The magician continues to deceive and weave her spells - part three

Pictures to cull through today make it difficult to do this, as all the thumbnails look pretty good. This is the danger of a burgeoning form of art, and generative art work in particular, is that it makes it very difficult to know when something is done, as you think you may get an even better outcome the next time you click that generate button.

It is a hit of dopamine, I won't lie.

I like a lot of these, going by the thumbnail alone. There's interesting poses and shapes throughout the selection, and I think, by looking at the thumbnails alone, it would be difficult to eliminate any of these images as being non suitable, but I must. I must select only those that are the very best compositionally and thematically.

The first three two, when viewed up close have issues with the facial features and composition,but hte third one eliminates all of that, but I am not quire sure that she might be a magician.

I really like the circular window behind the head, it reminds me of a lot of renaissance style paintings where this feature was a way of replacing the halo, or showing the innocence of a character. It also framed the face, which helped to bring the eye to the point of most emotion in an image typically being the facial features.

Now here, we are missing an entire head. That would be okay if the character wasn't important, but they are. That is something that can be fixed, using a feature, or script leveraging Stable Diffusion to do something called "out painting", which expands the canvas in new and interesting ways.

Here, I've done that, and this was the outcome of a single run. Running on my M2 macbook air, this took about two and a half minutes to generate the head and the additional bacgkround detail, with little no input on my behalf. It would have been much quicker using my 4090, but I am typing this post out on my laptop. It is pretty incredible that this can run on ARM based silicon at all, given the underlying libraries are designed with nvidia's CUDA in mind.

I could generate more variants, but why would I do that when I already have so very many images encompassing this theme to go through? Speaking of which, let us have a look at another image.

This is probably the strongest of the remanning representations in terms of getting to the point of my underlying prompt, and well, it is a complete image overall. There's decent facial structure and detail, well rendered hands, and importantly, many of the representational elements that align the picture with that of The Magician. It is a good one to short list, to be sure.

The other images, while having striking thumbnails, were not quite what I was looking for in terms of composition. Not every generation is a winner. Not every picture is a winner. It is all part of the experimental journey, which is constant and ongoing.

Looking at these images in review teaches me about what I want the final image to actually look like, and that's awesome; because I can improve and iterate my art direction skills enormously as I go along.

And here is what I will contend with next time. Each of these images deserves a closer examination, and you'll probably see a few familiar images here that I decided to upscale on the basis of them being in the short list last time.

---

Want more content from me?

Witness my futile efforts to play my Steam Game collection in alphabetical order.

Are you aware that I love photography? Check out my work in a collection.

---

If you haven't started playing Splinterlands, you should do that immediately. It's very good fun.

If you want to see my Splinterlands antics and rants live, Find me on Twitch

If you prefer sleeping in your designated time zone, go watch replays on YouTube.

---

Thanks as always for your time!