Blender 3D: Noob to Pro/Perspective Views
As you know, the main reason for modeling 3D objects in Blender is to render images that exhibit the illusion of depth.
Orthographic views are great for building a house, but seriously flawed when it comes to creating realistic images of the house for use in a sales brochure. While a builder wants blueprints that are clear and accurate, a seller wants imagery that's aesthetically pleasing, with the illusion of depth. Blender makes it easy to use tricks like perspective, surface hiding, shading, and animation to achieve this illusion.
How does perspective work?
The essence of perspective is to represent parallel edges (in a 3D scene) by edges (in the 2D image) that are not parallel. When done correctly, this produces foreshortening (nearby objects depicted at a larger scale than faraway ones) and contributes to the illusion of depth.
Perspective is challenging to draw by hand, but Blender does it for you, provided you give it a 3D model of the scene and tell it where to view the scene from.
If you're confident that you understand perspective, you can skip the rest of this module and proceed to the "Coordinate Spaces in Blender" module.
Blender only does 3-point perspective, not 1-point or 2-point.
Drawing classes teach several different kinds of perspective drawing: one-point perspective, two-point perspective, and three-point perspective. In this context, the word "point" refers to what artists call a vanishing point.
One-point perspective works when you're looking at a 3D object head-on and it's centered in your view.
Imagine looking down a straight and level set of train tracks. The tracks will appear to converge at a point on the horizon. This is the vanishing point.
The image on the right is a 2D image of a cubic lattice or framework. Like any cube, it has six square faces and twelve straight edges. In the 3D world, four of the edges are parallel to our line-of-sight; they connect the four corners of the nearest square to the corresponding corners of the farthest one. Each of these edges is parallel to the other three.
In the 2D image, those same four edges appear to converge toward a vanishing point, contributing to the illusion of a depth. Since this is one-point perspective, there is a single point of convergence at the center of the image.
Now the cube is at eye level, and you're near one of its edges. Since you're not viewing it face-on, you can't draw it realistically using one-point perspective. The horizontal edges on your left appear to converge at a point on the horizon to the left of the cube, while those on the right converge to the right. To illustrate the cube with a good illusion of depth, you need two vanishing points.
Now imagine that you're above the cube near one of its corners. To draw this, you'd need three vanishing points, one for each set of parallel edges.
From this perspective, there are no longer any edges which appear parallel. The four vertical edges, the four left-right edges, and the four in-out edges each converge toward a different vanishing point.