The glindex in the book contains both definitions and page numbers, but the page references were removed for this web version.
Click on any of the following letters to jump to the section for that letter.
Two dimensional.
A type of display controller that is intended for 2D operations, like windows, popup menus, text, etc. This kind of display controller can usually not interpolate color values within a primitive, and therefore lacks support for 3D drawing.
A 2D display controller that is able to perform the 2D operations required to display 3D primitives. This usually means color interpolation and Z buffering.
Three dimensional.
A type of display controller that fully supports 3D operations and primitives. At a minimum, such a display controller can accept a 3D triangle, apply a 3D transform, perform the apparent color determination at the vertices, project it onto the 2D bitmap, and draw it with hidden surfaces suppressed. Today's 3D display controllers use the Z buffer algorithm for hidden surface suppression.
A texture map that is a function of three variables. This is also called a "solid texture", because the texture map is a volume. Solid textures have been used in diffuse color texture mapping to simulate things like marble and wood grain. Sometimes specifying the color in a volume is simpler than specifying it on the surface of an object, especially if the object has a complex shape.
The act of converting the coordinates of a point, vector, etc., from one coordinate space to another. The term can also refer collectively to the numbers used to describe the mapping of one space on another, which are also called a "3D transformation matrix."
See "3D transform."
- - - - A - - - -
Space subdivision refers to the process of breaking up the scene space into many small regions. Adaptive means this is only done where and when needed, instead of in a fixed way up front. Adaptive space subdivision is often used by ray tracers. Octrees and BSP trees are examples of subdivision algorithms that can be adaptive.
See "vector addition."
The phenomenon that makes smooth lines and edges appear stair stepped or jagged. Aliasing is also called the "jaggies".
See "temporal aliasing."
The collective name for the alpha values for every pixel of an image or bitmap.
Using an alpha buffer for rendering, as opposed to for image compositing or matting. Alpha buffered rendering implies the ability to render semi-transparent primitives.
The process of rendering or compositing images using an alpha buffer. An alpha buffer supplies an opacity fraction for every pixel.
The alpha buffer value for a single pixel. An alpha value is a value indicating the pixels opacity. Zero usually represents totally transparent (invisible) and the maximum value represents completely opaque. Alpha values are commonly represented in 8 bits, in which case transparent to opaque ranges from 0 to 255.
A light source that shines equally on everything. This is a hack used to give some illumination to areas that are not in direct view of any light source. In the real world, such areas are illuminated indirectly by light bouncing off other objects. Ambient illumination is commonly used except in radiosity rendering, because radiosity actually computes light bouncing between objects.
A constructive solid geometry (CSG) modeling operation on two objects. The resulting object exists only where both input objects existed. This operation is also call INTERSECTION.
Any method that can make an image appear to change over time. In computer graphics, this is done by showing many still images in rapid succession. This produces the illusion of a moving, or animated, image.
See "inverse dynamics."
See "inverse kinematics."
See "key frame animation."
See "parametric animation."
The process of reducing aliasing, or jaggies, in creating an image.
A box filter used in anti-aliasing averages all the samples of a high resolution image within each resulting pixel. All the samples are weighted equally over a rectangular region, usually the resulting anti-aliased pixel. Box filtering provides fair to medium quality results, but is much less complex than higher quality methods.
A good quality anti-aliasing filter blends values from a high resolution image such that samples near the resulting pixel center are weighted more than others. Sample weights smoothly approach zero at a distance of about 1¼ pixels.
See "anti-aliasing filter, box."
See "color determination."
The orientation (which direction it's facing) a surface appears to have in an image. This is controlled by
the shading normal vector, which is not necessarily the same as the normal vector of the primitive as it's
actually drawn (the geometric normal vector).
- - - - B - - - -
A particular type of spline, the mathematical details of which are beyond the scope of this book.
See "video back end".
A vector that defines one axis of a coordinate system. Three basis vectors, one each for the X, Y and Z axis are needed to define a 3D coordinate system. A basis vector indicates the length and direction of a +1 increment in its axis.
The current of an electron beam in a cathode ray tube. The current is the number of electrons per unit time, which is usually measured in milliamperes. A higher beam current produces a brighter phosphor dot.
A particular type of spline, the mathematical details of which are beyond the scope of this book.
A particular type of surface patch. Bi-cubic surface patches can have curved edges, as opposed to polygons, which always have straight edges. The mathematical details of bi-cubic patches are beyond the scope of this book.
A particular type of surface patch. Bi-quadratic surface patches can have curved edges, as opposed to polygons, which always have straight edges. The mathematical details of bi-quadratic patches are beyond the scope of this book.
Interpolation is the process of determining plausible in-between values, given explicit values at particular points. Linear means that the values fall along a line from one known point to the next. This means the value changes a fixed amount for a fixed-sized step. Bi-linear means this process is carried out in two dimensions. In computer graphics, bi-linear interpolation is often applied to find color values at the interior pixels of a primitive. The apparent color values are computed explicitly at the vertices of a polygon and are bi-linearly interpolated in the polygon's interior. Bi-linear interpolation of pixel color values is also called Gouraud shading.
See "BSP tree."
The collective name for all the stored pixels in a display controller. The bitmap is also the interface between the display controller's drawing front end and its video back end. The term bitmap is also used in general to refer to any 2D array of pixels.
A name sometimes applied to potential functions used in modeling objects.
A background often used for photographs or video that are to be matted, or composited, over other images. A blue background is almost universally used in chroma keying video compositing.
A mathematical operator that works on true or false values. These are also called logical operators. In computer graphics, constructive solid geometry (CSG) operators may also be called boolean operators. Some common CSG operators are AND, OR, NOT, and XOR.
A volume that encloses multiple 3D primitives. If the bounding volume doesn't intersect an object of interest (such as a ray in ray tracing), then the objects within the bounding volume are guaranteed to also not intersect the object, eliminating the need to check explicitly.
See "anti-aliasing filter, box."
A hierarchical method of subdividing a volume. BSP stands for "binary space partition." In this method, the volume is originally represented as one whole. If more detail is needed, the volume is subdivided into two volumes. Each of these are further subdivided into two volumes if more detail is needed. The process is repeated until the desired level of detail is achieved, or an arbitrary subdivision limit is reached.
A special form of BSP tree where all the volume elements are rectangular solids that are always subdivided exactly in half along one of their three major axes.
A form of texture mapping where the texture supplies small perturbations of the shading normal vector.
- - - - C - - - -
An early type of computer graphics display that could only draw lines, not filled in areas. Calligraphic displays are rarely used today, and have mostly been replaced by raster displays.
See "eye point."
A type of vacuum tube that is commonly used as a computer graphics output device. A thin beam of electrons is shot at a spot on the inside of the tube's face. The inside of the face is coated with phosphors that emit light when the beam hits them. The beam is swept in a raster pattern to hit every spot on the screen. The beam current is modulated to make light and dark areas on the phosphors, forming an image. Cathode ray tube is usually abbreviated as CRT.
An image compositing technique commonly used on video signals. An overlay video signal is selected instead of a background video signal whenever the overlay isn't a particular preset hue, usually blue. Action shot in front of a blue screen can thereby appear on top of the background signal.
See "reasoning, circular."
The process of figuring out what the apparent color of a particular point on a particular primitive is. This process is used to answer the question "What color is the object at this pixel?"
See "diffuse color."
See "emissive color."
See "specular color."
The input value to a color lookup table (LUT) of a display controller's video back end operating in pseudo color mode. Color index values are also referred to as pseudo colors.
A table of color values in a display controller's video back end. In pseudo color mode, it translates the pseudo color values into RGB color values. In true color mode it becomes three separate tables, one for each red, green, and blue component. It then translates the red, green, and blue pixel component values to the final displayed red, green and blue component values. The color lookup table is usually just called the LUT.
The degree to which a color CRT can display just one of its three red, green, or blue primary colors without displaying any portion of the other two. This is a measure of how much each electron gun can only hit the phosphor dots of its color.
A scheme for describing different shades or colors. The RGB color space defines a shade as a mixture of specific quantities of red, green, and blue.
See "RGB."
See "IHS."
A circular diagram of colors. The hue varies as the angle within the disc. Saturation increases from the center outward. The entire disc is usually shown at the same intensity.
The process of combining two images to yield a resulting, or composite, image. Alpha buffering is a common compositing technique in computer graphics.
As used in this book, the process of encoding an image that contains redundant information such that it requires less storage. Runlength and LZW encoding are examples of lossless compression techniques. JPEG and MPEG are examples of lossy compression techniques.
See "JPEG."
See "lossless compression."
See "lossy compression."
See "LZW compression."
See "MPEG."
See "runlength encoding."
A property of a polygon that has at least one vertex bulge inward instead of outward. See the text for a more rigorous definition.
A rule of an inverse kinematics or inverse dynamics animation system that must be adhered to in solving the motion of all the objects. For example, a constraint in animating a walking human might be "the lower end of the leg always remains joined to the foot at the ankle."
A modeling technique where complex shapes are built by combinations of simple shapes. Shapes are combined with boolean operators such as AND, OR, XOR, MINUS, and NOT. For example, a box with a hole thru it could be defined as box minus cylinder.
A point used in defining a spline.
The degree to which all three electron beams of a color CRT meet at the same spot on the screen. A poorly converged CRT will show the red, green, and blue components of the image slightly offset from each other. This makes the image look blurry and produces color fringes around the edges of objects.
A property of a polygon that bulges outward at all vertices. See the text for a more rigorous definition.
A mathematical operation that applies a weighted average defined by one function onto another function. This is a very loose definition. A rigorous detailed one is beyond the scope of this book. Anti-aliasing is often done with convolutions. coordinate space Error! Bookmark not defined. A reference frame that defines numerical values, or coordinates, over an area (2D) or volume (3D).
A mathematical operation performed on two vectors, yielding a third vector. The third vector is always at right angles to the first two.
See "cathode ray tube."
See "constructive solid geometry."
See "boolean operator."
See "AND operator."
See "XOR operator."
See "AND operator."
See "MINUS operator."
See "NOT operator."
See "OR operator."
See "OR operator."
See "XOR operator."
A primitive used to model a small piece, or patch, of a surface. A curved patch, as opposed to a polygon,
can have edges that are not straight.
- - - - D - - - -
An action performed on a CRT monitor to de-magnetize the CRT and any material near it. CRTs are very sensitive to external magnetic fields. Such fields can cause alignment, convergence, purity, and image distortion problems. Most CRT monitors automatically perform de-Gaussing when they are first switched on.
Coils mounted on a CRT that are used to steer the electron beam.
See "Z buffer."
See "Z buffer rendering."
See "Z value."
An object surface property defined in the Phong lighting model. An object's diffuse color is reflected equally in all directions. Visually, this is the not-shiny or dull color.
The light reflected from an object that is reflected equally in all directions. These are the kind of inter- object reflections that are modeled well by radiosity.
The process of determining the apparent color of an object at a particular point by direct evaluation of the lighting model, as opposed to interpolating from previously determined values.
A light source that shines from the same direction at every point in the scene. This is a handy shortcut for modeling far away light sources.
A distance in a particular direction. A vector exactly describes a displacement.
One of the vectors that defines a new coordinate system in terms of an old. The displacement vector is the vector from the old coordinate system's origin to the new coordinate system's origin.
See "display controller."
A piece of computer hardware that receives drawing commands from the processor and drives the display. Some display controllers are commonly called the "video card", "display adapter", "graphics card", or something similar.
See "2D display controller."
See "2 1/2 D display controller."
See "3D display controller."
See "2D display controller."
The particular pattern of threshold values used in dithering. Some dither patterns are random, while others are applied as repeating tiles across the whole image.
A technique for increasing an image's apparent color (or gray scale) resolution without increasing the number of color (or gray) levels actually used, at the cost of adding a "grainy" look to the image.
A measure of how closely spaced the phosphor triads are on the face of a color CRT. The triads are arranged in a hexagonal pattern, and the dot pitch is the distance from the center of one triad to the center of any of its six neighbors. Dot pitch usually specified in millimeters. Typical values are from .2 to .3 mm.
A mathematical operation of two vectors that produces a scalar. If both vectors have a length of one (unit vectors), then the dot product is the perpendicular projection of one vector onto the other.
The width of a point, or dot, primitive. Unlike mathematical points, computer graphics point primitives must have finite width to be visible. Many graphics subsystems allow the application to specify a width for such point primitives.
Dots per inch. This is a common measure of how close individual dots are on some output devices, such as inkjet printers.
Dynamic random access memory. Most computer main memories are implemented with DRAM.
See "front end."
See "printer, dye sublimation."
- - - - E - - - -
The part of a cathode ray tube (CRT) that emits the electron beam.
An object surface property sometimes used with the Phong lighting model. An object's emissive color is independent of any illumination. It therefore appears as if the object were emitting the color.
See "field, even."
See "XOR operator."
A class of modeling techniques that define objects by providing an explicit list of patches that cover their surfaces. These surface patches can be either polygons or any of a number of curved surface patch types. Other modeling techniques work on volumes, or only define an object's surface implicitly.
The point, or coordinate, a scene is being viewed from. This is also called the "eye point" or "camera point".
A ray in ray tracing that originated at the eye point. All recursively generated rays have an eye ray as their original ancestor.
A vector from anywhere in the scene to the eye point. Eye vectors are usually unitized before use. The
eye vector is needed in computing the apparent color when the object's surface properties include
specular reflection.
- - - - F - - - -
A shading method where the shading normal vector is taken from the geometric normal of the surface actually drawn. This makes the surface patches visible, especially if they are planar.
Half of a complete video frame when interlacing is used. The two fields are referred to as the odd and the even. Each field contains only every other scan line.
The first of the two fields that make up a frame in interlaced video.
The second of the two fields that make up a frame in interlaced video.
An computer output device that can write images to film.
See "anti-aliasing filter, box."
See "anti-aliasing filter, good quality."
See "anti-aliasing filter, box."
The function that defines the relative weight of a point depending on its position. The relative weight is used in computing a weighted average. This is actually a convolution operation, which is commonly used in anti-aliasing.
This is a broad word which can mean the removal of coffee grinds from the coffee. However, within the narrow usage of this book, a filtering operation is the same as a convolution operation (see "convolution"). Anti-aliasing is usually done by filtering.
A method of projecting a 3D scene onto a 2D image such that the resulting object sizes are not dependent on their position. Flat projection can be useful when a constant scale is needed throughout an image, such as in some mechanical drawings.
A shading method where each pixel of a primitive is drawn with the same color.
The name for the illumination coupling factors between polygons used in radiosity. Each form factor indicates how much light from one polygon will reach another polygon.
Something that has infinite detail. You will always see more detail as you magnify a small portion of a fractal. Mandelbrot set functions are examples of 2D fractals.
One complete video image. When interlacing is used, a frame is composed of two fields, each containing only half the scan lines.
The part of a display controller that receives drawing commands from the processor and writes the
primitives into the bitmap.
- - - - G - - - -
The direction the virtual camera is pointed in the scene description. The center of the image will display whatever is along the gaze direction from the eye point.
A normal vector of the primitives as they are actually drawn. This often differs from the normal vector of the surface that was being modeled. Facet shading results when the shading normal vector is taken from the geometric normal vector.
A file format for storing images. GIF stands for Graphics Interchange format, and is owned by Compuserve, Inc.
See "interpolation, bi-linear."
A modeling technique where complex shapes are defined as relatively simple, recursive procedures.
An object that the graphics system is capable of drawing into the bitmap. Examples are lines, points, and some polygons.
See "display controller."
See "2D display controller."
- - - - H - - - -
A copy of computer data that is directly readable by a human without using the computer. A printout is a good example of a hard copy.
Another name frequently used for the same thing as IHS. See "IHS."
Constant throughout.
See "refresh rate, horizontal."
Another name frequently used for the same thing as IHS. See "IHS."
HyperText Markup Language. The text formatting language used by documents on the world wide web.
- - - - I - - - -
A color space where colors are defined by their intensity, hue, and saturation attributes. This is sometimes referred to as HSV, which stands for "hue, saturation, value."
A two dimensional array of pixels that together form a picture.
A computer file that contains an image.
A class of modeling techniques that define an object by its surface, but without explicitly providing primitives that make up the surface. Examples are potential functions and iso-surfaces.
See "printer, ink jet."
The name for the global network connecting many computers to each other.
The mathematical process of determining plausible in-between values, given explicit values at particular points. Pixel values of polygons are often interpolated from values explicitly calculated at the vertices. Interpolation is usually much faster than explicitly calculating values.
A method for specifying motion in an animation. Linkages and other constraints are defined for the objects. A final state is then specified for some of the objects, and the computer calculates the motion of all the objects so that the final state is reached. Unlike in inverse kinematics, dynamic properties are taken into account, such as momentum, friction, energy loss in collisions, etc.
A method for specifying motion in an animation. Linkages and other constraints are defined for the objects. A final state is then specified for some of the objects, and the computer calculates the motion of all the objects so that the final state is reached.
An implicit surface that exists wherever a continuous scalar field in a volume is at a particular value (the iso-value).
See "bi-linear interpolation."
See "AND operator."
- - - - J - - - -
See "aliasing."
A platform-independent way of defining procedures for use with world wide web documents.
A lossy image file compression technique for still images.
- - - - K - - - -
A selected frame of an animation at which all the scene state is defined. In the key frame animation method, the scene state at key frames is interpolated to create the scene state at the in-between frames.
An animation control method that works by specifying the complete scene state at selected, or key, frames. The scene state for the remaining frames is interpolated from the state at the key frames.
See "printer, laser."
A vector sometimes used to define the virtual camera orientation in a scene description. This is more typically done with an up vector.
See "ambient light."
The property of light to get dimmer with distance from the light source. Real light intensity is proportional to 1/R**2, where R is the distance from the light source.
See "directional light."
See "point light source."
See "point light source with 1/R**2 falloff."
A ray in ray tracing that is launched from a point on an object towards a light source. The ray is used to determine whether the light source is illuminating the point. If the ray reaches the light source without hitting anything, then the light source is illuminating the point.
A scene object that illuminates other objects.
See "spot light source."
A vector from a point on an object towards a light source. Light vectors are usually unitized before use. A light vector is needed for each light source in computing the apparent color when the object's surface properties include diffuse or specular reflection.
Interpolation is the process of determining plausible in-between values, given explicit values at particular points. Linear means that the values fall along a line from one known point to the next. This means the value changes a fixed amount for a fixed-sized step. Sometimes the term "linear interpolation" is used to refer to "bi-linear interpolation".
See "bi-linear interpolation."
See "boolean operator."
A point in the scene that will project to the center of the image. The gaze vector points from the eye point towards a lookat point.
A compression scheme (see "compression") where all data is preserved. The data may be compressed and de-compressed any number of times without causing any changes. The compression ratio of lossless compression schemes is generally lower than that of lossy schemes. Runlength and LZW encoding are examples of lossless compression schemes.
A compression scheme (see "compression") where some data may be irreversibly lost, in favor of a high compression ratio. Many lossy schemes can trade off between amount of loss and the compression ratio. JPEG and MPEG are examples of lossy compression schemes for images.
See "color lookup table."
See "color lookup table."
A digital data compression scheme that works by identifying and compressing recurring patterns in the
data. LZW stands for Lempel-Ziv and Welch. Unisys corporation now claims that LZW compression is
covered by its U.S. patent number 4,558,302.
- - - - M - - - -
An optical illusion caused by a sudden change in the rate of change of the brightness (discontinuities in the brightness' second derivative). This can give the appearance of a light or dark line at the sudden change.
A popular mathematical function that exhibits fractal characteristics.
See "surface properties."
A constructive solid geometry (CSG) modeling operation on two objects. For the operation "a MINUS b", the resulting object exists wherever A existed without being coincident with B. This operation can also be expresses "a AND (NOT b)."
As used in this book, the process of creating a description of an object or scene for the purpose of subsequent rendering.
See "explicit surface modeling."
See "implicit surface modeling"
See "polygon modeling."
See "potential function."
See "procedural modeling."
See "space subdivision modeling."
A piece of computer hardware for the live display of output. A monitor is different from a CRT, in that a monitor is the complete user-accessible unit that includes a case, power supply, knobs, etc. Many monitors use a CRT as the main imaging component, but other technologies are also used, such as flat panels.
A lossy image file compression technique for motion pictures or animation sequences.
- - - - N - - - -
A vector pointing straight our from (at right angles to) a surface.
See "geometric normal vector."
A constructive solid geometry (CSG) modeling operation on one object. The resulting object exists only where the original object did not.
Non Uniform Rational B-Splines. A particular type of surface spline for making curved surface patches in
modeling complex shapes. These type of splines are supported by many computer aided design (CAD)
systems.
- - - - O - - - -
See "ray tracing, object hierarchy."
A hierarchical method for subdividing a volume. In this method, the volume is originally represented as one rectangular box (parallelpiped, to be more precise). If more detail is needed, the box is split exactly in half along each of its three major dimensions, yielding eight smaller boxes. This process is repeated on each sub-box until the desired level of detail is achieved, or an arbitrary subdivision limit is reached. Octrees are the 3D equivalent of quadtrees.
See "field, odd."
A 3D graphics procedural interface. It was developed by Silicon Graphics, Inc., based on their earlier proprietary GL graphics library.
A constructive solid geometry (CSG) modeling operator on two objects. The resulting object exists where either or both input objects exist. This operation is also call UNION.
The point in a coordinate space where all coordinates are zero.
See "flat projection."
A simple image compositing method where the overlay image results wherever it's not zero, and the
background image results wherever the overlay image is zero. This method is rather simplistic, but is
sometimes supported in low-end hardware. It can be useful in overlaying text, for example. Alpha
buffering is a more general compositing method.
- - - - P - - - -
A format for storing an image that works much like pseudo color in a display controller. Each pixel contains a color ID instead of the actual color value. The true color represented by each color ID is defined in a table called the palette, which must also be stored with the image. A palette is much like a LUT in a pseudo color display controller.
An animation control method where scene state is determined by mathematical functions or computer procedures that take animation time as an input parameter.
A modeling technique where objects are defined by the collective tracks of many individual particles. Randomness is usually used to determine the details for each particle automatically, although overall guidance is supplied by the user.
The property of the human visual system to continue seeing an image a short time (fraction of a second) after it has gone away.
A method of projecting a 3D scene onto a 2D image such that distant objects are displayed smaller than near ones. A normal camera produces images using perspective projection.
A particular method for computing the apparent color of an object at a particular point.
A shading method where the shading normal vector is interpolated for each pixel, then used in a separate apparent color calculation for that pixel.
The material coating the inside of a CRT face. Phosphors have the special property that they emit light when struck by an electron beam.
The property of phosphors to stay lit a short time (fraction of a second) after the electron beam is cut off or moved away.
One red, green, and blue phosphor dot on the face of a color CRT. The dots are arranged in an equilateral triangle. Triads are arranged in a hexagonal pattern, such that each triad is the same distance from each of its six neighbors.
The smallest indivisible unit of a digital image. A pixel is always the same color throughout. An image is a two dimensional array of pixels.
A light source where all the light comes from one point. Although real light gets dimmer farther from a light source, a computer graphics point light source shouldn't be assumed to work that way unless explicitly stated.
A point light source where the light gets dimmer farther from the source. The light intensity is proportional to 1/R**2, where R is the distance to the light source. This formula is used because that's how real light works.
A graphic primitive that looks like a dot, or point.
A graphic primitive that is an area enclosed by a loop of straight edges. Triangles and squares are examples of polygon primitives.
A modeling method where surfaces are approximated with abutting polygons.
A graphic primitive of end-to-end line segments. A polyline primitive is more efficient than each of the line segments as separate vector primitives.
A type of implicit surface. See the text for details.
See "graphics primitive."
See "point primitive."
See "polygon primitive."
See "polyline primitive."
See "quad mesh primitive."
See "vector primitive."
Computer peripherals for producing hard copy on paper and other similar media.
A type of printer that works by evaporating controlled amounts of dye from a ribbon onto the page. The amount of dye can be accurately controlled, yielding continuous color resolution. Dye sublimation printers are relatively expensive, but produce output comparable in quality to the traditional wet silver photographic process.
A type of printer that shoots tiny droplets of ink onto the page. Ink jet printers are a relatively low cost way to get computer graphics output.
A type of printer that works almost like a photocopier. The image is created by a laser under computer control, instead of coming from an original document as in a photocopier.
See "printer, wax transfer."
A type of printer that works by depositing small specs of wax from a ribbon onto the page. The ribbon is pressed against the page, and wax is transferred wherever the ribbon is heated. This type of printer is also called "thermal wax."
An object model defined implicitly by a procedure that can produce volume or surface elements.
See "flat projection."
A scheme for mapping the 3D scene geometry onto the 2D image.
See "flat projection."
See "perspective projection."
A graphics system that stores pseudo colors, instead of true colors, in its bitmap. Pseudo colors are translated to true colors by the color lookup table (LUT).
See "color index value."
- - - - Q - - - -
A graphic primitive that contains a grid of quadrilaterals. A quad mesh primitive is more efficient than the equivalent separate quadrilateral primitives.
A hierarchical method for subdividing an area. In this method, the area is originally represented as one
box (parallelogram, to be precise). If more detail is needed, the box is split exactly in half along each of
its two major dimensions, yielding four smaller boxes. This process is repeated on each sub-box until the
desired level of detail is achieved, or an arbitrary subdivision limit is reached. Quadtrees are the 2D
equivalent of octrees.
- - - - R - - - -
A rendering method that takes into account diffuse reflection between objects.
The name for the pattern the electron beam sweeps out on a CRT face. The image is made of closely spaced scan lines, or horizontal sweeps.
A term sometimes used to denote non-recursive ray tracing.
See "eye ray."
See "light ray."
A rendering method that follows rays of light backwards to eventually find what color they are. Rays may be launched recursively when the color of a point on an object depends on incoming light from specific known directions. For example, when a ray hits a reflective object, a recursive ray is launched in the direction the reflected light is coming from.
A ray tracing speedup method where objects are kept track of in groups. These groups can be further grouped in a hierarchy. A bounding volume is maintained for each group in the hierarchy. If a ray doesn't intersect a bounding volume, then it definitely doesn't intersect any subordinate object in the hierarchy.
A ray tracing speedup method where the scene space is subdivided into blocks. A list is kept for each block indicating which objects a ray could hit in that block. As a ray is traced, it is walked thru the blocks, checking for intersection only with the objects listed in each block.
See "circular reasoning."
See "ray tracing."
A vector used in the phong lighting model to compute the specular reflection. It is usually unitized, and points in the direction light is reflecting off the object.
The rate at which parts of the image on a CRT are re-painted, or refreshed. The horizontal refresh rate is the rate at which individual scan lines are drawn. The vertical refresh rate is the rate at which fields are drawn in interlaced mode, or whole frames are drawn in non-interlaced mode.
The rate at which scan lines are drawn when the image on a CRT is re-drawn, or refreshed.
The rate at which fields are re-drawn on a CRT when in interlaced mode, or the rate at which the whole image is re-drawn when in non-interlaced mode.
See "BSP tree, regular."
The process of deriving an 2D image from the 3D scene description. This is basically the "drawing" step.
The measure of how closely spaced the pixels are in a displayed image. For example, if 1,024 pixels are displayed across a screen that is 12 inches wide, then the image has a resolution of 85 pixels per inch.
A color space where colors are defined as mixtures of the three additive primary colors red, green, and blue.
A vector sometimes used to define the virtual camera orientation in a scene description. This is more typically done with an up vector.
A lossless digital data compression scheme. It identifies identical consecutive values, and replaces them
with just one copy of the value and a repeat count.
- - - - S - - - -
A regular, single number, as opposed to a vector.
One line in a raster scan. Also used to mean one horizontal row of pixels.
A way of arranging image data so that all the pixels for one scan line are stored or transmitted before the pixels for the next scan line.
See "refresh rate."
See "secondary scatter."
The complete 3D description of everything needed to render an image. This includes all the object models, the light sources, and the viewing geometry.
Light that is reflected from a non-emitting object that illuminates other objects. This is the kind of inter- object illumination that is computed by radiosity.
This term is used several ways in computer graphics. However, shading always has something to do with figuring out pixel color values, as opposed to figuring out the geometry or which pixels are to be drawn.
See "bi-linear interpolation."
See "facet shading."
See "flat shading."
See "bi-linear interpolation."
See "bi-linear interpolation."
The normal vector used in determining the apparent color of an object. In facet shading, the shading normal is the geometric normal of the primitives used to model the surface. In smooth shading, the shading normal is the normal vector of the surface that was modeled. The shading normal vector may be further modified by bump mapping.
A thin layer in a color CRT that the electron beams can not penetrate. It is suspended just in front of (from the electron beam's point of view) the phosphor screen. The shadow mask has one hole for each phosphor color triad. Due to the position of the phosphor triads, the shadow mask holes, and the angle of the three electron beams, each electron beam can only hit the phosphor dots of its assigned color.
The Special Interest Group on Graphics of the Association for Computing Machinery (ACM). SIGGRAPH is the premier professional association for computer graphics.
Any shading technique that attempts to make the rendered surface look as close as possible to what was modeled, instead of the primitives used to approximate that model. This is usually done by taking the shading normal vector from modeled surface, instead of from the geometric normal vector.
See "3D texture."
See "adaptive space subdivision."
A modeling technique where the scene is broken into small regions of space. A list of the objects present is kept for each region. Examples of space subdivision modeling are BSP trees and octrees.
See "ray tracing, space subdivision."
The color of an object's shiny highlights. Specular reflection also depends on the specular exponent. Both these parameters are part of the object's surface properties.
This is a scalar value that controls the "tightness" of the specular highlights. A value of zero causes the specular color to be reflected equally in all direction, just like the diffuse color. An infinite value causes it to be reflected the same way a mirror would. Common values are about 5 to 60.
See "B spline."
See "beta spline."
A curved surface which takes its shape from the placement of a set of control points.
A light source that does not shine in all directions, usually in a cone. This is a convenient hack for modeling lamps with shades or reflectors, without actually having to compute the effect of the shade or reflector during rendering.
An input pixel to an operation that uses multiple smaller pixels at higher resolution to compute each resulting pixel at the final resolution. This is done, for example, in an anti-aliasing filtering operation.
See "apparent surface orientation."
The collective name for all the object-specific parameters that are used to compute the apparent color of
a point on that object. These include the diffuse color, specular color, etc. The term "surface properties"
is common but not standard. Other names for the same thing are "visual properties", or "material
properties".
- - - - T - - - -
Aliasing in time by animation frame, instead of spatially by pixel. The visual effect of temporal aliasing has been called "strobing". Moving objects appear to jump thru a sequence of frozen steps, instead of in smooth paths.
The relative fineness or granularity of the patches used to model a surface. A model with smaller, and therefore more patches is said to have a higher tessellation level.
The act of tiling a surface with individual surface patches.
One pixel of a texture map, where the texture map is an image. In diffuse color texture mapping, the texture value is a color. The texture color values are therefore often supplied as an image. A texil is one pixel of this image, as opposed to one pixel of the final image being rendered.
See "3D texture."
The external function that supplies the texture value in texture mapping. In diffuse color texture mapping, the texture map is a set of color values in two dimensions. This is usually specified as an image. Note, however, that not all texture maps are images, because not all textures are two dimensional color values. Bump mapping is a form of texture mapping where the texture is not an image, since the texture values are shading normal vector perturbations instead of colors.
The process where some parameter of the apparent color computation is replaced by an external function, called the texture map. A common example is texture mapping an object's diffuse color from an image. This gives the appearance of the image painted onto the object. Bump mapping is another form of texture mapping discussed in this book.
A form of texture mapping where the texture defines the object's diffuse color. This gives the appearance of pasting the texture image onto the object.
See "bump mapping."
See "printer, wax transfer."
Tag Image File Format. A common image file format. Just about all applications that can import image files support the TIFF format.
The material that is deposited onto the page to form the image in a photocopier or laser printer.
See "3D transform."
See "3D transform."
See "alpha value."
See "phosphor triad."
A graphic primitive that contains a set of successively abutting triangles. A triangle strip primitive is more efficient that the equivalent separate triangle primitives. It is also called a "Tstrip."
A graphics system that stores true colors, as opposed to pseudo colors, in its bitmap.
See "triangle strip primitive."
- - - - U - - - -
See "OR operator."
A vector of length, or magnitude, one.
To make a vector a unit vector. This means adjusting its length to be one without effecting the direction it's pointing.
A vector used to define the orientation of the virtual camera in a scene description.
- - - - V - - - -
A value that has a direction and a magnitude (length). Plain numbers can be called "scalars" to explicitly distinguish them from vectors.
The process of adding two vectors, which results in another vector.
See "basis vector."
See "cross product."
See "dot product."
See "eye vector."
See "geometric normal vector."
See "left vector."
See "light vector."
A graphics primitive that looks like a line segment. Unlike mathematical line segments, vector primitives have finite widths.
See "reflection vector."
See "right vector."
An operation of a vector and a scalar. The resulting vector is the same as the input vector, except that its magnitude (length) is multiplied by the scalar. Scaling a vector by the reciprocal of its magnitude results in a unit vector.
See "shading normal vector."
The width of a vector, or line segment, primitive. Unlike mathematical line segments, computer graphics vector primitives must have a finite width to be visible. Many graphics systems allow the application to specify a width for such primitives.
See "unit vector."
See "unitize."
See "up vector."
See "refresh rate, vertical."
The part of a display controller that reads the pixel data in the bitmap and produces the live video signals. Among other components, the back end contains the color lookup tables (LUTs), the digital to analog converters (DACs), and the logic to generate the video timing signals.
See "display controller."
See "gaze direction."
See "eye point."
A name loosely applied to systems that attempt to immerse the user in a virtual world generated by the computer. This is often done with a stereoscopic display that is updated based on head position, and usually includes some sort of 3D pointing device. More advanced systems include 3D sound and some form of touch feedback. This is still a rapidly changing area.
See "surface properties."
A unit of subdivided space. Voxel stands for "volume element". Nodes in an octree, for example, are referred to as voxels.
Video random access memory. This is really DRAM with additional features specifically for use as bitmap memory in display controllers. VRAM typically costs twice as much as DRAM, but allows the drawing engine full access to the bitmap independent from the video back end. This can increase the hardware drawing rate.
Virtual Reality Modeling Language. The standard description language for 3D graphics in world wide web
documents.
- - - - W - - - -
See "printer, wax transfer."
Software that can read and display world wide web documents.
A rendering method where only the outline of objects and primitives are drawn.
A set of interconnected documents on the internet that adhere to the HTML standard. The world wide web is often abbreviated to WWW.
See "world wide web."
- - - - X - - - -
A constructive solid geometry (CSG) modeling operation on two objects. The resulting object exists where each object existed alone, not coincident with each other. XOR is an abbreviation for "exclusive or."
A window management and 2D graphics environment that is very common on Unix systems.
- - - - Y - - - -
- - - - Z - - - -
The collective name for all the Z values in a bitmap.
A rendering method where an additional depth, or Z, value is kept per pixel. A pixel is only overwritten by a new primitive if the new Z value represents a shorter distance, or depth, from the eye point. The end result is an image of the front most surface of the front most objects.
The name for the value that represents distance from the eye point in a Z buffer rendering system.