The Story
What's with all the kitty cats with big eyes? Or all the buxom fantasy women dressed up like Vikings or Angels who look to be strutting along some ancient red carpet event doused in overly dramatic lighting?!?
That's kind of what I am seeing in a LOT of the AI art out there (especially on the service I've been using of late... Night Cafe... the site is just rife with buggy-eyed kittens and Amazonian-looking gals in various degrees of undress). So be it, I guess. Wouldn't hang any of it on my walls, but hey, there's no accounting for taste, right? There's people who love that gawdawful painter of light guy, too. So be it.
However, I have taken note that there are some VERY interesting works coming out of all this stuff, like those of my contemporary DBFISH, such as his Mechanical Marvel. Groovy stuff! I like it!
As I set about creating the cover art for Smokin' Dirty Sock, I put into play some more traditional looking photograph-based pieces for the track ...
The track has samples of thunder and rain and other sounds in it, so I did some shooting during one of our atmospheric river rains and thought I might try to use one of those shots for the cover (the second one above is an obvious bow to the record covers of early English proggers Henry Cow from the '70s).
...but neither of those WIPs had the loopy sense of fun I was looking for for a track with such a goofy title (maybe I'll use them for something else!).
I get an email from Night Cafe every evening giving me a few free credits (what you 'pay' with to create works on their site... you will get these nightly emails too if you sign up for a free account), so with 300 or so Credits available to me, I turned my thoughts to trying to tell the Smokin' Dirty Sock story with a piece of AI generated art. I only used up about 18 Credits total to create what became my final piece (seen below). I wasted a few credits bumping around on my initial idea of a cabin in a bayou with frogs playing instruments on the porch in the middle of a storm...
Hm. Weird fish in the sky (kinda frog-looking?...)... a great big Adirondack chair (well, of course!)... a green moon? I kinda like it, but it's all just so absurd!
Obviously I have no idea (yet!) how to write a good prompt to create AI artwork via the various AI art engines out there. The prompt for the cover I eventually ended up with reads as follows: A dirty athletic sock stuffed into the top of a large bong, smoke trailing up from it, hyperrealistic, 8k, cosmic, noisy, splashes.
I had you at BONG, didn't I?...
In any case, along my way I did learn an interesting bit regarding stuff you DON'T want to see in your images. You achieve this by including a 'negative prompt' that drops stuff OUT of your final rendered art. Many users use EXACTLY the same negative prompt, so I attached same to this work: ugly, tiling, poorly drawn hands, poorly drawn feet, poorly drawn face, out of frame, extra limbs, disfigured, deformed, body out of frame, blurry, bad anatomy, blurred, watermark, grainy, signature, cut off, draft. You make this second prompt a negative to the art to be rendered by the machine by giving it a negative weight. Most folks use a weight of "-.3" for this secondary prompt, so that's what I used, too.
What came out was this image:
I quite liked that! But as with everything I touch, I had to toy with it in Photoshop to get to my final cover art image: I desaturated the shoe (and gave the back of the shoe a lip, which it certainly needed!), added grain, gave it a vignette, brought in the title and my name (of course), etc.
I do find the whole AI process interesting and intriguing, but don't see the final products created to be 'final' to my eye (and brain). Not yet anyway. I think I will always futz with them beyond what the AI engines do. It's in my DNA to futz. Maybe things will get better for me as I progress with prompt-writing skillz. We'll see.
FINAL:
Done. /\/\
