Abstract: Component-based rendering and lightning models process scene data to generate rendered images. Components include input and output parameters, data, and/or instructions for evaluation. The component parameters are dynamically typed, so that each component can output data of any arbitrary type and/or quantity to receiving components. The components can also receive data of any type and/or quantity without any advance knowledge of the existence or functionality of the sending component or its output's data type. The components are dynamically bound, so that the dataflow between components is expressed without recompiling components or scene data. Components can be classified into categories, typically representing a discrete functions, operations, or related properties.

Abstract: Visual representations of different versions of an object are automatically generated and presented to users. Users can compare these visual representations to determine appropriate level of detail transition points based on the visual appearances of different versions of an object. The specified level of detail transition points are converted into level of detail parameter values to be used to select one or more versions of an object for rendering, simulation, or other tasks. The visual representation of each version of an object can include an image sequence of the version at a range of distances from the camera. Each image corresponds to a view of the version at a specific level of detail parameter value. A user interface allows users to view the image sequences associated versions of an object as still images or animation. Users can select images in image sequences as level of detail transition points.