Please remember to state if your entry is pure, open or non-competing.
“There are two main goals for these contests: Fun and Glory.
Fun: You get only 4 days to create a cool image, based on the theme for the week.
Glory: After 2 days of voting, a winner is declared. The winner picks the theme for next week and will be featured in the Feature Row.”
open (made with the help of ai, tho I have not used it just to make the Image because the Base like the Houses and hills and tree bases, fog and sky are made in blender, I used Ai just for rendering textures and light on the Image) I used stable diffusion automatic 1111 for the ai and git to download it, for the Image creation you will need the Controlnet addon and a depth map model that works with controlnet.
After that you will need a stable diffusion model that fits the style you are looking for.
Then open Blender create a Scene like you would normaly do but without textures,
after that render a depth map of the scene in the compositing tab.
Important is the Size of the Image best works 512x512 or 1024x512 or 1024x1024.
But its possible to resize the Image, for high res results.
For resizing I use the ultimate resize script with the Ultra Sharp 4x model for the script.
You can also tell the Ai how much it should stick to the promt.
So you can control what and how exatly you want your result.
Throw it into stable diffusion tweak settings write a prompt and hit Generate.
Its not quite as simple as that but you can find many tutorals on youtube
As I noticed the filename and quite a few strange artifacts in the image, I was just going to write you a message about this.
Depending on the amount of AI-contribution, it might be better to have this one in the “open” category. In order to keep the challenge fair for everyone, it would also be helpful if you could post the original Blender image as well. (And maybe describe the process/software used when creating this image.)
This way, everyone could form their own opinion when voting starts next week. (And besides that, it would probably be an interesting read.)
Thats how a depth map looks like
It shows the ai distances of object, so the ai can see where objects are.
Note: If you have curved Objects it can break the Image output
Thanks for updating the post and writing all of this down. It is a lot clearer now. And I guess, making the entry ‘open’ was definitely the right choice.
Btw. I tried using the height map as a displacement map, which worked really well (from that perspective):
Thanks for the reply.
Btw if I can still do Ai Images pls tell me, I well say that it is made with ai and make the entry open.
Tho if I dont do something with Ai, I am just gonna do a normal entry
If I am doing something with ai should I also send the promt i used?
Was gonna make Trump and Putin playing PS5 in a bunker, but realised I’m not good enough at making accurate likenesses.
And then tried making a T800 dance for a TikTok video, but realised I wouldn’t have enough time. Even tried making a few dolphins dancing for a TikTok, but again, gave up due to time.
Edit* Forgot to say why it’s open: ant mesh, phone, crack png, and house image were existing/downloaded. I made the ground/rocks, leaves, skull and gas mask (which are barely visible), and rigged and retextured the ants.
So far, there is no weekend-challenge-rule against using generative or any other kind of ai in the process of creating an entry image. In most cases it will probably make the entry open but there is nothing wrong with entering open entries. In the end, it is all about fairness, so adding a good description on how the entry was created is always appreciated. (Publishing the prompt used, is also a great idea.)
That being said, generative ai in art is a very controversial topic and hopes and fears might collide in rather energetic ways. (Not even counting copyright and other legal issues, which (in one way or another) apply to other entries as well.) Not everyone feels the same way about it. So don’t be disappointed if this leaves its marks in the voting.
Now that I think about it, I have a feeling that there might have been a policy against AI in the feature row. Unfortunately I can’t remember the specifics and searching for it didn’t really work out. But, maybe @joseph can help with that?
Work on the featured row must be predominately done in Blender, or Blender must be a major, or significant, part of the creation process. A purely AI-generated image would not fall into this description, and would not be eligible for the feature row. A case like this- where Blender is used for modeling and rendering, and AI is used for texturing- would be eligible for voting, but as you said:
Generally, works with AI-generated elements haven’t made their way through voting to the featured row. This is less of a policy statement and more just an observation on the current voting habits of the current regulars
Would be interesting to get one of the “AI” things to generate challenge themes. Then get one to generate entries for that theme. Not for the weekend challenge but as an experiment.
Update: I scorched the earth! It’s based on a NASA photo, with colour remapping on the shader to destroy all greens, plus glowing toxic greens in the ocean, and clouds with some glowing fire clouds, and an atmosphere of pure smog.
I also added a clean blue atmosphere to the ring.
Update 2: I removed the spokes because I like the composition better with a hole there. I added some thrusters, tweaked the lighting so the earth drops more into the background, and added some buildings and roads inside the ring! The terrain & buildings are a displacement shader on the ground, with various voronoi patterns.
Remaining nomadic civilization after world ending catastrophe finds a one in a million standing building from before the war, uses its resources, plumbing, and technology to maintain and grow a small culture :>
This is one of my first times using blender and had a lot of fun with this, thanks for the prompt!