What the new technique does is steal a bit of light from the core and use it to suppress some of those ripples. It interferes with them, damping them down. It only works on half the image, as you can see: on the right the ripples are obvious, and on the left they are essentially gone. If you’re looking for a planet, it means you have to observe it several times rotating the detector so you can clean up the halo all the way around the star.
The Steward Observatory team used a machined piece of infrared optical glass about the size and shape of a cough drop to introduce the ripples. Placed in the optical path of the telescope, the APP device steals a small portion of the starlight and diffracts it into the star’s halo, canceling it out.
I really like this because it’s a hack. Astronomers have been trying for hundreds of years to get rid of excess diffracted light, and you can’t do it. You just can’t. But what you can do is get rid of some of it, and shove the rest into parts of the image that you don’t care about looking at. And that turns out to be better than good enough. (Technically, what they’re doing is trading spatial resolution for time, since you have to observe for much longer to image all angles around a star, but that’s OK too. And, as with most things that have thresholds, you’re not going from mediocre signal to slightly better signal, but from no signal to signal at all…)