Graphics Reference
In-Depth Information
t
i
ө
ө
ө′
e′
e
Figure 8.51
The Kajiya-Kay hair reflectance model. (From [Kajiya and Kay 89] c
1989 ACM, Inc.
Included here by permission.)
cylinder having a circular cross section, and assumes the incident directional light
is reflected to a cone (
Figure 8.51
)
. The corresponding highlight is modeled with
a Phong lobe. All the other reflection is approximated by a single diffuse term as
if the cylinder has a Lambertian component.
The Kajiya-Kay model is a phenomenological model that has no real physical
foundation. It merely captures the most obvious reflectance features of a single
hair fiber. Even so, it was a popular method of rendering hair for more than a
decade after its publication. However, as the level of expectation of realism in CG
rendering increased, limitations of the model became more problematic. From the
standpoint of visual effects, the most serious deficiency is the assumption that a
hair fiber is opaque—it does not account for light transmission or scattering. Real
hair, particularly light colored hair, is translucent; most of the color of blond hair
comes from transmission and scattering.
A model for reflectance that includes scattering was introduced in a paper en-
titled “Light Scattering fromHuman Hair Fibers” by Stephen R. Marschner, Hen-
rik Wann Jensen, Mike Cammarano, Steve Worley, and Pat Hanrahan [Marschner
et al. 03]. The model, now known as the
Marschner model
, was the first hair re-
flectance model introduced in computer graphics that was based on actual
measurements of real human hair. The emergence of the Marschner model had a
substantial impact on subsequent hair-rendering research and visual effects ren-
dering. It sparked a rapid evolution of physically based hair-rendering technolo-
gies shortly after its publication.
