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
© 2025 Daz Productions Inc. All Rights Reserved.
Comments
There must be hundreds of public domain photographs of that bust of Nefertiti. And it's sort of the most famous head shape in the history of the world.
Indeed, but we're also told it can do celebrities too. Although I've confirmed that V3D is safe, at least. It thinks Lana Del Rey is Kathy Newman with a nose job.
AI really is a catch-all phrase that carries a lot of negative baggage with it.
Personally, I'm of hte mind that it depends on the flow. Those goofy web models that generate images? I'd never pass one of those off as art in a billion years. I may use one to create a D&D character for when I play with my buddies, or if we're talking about somthing absurd in our WhatsApp, I may use it to visualize what we're talking about for a laugh. What I see some people do on YouTube (not naming names), but thinking that 'art' is writing prompts to get what you want. It isn't. You're just pushing random buttons to get random results.
Now, on the flipside, when it comes to generative features to help augment art... One case I agree with was a guy on YouTube who does wildlife photography. He said that it's not cheating to use GenAI if you take your photo of something, maybe the crop is wrong or something and you want to fix it. You can't just go out and take another picture of a snowy owl, it's gone. But, you can use GenAI to re-crop and fill in the image to match more your vision. You, the human still put in the grunt work of taking the photo. You, the human, just made a typical human error and cropped wrong, or maybe you had one vision in your head, but realized after taking it, another one comes up.
Honestly, we've been using AI ever since the advent of computers. A simple "If/Then" statement is artificial intelligence, artificial logic that takes an input, and generates an appropriate output. Using PhotoShop to adjust lighting, exposure, etc, is still using AI.
Now, a character shaping program? I worry that it will take off the plate of some really talented PAs. Would it be great if it could quickly 'tweak' exisitng models? Of course! Sure, it's all stuff that we could be dialling in in the menus...but why do cameras have automatic mode? I wouldn't see it as cheating if you're using a tool to adjust what you already have and make your flow quicker. If it can take that "80% of the work is perspiration" out of the mix, wouldn't our lives be better?
Lost one token. And even though I used different prompts, the characters look as if they're from the same family. No randomness to it.
I honestly don't think the PAs have got anything to worry about
It's not like we don't have a lot of products doing exactly these things with full control from the user and actual support toward the PA's.
Exactly. The thing is that it could take more time to prompt an AI to do an adjustment compared to simply using the product yourself and doing it. Its plugins like Yellow that ironically highlight the work the PAs do, and after experimenting today, I feel they're still in a pretty good position.
The "must be logged in and online to use" was already a hard no from me. The bit where you have to use up a (usually paid) token just to roll the dice on whether it manages to get anywhere near what you're thinking of... that's another layer of nope.
Exactly! As a pretty visual person, I just know I would spend time trying to figure out how to describe what I want, only for it to not match what's in my head. In that time, I could have found a nice mix of shapes from products I already have, or perhaps looked through wonderful PA catalogues to find a good option. Half-joking: I wish the language-interpretive effort here was instead helping to make the search options in the store work better. (I think of DAZ as a grocery store, not a restaurant. Help me find the shapes & ingredients I am looking for so that I may select them, rather than serving me a recipe I may or may not want, and had, after all, intended to prepare & cook myself.) But I do understand there might be some fun in seeing what it comes up with, just probably not for me :)
Same. I am heavily into AI and many of the online AI apps require a purchase, which is why I use stand alone AI apps. I can generate freely until I get what i want. Having to purchase toekns each time i want to use Yellow is a no go for me, just more attempts at keeping creativity behind a pay wall
Yellow claims to be trained ethically with DAZ assets, but I can do that on my own without the help of AI and personally I want unique characters that DAZ morphs can't create, Any how does it handle HD morphs?
I get that DAZ wants to get on the AI bandwagon and I am all for it, but it seems like the people that are making the decisions are really way behind the curve, especially after my exeriences with DAZ AI Studio compared to other AI apps. DS users need AI tools to make their renders better more than anything IMO and which would be easier than coding AI to create everything in DS for a user.
What we really need is a stable release of DAZ Studio built with the current version of Qt and feature parity between macOS and Windows. And maybe a Linux version. But, as far as Yellow is concerned ...
I thought I saw somewhere that it would create textures but it seems to only create a model. Is that right? Because if it is then wouldn't it have been better to release a tool that autocreated textures first? Textures really help differentiate between characters. I've tried to put the breaks on buying characters because when I compare a new releases from a vendor and the previous works I already bought from that vendor they just don't seem very different all too often.
Okay that sounds like a more useful AI prompt system. It would be like describing someone to a sketch artist. Example, make the nose more pronounced more regal, the eyes need to be less shifty. BUT tokens wouldn't work for that creation model because of all the iterations for refinement. You would have to pay a final fee when you accept the final "sketch" and it yields a morphed character to you.
oh, I missed this post and started my own thread. sorry...
oh, I missed this post and started my own thread. sorry...
you could move your thread to the Art Studio forum (by editing your first post) since it's your personal blog on using the product and nothing wrong with that
I tried but the sub headings don't go as far down as to the specific fora fro Yello.
My experience with Yellow is here: https://www.daz3d.com/forums/discussion/722996/i-tried-the-yellow-ai-to-make-a-male-k-pop-idol
Here's my tests. I have no intention of using this, I've dabbled with AI in the past but I wanted to see what kind of variety of morphs Yellow could handle using a range of famous faces. So I just did one prompt for each of 10 different past and present Hollywood celebrities with what I thought would be recognizable facial features (depending on your age, maybe not). Probably not the ideal way to do it, since I didnt do any tweaking of results at all. Just 10 different faces omne after the other. And I only own 1 G9 character, I'm too heavily invested in G8, so I just threw on some default materials from the free material sets included in DS for Gen 9.
Prompts that I used below.
"Audrey Hepburn at age 26"; Brad Pitt how he looked in the movie "Fight Club"; Danny Trejo from the movie "Heat"; Jo Raquel Welch, how she looked at age 26, in the movie "One Million Years B.C."; Lance Reddick; Actor Paul Walker, caucasian male,"The Fast and the Furious"; Rober DeNiro, how he looked in the movie "Heat"; Samuel L. Jackson, Sean Connery at age 34, how he looked in the movie "Goldfinger"; Actor Sylvestor Stallone as the Italian boxer Rocky Balboa.
Needless to say I won't be spending any more time with this. If I use AI at all in the future as part of my workflow, it will be to inpaint my own renders, fixing hair, or correcting bad bends in the thighs, upscaling, etc. I don't want something to do the creative part for me. The creative part is why I do this. I've been playing Krita, they have a plug-in for AI that looks like it is going in the right direction for using AI as a tool and not a replacement for creativity.
You know, I'm not inherently opposed to this.
If I'm being honest, my intuition tells me that it's gonna be a big ol' nothing, but I could be wrong. It could be something.
Haven't tried it yet, curious to take it for a spin.
@FrankTheTank: these results are extremely underwhelming. To be fair, the texture plays a huge role when it comes to making a character recognisable. But still.
...that makes my decision much easier as I am still on 4.21.0.5 which is the last version where Iray supports Win 7 for GPU rendering (not going back to glacial CPU rendering)
I too have a fairly diverse stable of base characters as well as an extensive library of Morph, Shaping, and Merchant Resource content. for G2 - G8. I've been creating custom characters since the Gen4 days, which primarily is for my own story lines, illustrations, and RP game campaigns rather than known celebrities Like someone mentioned above, a character has to provide something unique for me (for what I cal my "gene pool") to purchase them. It isn't unusual to combine body and head shapes of several different base Daz chanters for the base "figure" to which are applied additional morphs and shaping.
Yeah it is often painstaking work but the results are worth it. As I am pretty bad ar at writing prompts I don't think trying the same with AI would come close to what i can do in Daz with the tools I have. Then again, I'm not working at a "pressure cooker" production job with a tight schedule with the dreaded "deadline doom" hanging over my head.
I'm sure that given time, AI tools like this will improve But i do agree it takes the craftsmanship (which I find is the enjoyable part) out of the process.
[...well that was interesting, just after I clicked Post Comment,, I got an "Internal Server Error 500" instead of the usual 502 Bad Gateway message.]
Nevermind the GPU/CPU! What, in God's name, are you still doing on Win7?!?!
:: dramatically points to Win10 ::
Have you at least considered upgrading??
It really doesn't work that well. I made some attempts of ethnicities missing from Daz's assortment. Not great, though. I wouldn't buy tokens as it isn't good enough.
Yeah, i just posted those so people can take a look-see without wasting time with it. I didn't even save those results as characters when I was done. I uninstalled the Yellow plug-in already and jumped back into Daz 4.16.
Anyway, here's my own attempt at a "Brad" just by spinning dials for a few hours with G8M, he is still a work in progress. But I think I beat AI by a mile if I do say so myself.
I agree completely ^^
It would seem that generative AI is a direct threat to their business model as the majority of their users create still renders.
At this point AI can produce still renders of a quality that far exceeds what one can do with Daz content & Iray renders.
Daz is trying to incorporate AI in a way that keeps users inside their eco system buying content,but here is the problem.
1) They have to monetize their AI with Tiers/fees which could work if their AI solutions were not light years behind
other online AI apps in quality or versatility.
2)The newest most powerful GPU hardware is optimized for AI computing hence people like @FSMCDesigns
running local AI apps without paying any fees.
Even if “Yellow” was producing amazingly unique results that were not possible with old fashion morph dailing/sculpting you are still stuck rendering them in a version of Iray that does not even support modern day RTX GPU’s giving existing user seven more of an incentive to just pay for a Midjourney subscription .
3) Finally, At this point what is incentive for NEW users who can afford the latest RTX GPU to buy into the Daz eco system with IRay.
What is incentive for NEW users who canNOT afford the latest RTX GPU to buy into the Daz eco system at all when he can get such amazing still image results with a service like ,Midjourney or Leonardo AI etc..