Once I had the initial idea, I used generative AI (I can’t remember which, but it was free) to create some concept images to give to the artist. The following are some cherry-picked results:
This (above) was pretty dang good. This is literally how I imagine Alexis looks in my head, and the blood-soaked bottom of the coat adds a nice edge. However, I purposefully try to avoid describing the way a character looks in too much detail; I want the reader to form their own images in their minds. Thus, for the cover, I want something more ambiguous. Enter: the silhouette.
This one is also pretty great. I started to realize that I wanted the coat and the whistle to stand out in contrast to the figure in silhouette.
This one I painted out by hand to make it fit a wrap-around book cover. The character is in deep silhouette (except the hands), and the coat stands out. Unfortunately, the jeans are blue, the coat is too short, and the image is a bit less dramatic. Still, just for fun, I made an author’s print of the book to see what it would look like.
Not bad, but I want something more impactful, and I don’t want to create actual art with AI anyway; this was just for preliminary concept work. These were a good start, but I came to realize it’s impossible to get AI to generate an image of a long coat that is *really* blowing up behind the wearer (a la the Face/Off scene referenced in my last post). I also couldn’t really find any images like this either, which is probably why AI couldn’t generate it. There are plenty of examples of capes flapping wildly, but not coats.
It turns out my daughter owns a long coat similar to what I imagine Alexis wearing (wrong color, but otherwise pretty close). I took a few test pictures with a powerful fan, to see what we could get.
It takes a lot more air to blow a coat up like that than I thought. Here’s another test, blowing air through the coat without the wearer:
Not much better. It kicks it up a little, but not in the same league that I wanted. I sent all this off to the artist, who had some ideas (particularly, an electric leaf blower). We agreed to get together for a photoshoot to see if we could get what we wanted.
Of course, I could just settle for the imagery I’ve put together so far, but the fact that I can’t find much on the internet (or in AI) like my vision makes me think I’m really onto something and should see it through.
Jonathan Mann 