|
||||||||||||||||||||||||||||||||||
 Sketch Reward for
Caffeinated Critter!!!
0 Comments
 The journey to learning how to create the ideal Female Werewolf with A.I. has been a pretty chaotic one. Stable Diffusion has proven to overall be a wash for me, with how unintuitive its systems are. At some point I'd like to attempt to train a model for werewolves on it with my better results elsewhere but for now I feel I'm starting to get close with Midjourney... and I'd likely have succeeded already if not for the dam censorship. Either way I'm starting to find some fun poses that react well to some of the better models I've created so far as inputs. Which as been no easy task with most of my input attempts being blocked for censorship. Decided to draw over one I liked the energy of below adding my own unique style to it. Cant ever just use the base 1 for 1 due to how inaccurate the anatomy often is, but is pretty good as a general guideline to follow. I'll be posting the better of my results to a new page under the concept art section. Some of them make for very nice inputs tho most are blocked by the system. ;0 Not sure if i'll take this pic to completion yet but so far its a fun pose. Tell me what you think. ^_ ^
 Finally complete!
One of my first illustrations created with the aid of AI. in this ones case Midjourney. Were I took inspiration from a prompt result I created accidentally while trying to get completely different results. It was an alluring yet spooky sort of result that I fell in love with and used the better poses I found there fixing up the anatomy along the way. So with this one no actual AI is directly represented in the finished work it was ultimately all hand drawn in Photoshop. In the future though I want to experiment with developing backgrounds for my illustrations to hopefully vastly speed up my workflow. I'll experiment with trying to generate a background image to my liking and drawing my characters onto it. whether I get a great singular piece or have to combine several good elements from different outputs I will ultimately be going back over it by hand to strengthen or correct various components to ensure things match my style, Am excited too to see how it helps improve my ability to render such myself. Already even with this piece I feel that my ability to do a more painterly background has already been improved upon through analyzing the styles and mediums of AI results I was very fond of. I'll make a space later to show off the AI results I liked best though none of them have yet been anything on a level I'd say was worth of being considered a finished product. even the best of them would require heavy touch ups by hand to bring up to my quality standards. I'm very excited to see where this tech goes and how it can help me to start bringing these stories to life in a more timely manner. ----- Story wise this piece features Weeping Babette, a person whose exquisite genetics afforded her the chance to become an exceedingly massive female werewolf. The events of her first transformation were sadly tragic, like most in my worlds setting. Her twin daughters Mariette and Cosette ultimately met their demise at her jaws, when she entered a berserker state during her first transformation. She is now followed by the phantoms of her daughters to torment her nightly, she cries out for them, trying to bring them close but she cannot ever seem to reach them. Babette though we explore her tale only for a brief moment, its her dual blood line that eventually leads to our Red Riding Hood Orphan becoming cursed and sharing a similar hyper transformation. ----- The final pic is loaded up into the gallery already, along with the ultra high res files for copper tier Patrons and above, as well as the original .PSD available for Gold Tier Patrons. (For learning purposes only plz) :p Would love to hear your feed back on how it came out!  Became rather captivated by Midjourney recently (to put it lightly) and wound up spending a ton of time trying to learn its ins and outs trying to figure out how I can utilize these new tools in helping to speed up my work process. Sadly learned that Midjourney is highly restrictive with their censorship and seem to have no interest in making an uncensored variant. Got banned periodically several times already. Prompting me to upgrade my video card so I could get stable diffusion set up so I could learn to use that instead. Sadly Stable Diffusion prompt system isn't nearly as intuitive as Midjourney's set up is. MJ will work in some of its own style in fascinating ways that ensures you overwhelmingly get results close to what you wanted or wildly better. Allowing its users to make new discoveries for ideas and concepts they didn't even know they wanted. SD on the other hand, wont give you nearly so much if your not exceedingly specific with what you want. Positive accidents from what I've observed are much rarer and it takes a ton of work to get something even vaguely as good as MJ. But its local variant is free and uncensored. If you can figure out how to install it. Tons of guides on how to do so but they all seem to be uniquely bad in their own way. Will take someone with zero know how in this realm considerable effort to get the app running on your rig efficiently.
While I've not yet created anything that I could call a finished product I've gotten many concepts created that has me dying to render them into a full illustration. Take this new illustration I'm working on. I had accidentally created these 2 haunting women in a forest with Midjourney so I decided to take 2 of the better ones I had found there to work into a new drawing. I had hoped it would help speed up the process, but its for the now a still rather plodding process for me. You can't trust the AI's sense of anatomy either so you have to know how to improve upon it when its ever so slightly off, or intensely so.
Overall I'm deeply disappointed with both of these systems in different ways. But the tech is advancing so quickly that hopefully in a few months its much closer to what I need to make it sing. I also must apologize for not posting enough, the holidays, various appointments, and necessary computer upgrades have my slow output even slower of late. Hopefully I can get sped up once the new year roles in. Here are a few highlights I've come up with in the past weeks. Sadly I've not made anything in SD worth showing off in the slightest yet. and its been quite aggravating to me.
They are all uniquely bad in their own ways, but also all great seeds of ideas to pick up and fly with. I wish I could more easily learn how to capture some more of this essence into my own style. For the now I'm experimenting with redrawing the elements I like most and combining them in unique original ways. I'll make a space at some point in the near future to show off more of these. For each of these decent ones I've made thousands of variants that aren't worth a dam. They've definitely helped to spark new ideas tho that will be getting implemented into my overall world of Crimson Dames.
I especially love how these prompts can widely mutate causing some surprising results that with a bit of refining can be made into stunning new characters. This lovely for instance!!!
Sketch Reward for Emperor Nihilus!!!
 Sketch Reward for Caffeinated Critter!!!
|
|
RSS Feed
