I was wrong today I am unable to recreate the effect with these commands "Dark black background object on shiny black surface" with new beta cannot again get a dark background
Update results after updating beta version to 25.11 m.2679 which was available today.
Black suit jacket and tie result on first generative fill request. Offer up all three versions with a black jacket. Of course, it appears to give each option a Peyton Manning length neck but beggers can't be choosers.
Bug feedback or Discussion please is fine. You tell me what the ai tool is trying to tell me here.
Simple need today to remove shirt and tie, add black suit and tie. I do this normally with two images. Swap out clothing from one image to another. I will with this need. 5 minutes.
But I figure, let's also try the new 25.11.0 beta with what it said was improved generative fill vers.3 with minimal prompt words.
starting with:
Ok, Prompt 1. Add suit and tie. I didn't specify "black" suit, yet. I guess I figured one of the options would be a black suit. my mistake. (marque drawn around whole image space from neck down)
Three images options result:
1. grey striped suit and tie (great if wanted that)
2. white suit and tie (works if wanted white)
3. light grey suit and tie (another fine suit, but I need a black suit)
Ok, so it seems like this might work great at this point but I realized I need to add a descriptive word, such as "black" suit to prompt.
Prompt 2. Add black suit and tie. Three more options are presented.
4. perfect black suit and tie, but it cuts out or masks out most of the suit! I can' t seem to unmask it. it's not there. Just a random cut out shape of white left.
5. Another would-be usable black suit option. But it crops it out yet again!
6 For the third option, it shows me once again that it can provide a full suit jacket and tie. Just not a black suit! Now I'm starting to think Adobe developers don't like me, or this AI tool doesn't. It offers a White Tux like jacket and bow tie. Looks great if that was what you asked for. Again, I prompted "add black suit and tie"
I try one last time to simplify the prompt by removing the word "add". Prompt 3 = black suit and tie
7. Shows a black suit and tie(sort of a tie), only now the option seems to be showing it bursting out of the paper space as a graphic and so shows very little of the suit and tie area. Nobody asked for that, did I?
8. Another black suit and bow/tie, and yet another wild portrait bursting out of the paper graphic effect. Why?
Not sure why it only produced two results the third time. Maybe the photoshop gremlins built into the software now went on strike or started to feel bad about messing with my sanity.
So, should I ask it for a white suit and tie? That way maybe it will finally give me a full black suit and tie result? Should I add, the words pretty please to the next prompt?
That's interesting, so you tried it and received similar result with the beta where it avoids offering a black suit. I was surprised seeing how quickly I received a black suit result with the 25.9.1 public version.
Regarding layers: Yes, I did run subsequent generatives in the same file and it produces a new layer each time I change the prompt. I think I just turn the eye off the first generative layer, reselect the base layer I'm running it on and have it generate another layer if I'm changing the prompt. This seems to keep the original three offerings separate as I try a different prompt. I also may have been turning the visibility eye on or off or randomly selecting layers for the screen captures afterwards. So I can't say I'm certain if the correct layer was selected when I took the screen grabs. Only that I'm sure they came in the order and steps I showed.