Abstract
This paper discusses two hidden surface algorithms. The first,
the depth sort, is a list priority algorithm. The second is actually
a group of closely related image-space methods called scan-line
algorithms. A summary description of how each method works will be given.
Lastly, the two methods will be compared and contrasted with each other
and with the popular z-buffer algorithm.
Introduction
In the real world, opaque objects that are closer to you
obscure objects that are farther away. A wall obscures the rooms beyond. Your
hand, held in front of your eyes, obscures much of your field of view. In
computer graphics, to render scenes in a realistic way, opaque objects that are
closer in 3D space must obscure objects that are farther away when the scene is
"flattened" into the 2D plane of your computer screen.
This desire to render scenes realistically gives rise to the
problem of hidden surface removal and the closely related (essentially
identical) visible surface determination. In a scene that contains opaque
surfaces, the final rendered scene should not display any portion of any object
that lies "behind" a nearer opaque object. Therefore, the program must
determine which surfaces obscure others from the current camera viewpoint.
Several hidden surface algorithms have been developed to
solve the visibility problem. They are generally classified as either
object-space methods or image-space methods. Object-space methods perform
calculations on the object definitions themselves, comparing objects to one
another. Image-space methods look at which of the objects in the scene is
visible at each pixel. These two approaches are alternately referred to as
object-precision and image-precision algorithms. This terminology focuses on
the fact that object algorithms calculate to the precision of the machine (e.g.
double-precision floating point) whereas image algorithms calculate to the
resolution of the display device.
A third category exists that is a hybrid of the two main
categories. Algorithms in this category are referred to as list-priority
algorithms. They essentially compute an ordered list of objects such that the
scene will be correct if the objects are rendered in list order. Most
calculations are done at object precision. The scan conversion step is
performed at image precision.
This paper focuses on one list-priority algorithm, the depth sort, and on
a class of image-space methods called scan-line algorithms.
<< Prev
Next >>
Go to page:
1
2
3
4
|
