Late Reply, But from my experience (the hard way), dividing the projected z by w (after projection) is a technique that non-linearly scales depth, providing greater precision near the camera while compressing depth further away. This approach helps mitigate issues like z-fighting at close ranges, where it tends to be more noticeable. Alternatively, if needed, you could create a custom projection matrix that doesn’t require this post-projection division (for z), resulting in a linearly distributed depth between the near and far planes.