Down the rabbit hole, backwards.
I have been casting about, trying to get the engines going around ideas of images that are not stale, overdone, boring to me. To no avail. I have been trying new materials, new books, new sites, new thoughts. To no avail—and yet, I saw someone posting this very thought provoking image citing “ai” (artificial intelligence) and the engines they used (NightCafe Studio and MidJourney). I am on Dall-e 2’s waiting list and would love to be considered for the engine that Google is working on—as their products have a nice level of simplicity and sophistication. However, I figured hey, here is a pointer to something new—dive in with NightCafe and MidJourney.
NightCafe Studio is lovely and an intro to this world. It is simple and frankly, after a dozen shots, kind of predictable as to what to expect. You can upload an image or let the engine grab something, and according to the text you input and the prompts/and modifiers you indicate, you can create a single image with no real opportunity to tune it or change it except for trying the same text again, and seeing what evolves visually.
It’s easy in, easy to use, predictable. It features modifiers such as artists Thomas Kincaid, lots of pretty impressionists and Van Goph. It references what is trending on ArtStation (new one for me)— The options and where you can take an image is very light, airy, and very topical—with this work having no real soul—just eye candy. No problem if that is your intent, or no problem should you want to stick you toe into the ai world. Additionally, it is linked to the land of NFT which, I am absolutely clueless about, but if this is your intent, NightStudio might be fun.
Now for MidJourney. First off, MidJourney is set up on Discord :
“Discord is a VoIP and instant messaging social platform. Users have the ability to communicate with voice calls, video calls, text messaging, media and files in private chats or as part of communities called "servers".[note 2] A server is a collection of persistent chat rooms and voice channels which can be accessed via invite links.”
Wikipedia
So, there’s the familiarity with this new environment that is the first test—no biggie, but something to learn. Discord has a good online support team, so if you have questions or are experiencing issues—they are immediately answered which is terrific. Then, there is the “server” for MidJourney—with all sorts of spaces for newbies and for general work, along with a full fledged manual, corporate and individual support.
MidJourney is all about writing text which then the engine manifests into a pane of 4 takes on an image. The user can upscale the image(s) and/or get further iterations on the selected image to keep driving the idea further. You cannot add/or subtract (my understanding now) to tune the image as you go—but you can easily copy the text, modify it and re-send the idea to see what happens. The key to MidJourney (for me—about a week in) is that my background as an illustrator, designer and art director helps as I have always had to describe what I am looking for visually— and this engine will do things as it will not question hybrids, fusion, and odd combinations.
Some thoughts as to how I am learning about this medium:
How do you stack the words?
What are the words? What is the difference in “kawaii” and “cute” (as an example)? Or “Graphic design” and “simple”
Who are the styling references? The engine knows “Hildegard of Bingen” and “Mary Blair, but does not know the “Posada” or even that of “ the Provensens”. I am trying the same idea with a simple text entry, only changing the reference, the same with descriptors. I am keeping notes as to how others are dealing with ideas, or even how they are setting up the text.
Forget about hands, and be prepared to not love faces….but work with those.
Landscapes rock.
Architecture rocks.
Presentation prep (for branding etc) is invaluable.
.How do I talk about camera angle, lighting, tone, mood?
MidJourney is far more sophisticated but also has some nuances that one must get comfortable with. MidJourney is the tool I am committing to. I am seeing it as a strong comping tool, a strong ideating tool which also can create images that can be doctored in Photoshop, or just as a starting point for my own work.
I have notice that there is a slight soulless spirit to the images, they do the work, but they do not have the umph of emotion that I am questing for…but hey, this is the beginning of this exploration, and I am but a newbie, but the horizon is filled with opportunity to see where this goes. Not for NFTs, but more to fill my head and art with ideas, and my heart with creativity.
