The Rendering Process
Camera = viewing volume + projection
Two steps: geometry & rasterization
Notes:
A virtual camera defines a viewing volume. The objects inside this volume are projected on the viewing plane to generate the image.
The rendering process consists of two steps. The first one is the geometric one. The objects coordinates are computed relative to the camera. Then they are clipped according to the viewing volume.
On this drawing, the purple objects are completely out of the viewing volume. They are discarded. The green one is completely inside the viewing volume so it is kept. For the red object, which is partially in the viewing volume, the process is a little more complex. Some new polygons are created to define a new object entirely inside the viewing volume. Only this new object is kept and given to the rasterization process.
The rasterization consists in computing a depth stack for each pixel thanks to the Z-Buffer for occlusion treatment, and then filling the image. On the example, the image shown on the left side contains the green object and only the non-occluded part of the red one.
Generally, the first step - geometry - is faster than the second one - rasterization.