• Sign up or login, and you'll have full access to opportunities of forum.

Fallenmystic's AI-assisted works

Go to CruxDreams.com
fallenmystic

fallenmystic

Tribune
This is the place to share renders I've made with Stable Diffusion, an artificial intelligence (AI) model that can be used to generate high-quality images in any style.

Unless otherwise noted, all of the images I post in this thread were made with AI.

Disclaimer: Most of my renders deal with fantasy slavery involving extreme racism. It has no bearing on real life, and I do not condone or tolerate racism outside my fantasy world.

The same goes for other themes like sexual abuse, objectification of women, and so on, all of which should strictly remain in the realm of fantasy.
 
  • please help us keep our excess usage charges under control,
  • so don't post byte-heavy images if they can be reduced without significant loss of quality;
  • observe the limit of 5 images per day in any one thread.
 
Madiosi said:
  • please help us keep our excess usage charges under control,
  • so don't post byte-heavy images if they can be reduced without significant loss of quality;
  • observe the limit of 5 images per day in any one thread.
Click to expand...
Sorry, but I'm confused. All of the images I posted above are under 2mb in size and there are 5 of them in total. Am I missing something?
 
Video generation is AI's current frontier where they are still clearing obstacles to make it feasible to claim the territory as a solved problem.

Still, it's at least possible to render a short, flickering video like the one that I made, which I attached below. Considering how much progress Stable Diffusion has made since its first public release mere 10 months ago, I believe we may also expect a breakthrough in this front in the foreseeable future as well.

The video is loopable, by the way, and contains an audio track I made from free sound samples.
 

Attachments

  • 20230430204535.avi
    1.9 MB · Views: 719
Another remaster of my early render I did today:

00821-20230621212105.jpg

Edit: I only just noticed a few spots that I missed while fixing the problems, like the necklace or that unidentifiable creature inside the window. I’ve spent too much time on it already so I’ll just leave it as it is for now.
 
Last edited:
fallenmystic said:
Another remaster of my early render I did today:

View attachment 1316893

Edit: I only just noticed a few spots that I missed while fixing the problems, like the necklace or that unidentifiable creature inside the window. I’ve spent too much time on it already so I’ll just leave it as it is for now.
Click to expand...

Here's a new version fixing the problems from the previous image:

00895-20230622065054.jpg

And here's the list of changes:
  • Made the white girl smile slightly to highlight the contrast. I felt her expression was too gloomy.
  • Made the hotdog a bit more prominent, after someone said he mistakenly thought the slave girl was looking at the white girl's clothes.
  • Fixed the disfigured necklace.
  • Killed the creature behind the window, whatever it was.
  • Made the skin of the slave girl more grimy and detailed.
When remastering this render, I initially only intended to fix a few glaring problems. But as I worked on it, I decided to use it as an occasion to test my new workflow.

It turned out pretty well, enabling me to make a quite realistic and detailed render as shown above. But a technical issue prevented me from determining the overall tone and lighting condition as I intended. It's unfortunate because it made me actually prefer the original version in some aspects.

As such, I decide to make another remaster of the original with minimal edits to fix only the glaring errors:

00905-20230622071928.jpg

I haven't yet found a solution to the technical issue I mentioned above, but I think I can use a workaround. We'll see how well it works in my future renders.
 
In hindsight, I feel the added grit on the slave girl’s body is way over the top. I should have lessened the amount of change (I.e. denoising strength).

But I won’t spend any more time on it because I feel it to be a lost cause with the lighting problem anyway. At least I’m happy with the second minimally remastered version.

There are many more renders I’ve done since last year which contain interesting ideas but bad in quality since both the technology and my understanding of it weren’t at the same level as it is now.

I want to remaster some of them before I start working on new stuff. So, I’ll move on to a different render next time I post here.
 
Normally, I don't care for AI renders. They have always seemed...off. And, I generally don't like the way they tend to steal others work, and cheapen the various artforms.

That being said, what you are doing here is great! Extremely well done. It's realistic, isn't taking away from another artist (as far as I can tell), and is allowing you to personalize images that might otherwise be difficult to achieve.
I applaud you!
 
Leatherchain said:
Normally, I don't care for AI renders. They have always seemed...off. And, I generally don't like the way they tend to steal others work, and cheapen the various artforms.

That being said, what you are doing here is great! Extremely well done. It's realistic, isn't taking away from another artist (as far as I can tell), and is allowing you to personalize images that might otherwise be difficult to achieve.
I applaud you!
Click to expand...

Thanks much for the kind words! I think the way AI enables easily generate low-quality images has given rise to a perception that it's all that it has to offer.

IMO, it's not too different from how Daz3D has opened up doors for many who have great ideas without 3D modelling skills to express their creativity. Even though a lot of Daz3D renders found on the internet tend to look alike, we have a few great artists using the tool whose style is readily recognizable.

It's not too different from how people use AI nowadays, as we have a relatively small number of people who are making truly amazing art with it while the rest is flooding the internet with low-effort renders.

In the end, I think it all depends on what you want to achieve with the tool and how well you understand it.

And I decided to make photoreal kink images - be it "art" or not - with it, and I'm quite enjoying the process. :)
 
Here's the latest remaster of one of my earlier renders:

00307-20230624125855-captioned.jpg

Because I felt annoyed that I couldn't control the tone and lighting as usual with the last render due to the aforementioned issue, I used a different workflow and focused on making the style exactly as I intended this time.

In particular, I wanted to make it look like a restored historical photograph and I feel I achieved the goal reasonably well.

However, I struggled with the upscaler this time and couldn't make it work despite having spent more than a day in an attempt to fix the problem.

So, I'll just leave the original resolution as it is for now and try to revisit it once I learn why it didn't work as intended.
 
Last edited:
RedOrc said:
I am glad to have you back posting here again. Congratulations on turning AI from "The Death of Art" to "A New Way to Create"!
Click to expand...
Thanks much for your kind word. But I don’t feel it fair for me to take such a credit because I’m a mere early adopter rather than a pioneer of the technology.

I would feel happy, though, if what I post here may help encourage others to become more open-minded towards AI so that they could benefit from it in expressing their creativity or enhancing their art like have.
 
You must log in or register to reply here.
Back
Top Bottom