Late last night I got into the Adobe Firefly Generative AI Beta and started playing around
This week Adobe announced their new Firefly Generative AI Beta that you can sign up for. I did on the first day and got in around 8 PM last night. I was playing on my iPad last night (which was fun) though the download buttons seems a little funky on iOS and it didn’t download everything that I asked it to. So I played some more on my Mac today.
There currently 2 active modules, Text to Image and Text Effects.
The type effects are very cool and fast. My biggest complaint was it certainly can’t do one plant or animal per letter, it uses them more as textures than objects, but still I will be using this quite a lot for certain things.
And as for Text to Image, you certainly have to get the prompt right.
For 1960’s psychedelic guitar playing man with beard onstage at woodstock, it went for a modern and disturbing image.
Fairly impressive hands, though those sunglasses, and not sure about the hair in some.
I then tightened it to psychedelic guitar playing man with beard on stage at woodstock in 1968 and things got more interesting.
Love the first and 3rd, then went for graphic instead of photo of the same.
The 2nd one is much more interesting this way, and the 3rd has Floyd from the muppets eyes, or maybe the grateful dead, no idea, but creepy, and the 4th is pretty damn fantastic, though a strangely 7 string guitar with 4 strings.
And then went for an art variation.
These are all pretty damn amazing!
And I do love the controls for image refinement,
And I tried to do some imagery for my short film THE MISADVENTURES OF BEAR, while I do love what it came up with. It certainly isn’t listening to the “dark hallway at night” it is a hallway, But it is certainly lit, and I can’t seem to get it to go dark. The first 4 are art, and then photo afterwards.
And the next are photo.
And then i stayed on photo but tried to go darker, but setting color and tone to cool, and lighting to low lighting.
And then I changed the promp to down a hallway in the dark to see if that would do it, and it didn’t, but I do like the clothing variations it added.
So wow, a really powerful tool I will keep playing with, but it doesn’t seem to know how to do low light or unlit images, or day for night?