Part of the fun of working with AI is the ability to find your own personal favorites, and then give them a makeover. My re-worked image today is one such item.
Before we get to that, my DEVIANTART GENERATION TODAY is an old OTK favorite from this website, and it is an image that I personally own having paid $20 for it. The cartooned version of Al Siegel spanking Connie Russell.
To start with, today’s effort was the first piece of mainstream spanking art that I have been trying to create that was rejected. It wasn’t rejected for the spanking, it was rejected because the actor in the image, Bill Owen, is a household name. If you are from America, or do not know so otherwise, Owen played a TV character for years where virtually everyone in the UK knew who he was.
The funny part, though, after the AI chat had explained the whys and wherefores, it followed up by telling me what I should have done to avoid the censor, lol. So I took its advice, and I applied their suggestion on a different account, and this time it worked.
It gave me great pleasure to give a makeover to the lobby card spanking featured in “The Girl Who Couldn’t Quite.” The results of this effort can be seen below, and I think that a stellar job was done.
Hang on, though, this is AI, surely it isn’t prone to mistakes, is it? While I thought that the creation looked wonderful, there is one glaring error in this color rendition of that old lobby card. Can you see what that error is? I bet you a dime to a dollar that you will spot it quickly. Like always, click on the image itself for the full-sized version.
Did you spot it? Did you notice that old Bill seems to be missing something? The AI asked me if I wanted to work to fix it, but I decided no, I’m not going to fix it. In fact, the color really enhances it, eh? I’m pretty sure that you have noticed what’s missing already.
Now then, now that you can see that old Bill is missing something, let me share the original image once more that I have shared on here before. Before I do so, though, let me ask you a question. In the black and white version of this lobby card, did you ever notice before that the item we are missing above, was never there to begin with. Be honest, have you spotted the error previously?
Once I completed the task, and I said to the AI that I had never noticed it before, the AI then shared some of the old movie tricks and tips. It surmised that it was probably a problem area, and the editor said “just get rid of it, nobody is going to notice anyway.”
The color really highlighted what is obvious, once you realize that something is missing. So what do you think guys, do you like today’s effort?
I’m tempted to give a cryptic answer – is it a bit like Peter Cook and Dudley Moore on the “Tarzan” sketch… deficient, to the tune of one! LOL
I hadn’t noticed before either (if the missing leg is what you had in mind) but I see it a lot with AI models. The “Chase & Sanborn” image I shared on masto/bluesky earlier is a composite, because when I finally got the perfect composition and background, the model decided to “vanish” the woman’s legs entirely, and the man’s right hand disappeared!! I selected and blended them in from a previous version. Her hand also had to be heavily retouched, as the model absolutely refused to give her 4 fingers! I rebuilt it with a lot of copy/paste, warping, and blending in. Hands (a little bit in your picture here too) seem to pose a major challenge.
I think you’ve done a great job… again! I think it’s a fab scene. The overall movie, maybe not quite as good as that scene! 🙂
J,
You nailed it on the money. It took me several generations before I paused to check the original. I wanted to see why it wasn’t picking it up, and that’s when I saw that it was never there to begin with. AI then educated me on artistic practices of the past, and how the shadows of the black and white trick you into thinking it is there. The moment is goes to color, though, it is spotted immediately.
Oh, the limbs, the limbs. It is the bane of AI art, isn’t it? On two occasions I have literally had to draw a limb in using paint for AI to recognize it. There was another one where I had to falsely instruct AI that the guy’s hand was in the air, palm showing, and it worked, however, it also left the hand on her bottom, lol. Thankfully it had rendered a green flowered skirt, and it was easy to cut a segment of the skirt and paste it over the hand. It looks seamless.
Another issue I encounter frequently is footwear, though a lot of times I just accept it because nobody it looking at the feet anyway. There was one that I did that involved a male and a female spanking two girls with shoes, and two people in the background. For whatever reason, the AI also interpreted that one of the guys in the background was also holding a shoe in the air. I couldn’t get rid of it for love nor money, and it took ages before I finally got it. Then I was giving it the once over for any final edits and I was like, “hang on, where did he come from?” When the program removed the third shoe, it added an extra man into the photo. He actually blends in well, though, so I left him in there.
Oh, and finally, the image that I am sharing tomorrow on my blog, on the initial rendering it had the two clearly defined men in the picture, but for the lady in the dress, AI saw a man in the dress. Each to their own, I have no problem with that, but it wasn’t the aesthetic that I was going for. The problem, however, is how do I change that without triggering any ‘rejected’ keywords? It was quite the brain teaser, but I finally worked it out.
Rich
Yes as much as I have seen this spanking I never did notice the missing leg. Thanks for showing them both. I do enjoy the colorized one very much. Thanks for sharing and have a great day.
Jim
Jim,
That is good to know because I have a different color version for DeviantArt. In that one it is a yellow top, pink skirt and Bobby sox. 😉
Rich