Direct3D Mobile Rendering Pipeline
A version of this page is also available for
4/8/2010
The render target is a memory buffer that holds color values that correspond to pixels displayed on the screen. The rendering pipeline defines how input data, generally vertices, is processed into color values on the render target. It is the computational path that connects 3-D modeling data to the display.
The rendering pipeline offers a pre-defined set of options, which define how the incoming data is processed. These options can be applied in numerous simple and elaborate ways but the set of available options is finite. As a result, the Microsoft® Direct3D® Mobile rendering pipeline is referred to as a fixed-function pipeline. The current state of the rendering pipeline defines exactly which of these fixed functions will be applied to the current input data.
Conceptually, processing in the rendering pipeline can be broken down into the three distinct phases shown below.
During the transformation process, vertices, which only exist in abstract coordinate spaces, are mathematically processed to generate physical screen coordinates for lines, points, and triangles. As part of this process the geometric data for the Diret3D Mobile lighting model is transformed and the lighting values are calculated as well. For more information see Transformation and Lighting.
During rasterization, screen space coordinates for the various primitives are used to determine the exact set of pixels that make up the primitive. Vertex components, such as color and texture data, describing the primitives attributes are interpolated across a set of pixels. For more information see Rasterization.
During pixel processing exact color values for each individual pixel is calculated based on the general lighting levels computed during the transformation process and the attribute values calculated during rasterization. Per pixel depth buffering is used to perform visibility testing and alpha blending is used to write the pixel to the render target. For more information see Per Pixel Operations.