Another frontier is light/matter interactions. This is an area where a ton of progress has occurred in a wide range of areas. And that work often has direct implications. I worked on a microscope recently that coupled 6 different lasers into a single fiber for fluorescent illumination. The tools we got to work with... some are from 100 years ago and are still great, others, like spatial light modulators and other light-conditioning tools are really amazing. These days you can reproduce the Michelson Morely experiment (technically VERY challenging at the time) in an afternoon. Or do a basic quantum eraser experiment with $2K in optical components(!)
One of the other students in my lab put it this way (very paraphrased):
If you break up the Nobel Prizes and look at the optics based ones, ones that made up some new microscope, or used light in some new way, then Optics becomes a top 5 domain. In that a lot of Nobels come out of manipulating light. And a lot of the time, the sum total 'new thing' worthy of a Nobel is just a single optical object, sometimes it's five, not much more. But, the thing is, if you look at how long those experiments took to preform, from concept to working prototype taking some kind of data, then Optics varies wildly. Some, like DIC imaging, is just a single 1/4 waveplate in the middle, and it took Nomarski nearly a decade and a half to get it just right. Some, like STED, were built in Hell's living room over about a month. And that's the thing about Optics research, it's just blind luck most of the time. Sure, the equations help, but they tend to come after the apparatus is up and running. You can spend a few weekends in your living room in your underwear tinkering about while watching the game, or you can spend your entire career painstakingly laboring in the nauseating dark of the lab all alone. Incremental progress just doesn't seem to be a thing in Optics; you either get it right or you don't. Anyone that has been in totality during an eclipse knows that there is a huge difference between 99% and 100% coverage. Optics is exactly like that.
