my version of Batman, if you want to see the raw draft before, it's here
Each Ai Art software has its own lingo, its own system, and with the advancement on the technology, what used to be impossible a few months ago, is now easier than ever on Midjourney, they now offer a possibility to "Remaster" your work.
The remaster feature allows you to make new variations of the images created and upscaled with the v1/v2/v3 versions. Remastered AI images will be more coherent in the style of the new “test” version but will still have that artistic touch from the standard Midjourney versions.
Of course, you can create only “test” or “testp” images right away and will also have coherent and very photo-realistic results. The benefit of using the remaster feature is the endless creativity of the standard Midjourney version combined with the new cohesive look. The result is awe-inspiring AI pictures, which you will love even more.
I wanted to show you some of my work revisited in the last few months, and how I started including "remaster" as a part of the creative process:
I actually have never showed you this one before, probably because I didn't really like the way it turned.
As I was explaining to myself earlier and even in a previous post, that I am pretty sure that GFGPAN is somehow involved in the remasterisation of each atwork.
You can either decide that you have what you're looking for, or keep on going further down the "refining" process, that can often becomes a giant rabbit hole.
As you can see, it really adds another layer of realism.
Old Rasta Man with wrinkles
Funny to think that I was pretty happy about this one in its original state, something about the eyes.
I ended up doing a video for Instagram with the evolution of the faces:
This is 2 of the best 10 I got, but all the faces were so interesting, and I am hoping one day I can print them, some really beautiful texture and photorealism.
So the wolf god was one of the first time, I felt that I reached a new level when I prompted it.
Little did I know that there were few more levels to go through and that the end result will be even more interesting.
I think "the weight of the prompt" as we call it, has been modfied recently on Midjourney.
In other words, you words matter more than they used to. It will take more of your input into consideration, and less "artistic interpretation" by the AI.
How do I know that? Because I tested it on roughly 2000 images, and up until now, it was DallE that was leading the way in term of "the weight of your words", then Stable Diffusion, and Midjourney seemed to have finally modified their algorithm to something similar.
I've stopped working on my comic idea because I was finishing a single in studio in the last month, that I will be releasing soon.
But the prospect of working on it again excite me a lot, I think I want to go back to drawing, and take my time crafting something that I enjoy.
You see, in the prompt I was talking about the katana, and it wasn't even being interpreted using the original Midjourney algorithm.
I am also still thinking about Jack Stone, the main protagonist, and In term of face, something with 25% Brad Pitt in it, seems to work nicely.
Golden Panther Godesses
The remastering is also giving me issues, because I often forget what I wrote in the prompt, most of stuffs are a lot by accident, and in my prompt somewhere was hidden "Black Panther", and I forgot about it, but when remastering the artwork, it reinterpreted the Black Panther in the comicbook way, and in that case there is no way to claim it as original work anymore.
Red Bull God
Here is the same challenge, because there is the word "Optimus Prime" in my prompt, the first result is visually very different, using the oiginal Midjourney.
But when remastered and the weight of the prompt having a bigger influence on the outcome, it suddenly fleshes itself out a transformer, though I could argue that you could probably get away with this one, if you don't name it "Optimus Prime".
The first draft was never really what I wanted to achieve, and remastering it had exactly the effect of shaping it the right way.
The intricate elven details is something that I love to incorporate in the prompt, always so nice to look at.
And the religous aspect is reinforced by the beautiful rendering of the gold, cool stuff!
Yes, I put Elvish in my prompt, not elf, not elven, I wonder how much of it has an influence on the final result.
This one is a good example on how I am not even sure, which one is the best.
You see, the top one is the original, and I really enjoy the deconstructed perspective, and the blend between nature and architecture, it almost blends in like an hobbit house.
Each remastering add another layer of interpretation, it will sway in one artistic direction, and you can do that indefinitely, to the point that the end result can really be miles away from the original.
The last one, was my favourite, I can't really explain why.
Robot Biden Creative Process
Man, I could do an entire post about Robot Biden. I got a bit obssessed, and my creativity went in all different directions, this is when I started really understanding the power of the new Midjourney AI and how quickly we are evolving.
It was September, and I started realizing that what I did before was already becoming obsolete.
My eye also starting getting used to spotting "imprefections" and "incoherences" making my workflow a little faster.
Really happy with the end result. I didn't get the success I thought it would on Instagram, but at the same time, I am not even sure of my statement, I just find the idea of a robot Biden, funny af.
I don't always like the new results with the remastering, Soul Mate is a good example of this, I just prefer the rough texture of the top one, and the new interpretation is a liitle bit too smoothen out for me.
What do you think?
Nuclear explosions with balloons
Now I use my first prompt for having a rough idea, and once I have a vague sketch, I then use the remaster as a refining tool, getting the best out of each images.
Really cool intricate details that I didn't think would ever been translated the right way, but now it is!
Lion kid Creative process
Once again I get the rough form of what I want, selected between 4 pictures generated by Midjourney.
And then polish the one that I've selected...