I am reading the occlusion clipping section in Real-Time Rendering 3rd Edition, and I could not understand how it works. Some questions:
How is the "Z-pyramid"? Why do we need multiple resolution of the Z-buffer? In the book, it is displayed as follows (left): 
Is the Octree structure the same as Octree, which is used for general rejection and rendering of truncation? Or is it a specialized Octree, made only for occlusion rejection technique?
A more general question: in the previous section (and also here ), the term Query Occlusion Query is described as "rendering a simplified bounding box on an object and comparing its depth results in a Z-buffer, returning the number of visible pixels." What features in OpenGL are associated with this concept of occlusal request?
Is this technique the standard for cutting open world occlusion?
source share