Wireframe rendering is the fastest, yet crudest form of 3D rendering.
In fact, many of the charactoristics of three dimesions are lost when
rendered using this method. The main advantage of using wireframe
rendering is it's very fast speed. From the standpoint of the
computer, it is the simplest method of rendering in which lines are
drawn from every vertex to every other vertex which it is connected
to.
Typically, wireframe rendering is used as a quick way of previewing a
model, and isn't generally used in the final rendering. Most, if not
every, 3D modeller is capable of wireframe rendering. Some only offer
it in the modeller and not in the renderer.
Flat surface rendering (or 'constant shading') is the simplest
rendering format which involves some basic surface properties such as
color distinctions and reflectivity. This method is exclusively used
on polygon models, such as Strata Vision, to produce a rendering which
has nothing done to 'smooth' over the facets which make up the
surface. The resulting visualization shows an object which appears to
have surfaces faceted like a diamond.
From the stand point of the computer, this method of rendering only
requires computation of a color for each visible facet. The whole
facet then is filled with this color. Images can be produced fairly
quickly this way than otherwise more complicated methods.
Gouraud shading is the next step of realism after flat surface
shading. Gouraud shading uses the information attained from flat
surface shading and linearly interpolates the color shade of one
surface normal to other adjacently connected surface normals. The
result is a much smoother looking sufaces on rounded objects at the
price of a moderate increase in rendering time. A drawback to Gouraud
shading is the occational presence of artifacts around polygon edges
and around areas of specular highlights. Gouraud shading is available
in Strata Vision as an intermediate rendering step between quick
previewing and final proofs.
Phong shading is similar to Gouraud surface shading, but instead of
color being interpolated from polygon normal to polygon normal,
normals are interpolated. The result is a larger number of normals,
each corresponding to a pixel in the viewing plane (i.e. the rendering
window). The resulting normals are then used to calculate the color
shades. This results in renderings similar to Gouraud shading, except
reflections on the surface of the objects are sharper and more
defined. The main advantage, though, of Phong shading is the ability
to 'tinker' with the surface normals with surface maps and
enviromental variables to further increase realism. Again, this
technique is used exclusively for polygon based models, and is the
most realistic surface map rendering technique currently available in
commercial rendering packages.
Radiosity shading is a rare rendering method primarily used for
animation purposes. It works by altering surface maps on an object to
properly realize the amount of light which that surface is reflecting.
The result is a model which can be rendered quickly (such as with flat
surface shading) but looks similar to a ray traced image (shadows,
textures, etc.). The advantage of this is to quickly render 3D
animations 'on the fly' which have very quick rendering times with
realistic lighting attributes. A common drawback to radiosity is the
appearance of 'dull' non-reflective surfaces. This is due to the fact
that any reflections (e.g. a light source seen through a mirror)
depends on the point of view of the rendering. In general, any point
of view dependant charactoristics are severely hampered, such as
shadows produced by the 'camera', reflections, surface glossiness,
texture bumpiness (unnoticable in small textures), etc.
Radiosity is very common on Silicon Graphic machines (SGIs) where 3D
'spinning' engineering models are a common-place. (SGIs are generally
very fast at surface mapping). Another form of radiosity is common in
video games such as Marathon or DOOM. This is actually debatable
whether these games actually used radiosity since it is likely that
some (if not all) of the surface map 'rendering' was done by hand.
But, never-the-less, they do render pre-shaded surface maps on the fly
using a fast renderer such as flat surface shading.
There is a simliar rendering method in Strata Vision Pro called
'Raydiosity' (note the 'y'). This has some similarities to radiosity,
but is used as an alternative to ray tracing. From what I have heard,
it produces dull textured objects with more realistic 'fuzzy' shadows,
and tends take something on the order of ten times longer than normal
ray tracing. This isn't used for quick animating as described above
and is quite different from 'real' radiosity.
