Everyone went nuts for Adobe’s “content-aware fill” in Photoshop when it came out. The boring-sounding feature is in fact an incredibly useful tool, essentially an AI-powered clone stamp that intelligently brought in other pieces of the image to replace your selection. But it still failed in hilarious ways that only an AI is capable of. That should happen a lot less with the hot new tools Adobe is shipping soon.
Teased in a sneak peek video today, the new content-aware fill has a ton of new settings that shooters will love playing with. Photographers love tweaking things — that’s just a fact — and the more things they have to tweak, the better.
So, what used to look like this…
…and would occasionally produce results like this…
…now has a whole right-hand menu full of lovely options to choose from.
The most important difference is certainly the ability to choose which parts of the image the filling agent samples when it’s looking for stuff to put inside your lassoed area. If, as in the fail above, it decides to fill the field in with horse parts, you just exclude the horse from the agent’s consideration with a few brush strokes. No need to be exact — the algorithm is smart enough to work without the handful of pixels that are casualties of overeager mousing.
The improved algorithm also lets you tell the algo that it should be liberal with rotation and scaling of the elements it uses, or that it should mirror-image the content it finds to make it fit better.
Lastly, you can output the fill to a new layer — non-destructive editing is critical to any digital artist’s workflow. I can’t imagine why this wasn’t the case before.
Okay, so maybe some purists will scoff at those who use such heavy-handed means to fix a shot. But sometimes you have to go with the shot you’ve got, and sometimes there’s one too many cows in it. So I at least appreciate the option to do some major alterations without the hassle of manual clone-stamping and other techniques.
The new updates are “coming soon,” so keep an eye out for updates to your client.
Shiv has over 8 years experience working on Internet of Things and an avid user of Drones
