red = Observing // blue = Preloading // gray = Loaded image
The gray area around the center should just demonstrate a viewport, the white area would be the visible browser area.
I've added a random number after each image path for unique filenames (avoid caching)
Just play around a bit with the obSettings var in the JavaScript part of the pen.
Polyfill for IntersectionObserver