In practice, you are happy with all sorts of artefacts introduced by compression, and that's all this really is – a method of encoding data about what you are looking at in a smaller space.

It's actually part of AV1 https://norkin.org/research/film_grain/index.html