-
Sign up or login, and you'll have full access to opportunities of forum.
Art by Art
- Thread starter Art
- Start date
-
- Tags
- oiginal art
Art
Governor
Thank you, dear admins. That's exactly what I want. I am thrilled. Thank you again. The topic is now closed for me.
Also thanks to everyone who has supported me.
ianwarrick
Magistrate
wulf
God-Emperor
I'm sure we will when he's ready to share.
Art
Governor
Experiments - Interrogation by the inquisitor
The year has started quite well for me in terms of the number of releases. I have published 16 images so far. After that, I took a little break. Quite deliberately. I want to rethink my way of working and my style.
In my publications, I've noticed that more and more AI images are on the rise, especially on Renderotica. On RR this is almost inflationary. However, there is a large quantity, not always quality.
But what I have noticed with many AI images is that the images are very well color-coordinated in some themes. Especially in the fantasy genre. The results are wonderful.
You probably remember the wonderful CG artist Roadrunner. With his outstanding series "Malsavia 1335 AD", to name just a few of his works.
Roadrunner once said that about 50% of his work is 3D generated and the rest is postwork.
At the time, I didn't see it that way, because with DS and the possibilities it offers, postwork could be dispensed with in many cases.
Now, however, I have decided to take a little more time to post-process my images. I want to rethink the style of my images. I want to achieve the visual quality of AI images.
But at the moment that's the only thing I want to learn from it. At the moment, AI-generated images are in no way a replacement for our CG images.
I have created an image as a template for post-processing. See image 01 and my first result can be seen in image 02. I don't know yet whether I will like it in the long run. I will do some more post-processing.
The year has started quite well for me in terms of the number of releases. I have published 16 images so far. After that, I took a little break. Quite deliberately. I want to rethink my way of working and my style.
In my publications, I've noticed that more and more AI images are on the rise, especially on Renderotica. On RR this is almost inflationary. However, there is a large quantity, not always quality.
But what I have noticed with many AI images is that the images are very well color-coordinated in some themes. Especially in the fantasy genre. The results are wonderful.
You probably remember the wonderful CG artist Roadrunner. With his outstanding series "Malsavia 1335 AD", to name just a few of his works.
Roadrunner once said that about 50% of his work is 3D generated and the rest is postwork.
At the time, I didn't see it that way, because with DS and the possibilities it offers, postwork could be dispensed with in many cases.
Now, however, I have decided to take a little more time to post-process my images. I want to rethink the style of my images. I want to achieve the visual quality of AI images.
But at the moment that's the only thing I want to learn from it. At the moment, AI-generated images are in no way a replacement for our CG images.
I have created an image as a template for post-processing. See image 01 and my first result can be seen in image 02. I don't know yet whether I will like it in the long run. I will do some more post-processing.
Attachments
Dafmads
Magistrate
Thank you very much for your work, for sharing your art and for exciting us with your great visualizations.
And for your effort to improve (what is barely needful I.m.o.)
In the first picture, the transition from lighted areas to shaded ones is quite hard. Whereas the general lighting seems to work for most of the daytimes.
I prefer the second one, with the ‘warmer’ colors. It might refer to the early morning light or to the ‘golden hour’ near sunset.
But both are great- thank you!
Art
Governor
Prologue
I have created another AI image. "A western journalist in Afghanistan".
The advantage over CG images is that there are no significant costs. With CG images there are costs for the figures, hair, clothing, environments etc...
The disadvantage of AI images is that you can't make series because the AI does what it wants. If a character is described and created in one prompt, it usually looks very different in the next creation.
This picture here consists of four separately created pictures. These individual images were put together in a photo editing program. The interrogation room was the basis. The basis for the woman and the Taliban warrior were CG pose pictures. The facial expressions, skin, hair and clothing were created by the AI according to the prompts.
One advantage of AI over other photo manipulations is that if you don't depict celebrities, you don't have any copyright problems. But problems, maybe with ethnic groups. Maybe.
I have created another AI image. "A western journalist in Afghanistan".
The advantage over CG images is that there are no significant costs. With CG images there are costs for the figures, hair, clothing, environments etc...
The disadvantage of AI images is that you can't make series because the AI does what it wants. If a character is described and created in one prompt, it usually looks very different in the next creation.
This picture here consists of four separately created pictures. These individual images were put together in a photo editing program. The interrogation room was the basis. The basis for the woman and the Taliban warrior were CG pose pictures. The facial expressions, skin, hair and clothing were created by the AI according to the prompts.
One advantage of AI over other photo manipulations is that if you don't depict celebrities, you don't have any copyright problems. But problems, maybe with ethnic groups. Maybe.
Attachments
fallenmystic
Tribune
petelobo
Magistrate
From what I've seen, AI can make women, can make some facial expressions. I have yet to see it make a woman with electrodes on her body, with whipmarks, with any kind of torture occurring or the results thereof except for very unrealistic blotches of blood splashed about on her. Am I correct about that?There are multiple ways to achieve character consistency with Stable Diffusion. Try searching for keywords like IPAdapter, InstantID, ReActor combined with "Stable Diffusion character consistency", and you will see quite a few guides like this one.
fallenmystic
Tribune
Fetish elements are indeed difficult to depict with a base model since it wasn't trained over any such images. Fortunately, Stable Diffusion models can be finetuned or supplemented (with "Lora"s) by users, and we have Civitai, the largest community where people create and share such models.
As for whip marks, there are several models including my own. Unfortunately, they are mostly for SD 1.5. But I heard it's at least possible to use them for SDXL with something called "X Adapter".
For depicting blood, I know there's at least one good model for both SDXL and SD1.5. As for electrodes, I could only find a single model for SD1.5. It looks to be for anime-style images, but it shouldn't be too difficult to convert it to a photorealistic style.
