I'm struggling with Generative Fill. And for stuff that is unbelievably simple, like removing a cardboard box from a desk. Like stuff that must be at the absolute most simple/base of the technology's capabilities. So let's say I want to remove a box from the top of an incredibly simple surface, a desk. I select the object with about 4 or 5 pixles as a bit of a buffer, then hit Generative Fill, and almost without exception it just dreams up some crap and puts it there. It really is absurd. And Adobe has the gall to talk about charging people for this feature. So instead of just removing the box, it will replace the box with a bird or a baby or a racecar or something. Why? And yes I have tried the YouTube solution of not leaving the prompt box empty but simply putting a dot/period in it.... does the exact same thing.
@hightidemedia Have you tried entering a simple prompt (the dot/period workaround is outdated now)?
Not entering anything is leaving everything to chance. Even prompting "remove box" would help. I dont subscribe to the "dont tell it what to do" mentality.
"I let go of the steering wheel and the car drove off the wall" - how about pointing it in the right direction?
This is happening for me aswell, trying to do something as simple as removing a person, it just replaces it with some random rendering of a bird. Have also tried the youtube tip of leaving the prompt empty, still it renders 3 random objects instead of actually doing content aware filling in the background like the good old days.
4
Replies
AdChoices