Daz 3D is part of 
Connect
DAZ Productions, Inc.
7533 S Center View Ct #4664
West Jordan, UT 84084
Licensing Agreement | Terms of Service | Privacy Policy | EULA
© 2024 Daz Productions Inc. All Rights Reserved.
Comments
Is AI killing the 3D star?
I don't think, AI art is perfect for the people who lacks of patience, talent or creativity like me for example.
AI fails to do Sitting poses or Feet/Hands, you can't also create a scene with interactions between several characters.
3D art didn't kill 2D art, just AI art is another trend. It will keep on attracting people, talented 3D artists won't stop making us dream with their creations.
If you have Poser and the new plugin I've seen at R'osity, it is indeed possible to do things like proper hands, feet, sitting poses, character interactions and all that good stuff in Stable Diffusion running on your own computer. I am in hopes Daz Studio will have a version, or something similar. Only reason I haven't bought it is because it requires an upgrade to Poser 12 or 13 and I'm not sure how much sledge hammering is required to get newer Genesis figures working. Got tired of doing that a few years back, which is why fond as I am of my PoserPro 2014 (Poser 10), I made the switch to Studio for most of my 3d art.
Don't make the mistake thinking AI has gone as far as it can go.
Of course AI art is cheaper it's a big factor.
A lot of people can't afford to spend thousands of dollars in Daz3D / Poser content and in an expensive computer.
. yeah literally "art by numbers", like those old "paint by number" sets..
Great ideas to implement in Daz Studio.
Yes, I see it similarly. I want to do art primarly for me. I see AI as an extra tool: It is super for doing common stuff, background graphics, first inspirational images and so on. I let AI do the grunt jobs which I don't need so much details. But there are details that my brain is constantly nagging me about which I just have to do myself. AI is a gap filler, but the things that are important to me I want to do myself. And thats what AI can't do - be consistent with exactly these details. And AI can't do stuff that is uncommon... just yesterday I wanted to create a preview image for a character that I'm currentlys creating.... this character has purple skin color. Something that is just a little RGB slider is impossible. I think something like DAZ Studio, MBLab, Koikatsu etc. is just the goldilocks software for character creation. You don't have to start from ground up like in Blender (but can use it to add details) and you have still much more control that in Stable Diffusion (DAZ Studio morphs). I think parametrized humanoid creation is the way to go for people like me who don't want to spend 5 years learning human anatomy :)
https://bostondynamics.com/blog/robots-that-can-chat/?ref=futuretools.io
this terrifies me more than any AI art generator does
To be honest good AI art is comparable to good 3dart, and really good AI art is indistinguishable from a photograph ( which should please the 3dartist who were striving for photorealism no end ). The thing everyone seems to forget when they go off on some rant on how AI is putting "real artists" out of work is that:
1. The public doesn't care. They see art they like they buy it.
2. If you're making art because you love it, AI shouldn't bother you one iota. That's like me as an artist getting bent out of shape because Robin Eley exists and is obviously getting more attention than my work. The existence of any other artist DOES NOT prevent me from making my art.
If someone likes 3d, then they will continue to work in 3d. Just like you have photographers that still use film over digital. If the realization of your vision as you picture it in your head is your ultimate goal you will use whatever tools help make that a reality. The end result is all that matters -- that is unless you really like the process as well.
Making a living as an artist has NEVER been guaranteed. The term "starving artist" didn't just spring up when AI became a thing, but yet a lot of these people are whining like it was. They whined when the tools of photography became affordable to the average Joe. They whine whenever somebody that loves what they do gives their stuff away cheaply because they don't do it for the money, they whine whenever somebody comes along that is really good that didn't spend years and a fortune to reach their skill level. It's genuinely exasperating how much whining wannabe artists do, because they seem to focus on everything other than the art itself, which is the way grifters and opportunistics view things and not real artists.
Hello Lovely people!
I can't help but want to chime in on this topic because I hear everyone in every corner of media stressing about it. Whether in music, 3D, copy, video... AI certainly is going to change the way we create art. The legal and ethical questions are valid, but won't stop whatever is going to be. I'm going to continue to create art, because I need to. And I beleive in my heart there will always be a need for humanity in the creation of art in all it's forms. As good as AI is and I can only imagine how good it will get, it still isn't human. Never will be. I do think there is something special about human art that will always be needed and appreciated. There may be new ways to create and consume art, and new ways to profit from it, but I just hope I can find a way to keep up because it's all changing so fast, and I'm starting to get old! (LOL)
The potential of this tech is disruptive for sure. It does depend on regulation and general stupidness, as well as negligence, for how this will turn out.
It can yield "results" fast, and in comparison cheap, though not really for free, and training large networks per se, remains to be very expensive. It can be scaled to nowhere in a whimp. Fast + scale resembles the social network property, probably mind that for a comparison. Who sits on data, what can be used, and who can afford to train what with it... is not fully negotiated, but we've seen different approaches to those question in practice already (plus one or another lawsuit).
Since you might be able to mimic any other artist's works, once trained on their works, this has more disruptive potential than just to be fast and scaling well somehow.
Further it needs input to exist, but it's output kind of contests the art that has been used for input, which is a fun consideration to have, given how humans have been working on better weather for aeons.
Just for throwing in a short version of stuff that i haven't forgotten to mention in this post.
I recently dabbled with a couple of online AI text-to-image generators, but decided that when you have to pay for tokens in an app that might take dozens of attempts to get the image that you want, it wasn't worth it. I'm not saying that companies shouldn't be able to make money (especially when you are accessing their servers) but being charged for every iteration feels a bit unfair.
Anyway, I found an offline version of Stable Diffusion that works locally and had a play. Even as a hobbyist of 3D art, it is at first a bit concerning just how easy it is to type a few words and click the "make art" button. It is also concerning when you wonder where all of the images used come from. That said, there are a lot of limitations to balance out the advantages.
Advantages:
It's fast - I can generate four images per minute, change the prompts to refine the image and then run again.
The lighting quality is easy to get right compared to tweaking lights and running test renders in DS.
Disadvantages:
Consistency and repeatability. When you create your own scene or characters in DS (or any similar program), you can change the scene slightly and re-render. In DS you can take your character and put them in numerous scenes with the knowledge that they will look the same in all of them. With the AI text to image, you might come back to an older image and want to repeat part of it (such as the character), and find that you can't get quite the same result on the next attempt.
Specific characters - unless you are wiling to train your own models, you are somewhat at the mercy of others in what images are used. I was creating a series of superhero characters for an education project. Most styles of characters were fairly easy to get, but a black version of Dr Strange took a lot of attempts, presumably because the model was more biased towards asian and white characters. With DS or similar, you have full control over the character.
It's not really an artificial "intelligence", as evidenced by some of the truly stupid results that it generates, which can be frustrating at times.
Overall though, I do see one area where AI text to image generation could complement Daz Studio - the generation of background images. If the AI image generation thing doesn't fizzle out, perhaps in the future there could be a module for DS where you set up your characters and main, close-up props, then type in what you want for the background and the program generates it. I'm going to have a go at trying something similar with billboards and AI-generated background images to see how easy it is to match colours / styles etc.
I love stable diffusion, SDXL is pretty great. One of the top ten reasons I ended up gettin a 4090 is so I can locally finetune models on my PC. Another tool for my artistic toolbox. In the circles I am in, mostly the people bashing AI, are the same ones that bashed photobashing, or using daz3d renders. So I really pay them no mind. Not worth putting any time into debating with them. It goes in one ear, out the other. Daz3d still got it's place, although it's a lot less prominent in my toolbox. I will do a quick render to set up my composition is one use that is common in my works. I don't have to spend so much time getting out all the noise anymore though, which is a huge plus on indoor scenes.
LOL Automatic 1111 stopped working on my PC so after 3 days of deleting the venv folder, editing webui.sh and bat files I gave up
I need my iGPU too badly and won't disable it in BIOS so I cannot get it to install the correct pytorch wheel for my Nvidia card
no other workarounds I have Googled, read on GitHub, Stack Overflow of Bing Chat has tried to suggest works
I can still use Fooocus and NMKD for still images, will miss making my Deforum videos
so back to doing mostly 3D CGI videos with some AI backgrounds
Damn thats crazy video lol. I am mostly dealin with stills, that is some good stuff
thanks
I will post in other thread
...one of those old school photographers who still prefers to use film here.
Would love to get my hands on one of those old press cameras which use the 4"x 5" film packs.
Great videos, Wendy
Will have to try animating using ComfyUI and SDXL models.
Great, that I do not have to download again checkpoints, Loras and the other stuff.
Because I use Linux and Nvidia graphics card with 8 GB of VRAM.
ComfyUI works much better, than Automatic 1111, on my system.
Example cats generated with ComfyUI
Flying dolphins, why not...
And what about the Llamas...
And the Panther...
I have one of those in concrete
@Artini, are you expressing an opinion with these "artworks"? I don't get your point, unfortunately.
No, I just show off the examples of what ComfyUI can create.
Wish, I could create a similar images in Daz Studio, but so far I have not succeded, yet.
Oh ok, by all means carry on;) Now can we have a piglet pretty please?
...a DS version of flying dolphins, made long before AI was a thing lol..
I adore your art during a long time of my adventure with Daz 3D, Sven Dullah.
A 3D renderings is my hobby, and I am trying to use many tools available to me,
including AI.
Glad you asked,
so here comes Happy Piglets