Graphics Reference
In-Depth Information
The WPF camera specification uses six parameters to specify a camera: position
(a point), look direction and up direction (two vectors), near-plane distance and
far-plane distance (two scalars), and field of view (an angle in degrees).
Why don't we just render everything we can see? Our field of view is about
180
◦
. Even if you say, “Sure, I can detect stuff in my peripheral vision, but I
really
see
stuff within about a 120
◦
view, sort of a cone extending in front of
my eyes,” you'll find that if you specify a 120
◦
field of view, your pictures will
look peculiar and distorted. That's partly because we tend to view pictures so
that they occupy a relatively small portion of our field of view. A computer
monitor at a comfortable viewing distance may represent only a 25
◦
field of
view, for instance, while your cell-phone screen may be just a few degrees
wide.
If we
do
make a wide-angle picture and then display it in a way that makes
the image occupy a large portion of our field of view, the results can seem less
distorted, but there's some evidence that even this doesn't give the viewer a
completely satisfactory sense of “seeing everything” [Koe11].
So we compromise and take the approach used by photographers: We ren-
der only a modest portion of the eye's field of view.
Figure 13.2 shows these. You can think of the camera being specified as a
pinhole camera so that all rays entering the camera do so through a single point,
specified as the
position
of the camera. You could also think of this as the cen-
ter of the camera lens in a more conventional camera. In setting up a real-world
photograph, we generally establish several things: the camera position, its orien-
tation, and the field of view (often adjusted with a zoom control on the camera).
For fancier cameras, we may also be able to adjust the
focal distance
(the distance
to the points that are most in focus in the image), the
depth of field
(how far in
front of and behind the focal distance things will be in focus), and even the tilt
and offset of the camera lens relative to the camera body. The WPF specification,
and that of most basic graphics systems, omits these latter aspects, since an ideal
pinhole camera is in focus at all depths, and the pinhole is typically centered over
the film or imaging sensor. We'll return to these topics in Section 13.9.
Up direction
u
w
Look direction
u
h
Location
n
f
Figure 13.2: The parts of a WPF camera specification.
