I've decided to give images their own section to explore them in more detail
The main issue I have at the moment is the potentially vast number of variations each image could need in order to exactly match the way they are being viewed. Let's start with a relatively simple scenario - that you're probably looking at now - a two column grid within a container.