This topic provides definitions and links to more information for common computer graphics terms used throughout Insight3D®.
To "remove from a flock." In Insight3D, culling is used to improvement performance by not rendering objects that are not visible. Culling provides view-dependent performance improvements because the number of visible objects may depend on the camera's position. See also Horizon Culling and View Frustum Culling. The Real-Time Rendering slides provide a comprehension overview of culling techniques (see Spatial.ppt).
Data for a pixel before it is written into the frame buffer. A fragment may not become a pixel if it fails a test, such as the depth test.
Graphics processing unit. Loosely speaking, a video card.
A technique to improve performance by rendering an increasingly simplified object as the camera moves further away from the object (to be more precise, as the object occupies fewer and fewer pixels on the screen). This is used extensively by Globe Overlays. For more information see A Developer's Survey of Polygonal Simplification Algorithms and Level of detail for 3D Graphics.
An algorithm whose working set is not all in memory at once. This is common for terrain and imagery, since the datasets are so large. Globe Overlays use an out-of-core algorithm to only keep the most relevant data in memory.
To draw. "Draw the scene" and "Render the scene" are equivalent.
A texture element, similar to a pixel in an image.
A mesh made up of just triangles. This is the most common representation for objects in real-time computer graphics, as described in the Triangle Mesh Primitive topic.
A point. A triangle having three points is the same as having three vertices.