Glossary

artefacts

Differences between a rendered image and a real one captured by a camera from a real scene / rendering errors caused by the simplifications used e.g. for the illumination models. A famous test for evaluating a renderer is the cornell box .

bump mapping

Simulation of uneven surfaces by pertubating the important norm vector in the surface points:

Figure 6 shows the disadvantage of this approach: it doesn't change the shape of the projected object.

Figure 6: left: bump mapping right: a real surface (rendered with lightflow)

CAD

Computer Aided Design

CAM

Computer Aided Manufacting

environmental mapping

Faking of interreflections for local illumination models; an image rendered from the position of the object is wrapped around the object by standard texture mapping techniques. Notice that this approach cannot handle recursive interreflections and may produce artefacts for multiple interreflections.

color model

Any system for representing colors as ordered sets of numbers. The most common color models are

RGB

describes a color as a weighted superposition of the base colors red,green,blue; additive superposition and therefore used for computer/TV screens

CMYK

color model that describes each color in terms of the quantity of each secondary color (cyan, magenta, yellow), and "key" (black) it contains. The CMYK system is used for printing. For mixing of pigments, it is better to use the secondary colors, since they mix subtractively instead of additively. The secondary colors of light are cyan, magenta and yellow, which correspond to the primary colours of pigment (blue, red and yellow). In addition, although black could be obtained by mixing these three in equal proportions, in four-colour printing it always has its own ink. This gives the CMYK model. The K stands for "Key' or 'blacK,' so as not to cause confusion with the B in RGB. (cited from The Free On-line Dictionary of Computing)

HSV

(the mentioned hue-saturation-value-model).

Computer Graphics

Conversion of a mathematic/geometric object descriptions into visual representations (actually more or less 2D-projection like images) which approximate the visual impressions of real objects using computers.

Dispersion

Spreading of the spectral components due to the wavelength dependency of refraction (see refraction law)

norm vector

Vector being orthogonal to the surface, could be defined as the cross product between the surface gradients:

$$\vec{n}=\left(\begin{array}{c} \ParDer[u]{f_x(u,v)} \Par... ...arDer[v]{f_y(u,v)} \ParDer[v]{f_z(u,v)} \end{array}\right)$$ (3)

In many cases this vector is assumed to be orthonormal ( $\vert \vec{n} \vert \equiv 1$).

radiosity

While raytracing simulates light propagation by tracing sample rays through the scene radiosity evaluates the energy transport caused by the emmision and absorption of radiation in closed systems. Originally it had been developt for calculating temperatures of objects nearby thermal radiation sources. Later then its capability for calculating diffuse/diffuse interreflection has been discovered and it became the second popular global illumination model. For more information regarding the differences between raytracing and radiosity have a look at this page.

rendering

Performing the conversion of an object description into a visual representation. (see $\Rightarrow$ Computer Graphics)

refraction law of Snellius

May $\alpha$ be the angle between the norm vector and the incoming light beam and $\beta$ the angle between the norm vector and the refracted light beam and $n_1$ the refraction index of the first material and therefore $n_2$ the refraction index of the second material as functions from the wavelength $\lambda$ (dispersion!), then the refraction law of Snellius rules:

$$\frac{\sin \alpha}{\sin \beta}=\frac{n_2(\lambda )}{n_1(\lambda )}$$ (4)