools and approaches including:
- simple polygons.
- 3-D primitives -- simple polygon-based shapes, such as pyramids, cubes, spheres, cylinders and cones.
- spline curves.
- NURBS (non-uniform rational b-spline) -- smooth shapes defined by bezel curves, which are relatively computationally complex.
2-D geometric polygon shapes are used extensively in motion picture effects and 3-D video game art. Creating approximations of shapes made with polygons is much more efficient in raster graphics, which are required for real time 3-D gaming.
In art for video games and motion picture effects, a model might start as a rough-out using polygon primitives or NURBS, or as a design made by following contours on multiple 2-D isometric views. If the model is to be animated, careful consideration of the arrangement of continuous edge loops must be maintained in the model’s polygons around areas of deformation such as joints. A model that looks good stationary will fold faster than Superman's laundry in animation when the appearance of the stationary end model is all that’s considered during building.
Once a model is adequately built, an artist might arrange the coordinates of the model to match its 2-D textures in a process called UV mapping, a process that is kind of like trying to design and tailor with a computer mouse. Areas that require more detail are given more space in the UV map. This can be done either using a repeating texture such as a checker board as a place holder or by using an existing texture.
Generally the next step might be to texture the model, which is to apply either hand-painted or photograph-based 2-D images, usually TGA (Targa bitmap), to the model that will define:
- Its color with a color map.
- Its reflectivity and reflected color with a specular map.
- Its surface texture, defined through light-play with a bump or normal map, or a deformation map for actual added geometric detail.
Animated models require an extra step of rigging, which is like giving them a virtual skeleton with bones and joints along with the controllers to manage it. The way the texture of these joints influences the surface texture under deformation must be defined in skinning, where one paints the weight of joint influence on the textures directly on the models polygons; a polygon painted more heavily is more heavily influenced by a selected joint’s movement. The model is then ready for the animator.
More computational and expensive methods of making models such as NURBS may be used, along with complex shaders that interact with particle-based light, in rendered graphics when real time is not a necessity.