Rasterizing

Rasterizing is the process of converting vector graphics or geometric descriptions into a raster image composed of discrete picture elements, or pixels, on a fixed grid. It is a core stage in most raster-based rendering pipelines, especially for real-time 2D and 3D graphics. Vector primitives such as lines, polygons, and text are defined mathematically; rasterization determines which pixels are covered by these primitives and assigns color values accordingly.

In 3D graphics, the pipeline transforms vertices through model, view, and projection stages to screen space. The rasterizer then converts these primitives, typically triangles, into fragments or samples that map to pixels. Techniques include scanline methods, edge functions, and barycentric coordinates to determine coverage. Fragment attributes such as color, texture coordinates, and normals are interpolated across the primitive, enabling shading models like Gouraud and Phong during the rasterization process or in a subsequent fragment shader.

Depth testing and stencil testing are commonly used to resolve visibility and mask operations, while texture mapping supplies surface detail via sampling. Anti-aliasing methods, including multisample anti-aliasing and supersampling, reduce jagged edges by oversampling coverage. Rasterization is usually implemented in hardware on GPUs, delivering high throughput for real-time rendering, though some pipelines incorporate ray tracing for more physically accurate illumination.

Applications include video games, graphical user interfaces, and any system that displays vector-defined graphics as pixels. Rasterization contrasts with vector displays and with ray tracing, the latter being more computationally intensive but capable of more accurate light simulation.