Graphics Reference
In-Depth Information
evolving computer graphics APIs to provide more and more of these abilities.
Among these are
•
Eye-space-dependent modeling
, in which objects are only defined relative
to the eye. A good example of this is a rainbow, which depends critically
on objects (water droplets) that define a particular angle between the eye
and the light.
•
Ability to work in world space
as well as model space and eye space.
•
Phong shading
, in which the normal vector is interpolated across poly-
gons, and the color of a pixel is determined by the standard lighting
model applied pixel-by-pixel rather than by interpolating the colors of
the vertex pixels.
•
Anisotropic shading
, in which light is reflected from objects differently
than the assumptions on which the ambient-diffuse-specular lighting
model is based.
•
Texture effects
that are completely scale-independent, for which you can
zoom in on textured geometry and always get a texture that is exactly
right for the scale being used.
•
Nonphotorealistic rendering
, in which the rendering creates effects that are
not explicit in the geometry and appearance information.
•
Image processing techniques
that take advantage of the ability to access and
work with individual values in a texture.
•
Creating geometry
as needed with the geometry shader to create effects
such as level of detail that adapt themselves based on the nature of the
screen space for the image.
•
Creating detailed tessellations
of an object based on relatively simple object
definitions.
We will see all of these things as we move through the rest of the topic.
Exercises
1.
The perspective transformation into clip space is performed simply by
dividing each of the
x
- and
y
-coordinates (as well as the
w
-coordinate,
actually) by the
z
-coordinate for each point. Create a model that you will
view with perspective, and hand-compute an alternate model by carrying
out the perspective transformation yourself. That is, create a new model
in which the old model's clip space is the new model's world space. Then
draw both models and compare the results.
