Page 8 focuses on clips from diffusion based pipelines, the same family of models behind most modern imagery. Diffusion builds each frame from noise, which gives these clips their particular texture. No camera and no real person was ever involved.
Diffusion models build each frame out of noise, and that process gives these clips their particular grain and texture. Studying the page is a quick way to see how diffusion based video behaves on adult prompts. No camera ever touched any of it.
Keep browsing: open the full video archive index, or dig into the Ironline set and the Mainline clip collection. The archive currently holds 1740 clips across 35 pages, every one with.