Copy link to clipboard
Copied
Hi, I'm having a hard time getting decent results with generative fill. While I am probably a victim of all the fanfare about miracle-working AI, I was also thinking maybe I am just doing it wrong.
Most times I try to select the subject (human) and then ask AI to provide a whole new background.
Is this something I can expect from PS at this point? I've seen the video and most times it seems about "add this object there" or "expand this existing scenario".
Can you please help me understand if it's just a matter of prompts or maybe the tool is not meant to do this?
Thanks
Copy link to clipboard
Copied
Maybe you should post screenshots of what you want to achieve, so that users can replicate that.
Copy link to clipboard
Copied
You are absolutely right, and both the comments made me think that what I was actually expecting was from the AI to work some magic that I didn't even know in details my self, like "surprise me".
This was the outcome of a very simple prompt "star sky, art", but I found every element to be quite cheap and poorly refined. How could I "improve" this? Can I point the AI towards a specific reference if I find one?
Copy link to clipboard
Copied
You entered “art”, so you will get art…
Changing art to photo:
All depends on the prompt you use for getting a result in one or the other direction.
The initial image:
Copy link to clipboard
Copied
You did not provide a single example so is dififcult to assist. Would you call a mechanic and expect them to know be able to help with your car, without them being able to see your car?
https://www.youtube.com/watch?v=NvUZIm083P8&t=779s