If you are new to this series all the images are based on recent photos of me after basically being fed through a neural network (which can change age, gender, etc.), then edited and painted using various programs. None of the people actually exist, and thus they are like self portraits from alternate universes.
I know she’s got green eyes, but it’s even weirder that she’s got green hair. And I know a bunch of you from my generation are immediately playing this classic song in your heads (Hmmm. Someone should use it as a thumb for the song). Consider listening to the song while reading the post:
Everything’s better with a rock soundtrack. Speaking of classic rock, this image reminded me from the start of one of the portraits on the cover of King Crimson’s “In the Wake of Poseidon”. Check out the upper right:
This was one of the most difficult in the series so far, though that hopefully isn’t apparent in the final image. I’d rather I made it look easy and natural than a prolonged struggle. The reason it was difficult is because it’s an intermediary sort of image, not working from my imagination, and not working from reality. And because of the style, any and all mistakes are immediately and conspicuously apparent.
The impetus for this one was something the neural network did to one of my old digital painting self portraits (of which there is precisely one):
Whoa! Check out that hair! People other than me wouldn’t know this, but the AI customarily does a not-soo-good, or interesting, or convincing job of hair. This had a wild, romantic, sort of fantasy look about it. I wanted to use that hair, but these days I’m really bothered by anatomical inaccuracies, which I am becoming increasingly attuned to. Her right eye is a catastrophe. Fixing that eye is much more difficult that one would think. You can see that, below, where I was trying to work on it:
You may notice that this piece is a bit more illustrational, and that’s partly because in trying to correct errors I went sorta’ academic, as in I used techniques to create proper eyes and tried to integrate the result seamlessly.
A note for artists concerning drawing/painting eyes. It’s best to do the shape of the eye first. Get the lids and shape of the ball of the eye down before adding the iris, pupil, and lashes.
I’m not a portrait artist, nor a realist, and my favorite way to work is directly from my imagination, and without any source imagery. But some projects require different skill sets, and so I have to develop them. In the end I can integrate new skills and create new hybrids.
Here’s an early pic showing the beginning stages of just blocking in the general shapes:
Generally it’s best to move from the general to the particular. Here I’d already put in the pupils, and later I had to take them out and restructure the eyes.
Here’s the final image next to the original AI version:
Yup, it was a real struggle trying to keep the likeness, but correct the anatomy, make it more realistic, but keep the illustrational or painterly quality.
I was thinking this morning that this DOESN’T look like a work by me, unless you are already familiar with the series, in which case the visage might seem very familiar.
Very standard advice for artists is that you need to work in a consistent, recognizable, signature style. I find such a prospect appealing to a much younger versions of my self, with less experience, but by now I find it much more appealing to explore a range of possibilities and create new styles and images rather than be settled in my artistic ways.
If I were magically presented with this image a couple years ago, I don’t know if I’d connect with it as something I created, or would create, and I rather think that’s a good thing.
Here’s an early stage where I was struggling with the left eye:
Here are a couple pics showing the finished product, up close, next to the source of inspiration. The app creates a postage-stamp sized image for sharing on social media, but it is itself useless (at least for me) as a finished artistic creation. Not only are there lots of imperfections, but it’s also far too small to make a decent print. Here you can see how I recreate non-existent detail so that the image can be printing large (in this instant 24″x24″ at 300 dpi).
Note that I’m not drawing over the original, but drawing next to it. I had to invent a lot of hair. There are “hair brushes” one can use in PS, but I like to work with just one brush — a pastel brush — and tweak a few of the settings.
You may have notices how much I changed the mouth and nose. As with the eyes, I corrected the flaws by using standard methods (or my take on them) for realizing them.
Below is the work in progress, and I circled one of the automated layer names, which is 107. That means that during the process of making this image I’d created over 100 layers. Sometimes I work on just one layer, but here I needed to try out so many different things that I was constantly creating new layers and then combining them.
Here’s the final image again:
I have about 5 more I can do in this same style, at least. That’s one of the reasons I worked so hard on this one: to get the skills needed to do more.
Here’s the whole series so far in a slide-show.
Or you can see them in a thumbnail gallery. Just click anywhere inside to go into the screen-show mode.
And if you like the (experimental) sort of art that I do, and you don’t want me to have to quit or put it on a back-burner, please consider chipping in so I can keep working until I drop. Through Patreon, you can give $1 (or more) per month to help keep me going (y’know, so I don’t have to put art back on the back-burner while I slog away at a full-time job). Ah, if only I could amass a few hundred dollars per month this way, I could focus entirely on my art and writing. See how it works here.
Or go directly to my account.
Or you can make a small, one time donation to help me keep on making art and blogging (and restore my faith in humanity simultaneously).