Image processing for empathizing grain/pixel resolution without keeping semantic information?

Question

I am trying to process images to emphasize (and be able to tell apart) film grain vs. pixel resolution.

For example, in this image (from here):

I would like to ap