What happens to vendors now, we done?!

Padone

PLEASE EXPLAIN HOW TO USE AND WHAT IT CAN DO THANK YOU

Seem to me that all it can do is AI generate single images, at best they can be used for backgrounds if you don't need multiple views of the same subject.

As a storyteller, I start with some environments and characters, which need to interact on progressive shots. That's the very basic. Now first I don't see how it is possible to AI generate a full environment then reuse it in different shots, with different camera angles. Same I don't see how it is possible to AI generate a character then reuse it in different shots, interacting with the environment and the other characters. Thank you.

Squishy

this is so direly harmful to DAZ's whole business model lol

vectorinus

Daz studio does not alienate its clients, but tries to attract new ones. Currently, 3D companies, as well as 3D modelers, are experiencing severe pressure from visual AI, just as programmers and some other professions are experiencing pressure from Chat GPT. An experiment using AI is not a fad, but a desperate attempt by DAZ to stay afloat and keep the software free. In this way, DAZ continues to take care of its customers and hopes for their loyalty in this difficult time. Stand with Daz3D now!

foleypro

I must say this is a bummer and a blessing..

I have been a Vendor in the past for DAZ and other sites...been doing realistic 3D since 1995 when I got a donated tandy 1000 a glorified processor...So that said get used to it...the wave of the future is going to be prompts with pictures using different engines..pictures attached same prompt different engines...

vectorinus

I do realistic 3D... This is exactly the most vulnerable spot for every 3D artist! This is exactly where the insidious 2D AI hits. Look at these two pictures. If we were to put together a scene like this, when rendering we would try to make sure that all the content looked crisp and clear (even the background). We want the lighting to look realistic and even. But a prudent AI algorithm blurs the background, highlighting the main characters. And the background also benefits because it becomes more mysterious and unclear. It seems that it contains even more than what is actually depicted. The viewer's brain imagines something that is not actually there. But it's exciting for the viewer. Look at the lighting of objects. It's clearly not believable. But we are not outraged by this. Because people actually like implausible entities (as long as they are not dangerous). To further enhance the magical lighting, the AI darkens (also implausibly) some objects and even some parts of the main characters. But we would try to illuminate all areas of our three-dimensional scene evenly, right? And thus we would kill interest in her! And besides, many 3D artists are weak in the plot; they use standard, established situations, and corresponding poses and facial expressions of the characters. Therefore, such stories are uninteresting. They are sterile and lifeless. They don't attract viewers.

Squishy

vectorinus said:

Daz studio does not alienate its clients, but tries to attract new ones. Currently, 3D companies, as well as 3D modelers, are experiencing severe pressure from visual AI, just as programmers and some other professions are experiencing pressure from Chat GPT. An experiment using AI is not a fad, but a desperate attempt by DAZ to stay afloat and keep the software free. In this way, DAZ continues to take care of its customers and hopes for their loyalty in this difficult time. Stand with Daz3D now!

daz's software being free has always been a vehicle to market paid content (and that's okay!)

having seen how heavy gallery posters on sites like renderosity have been using generative AI to flood those galleries with this stuff (which is itself another vehicle for selling paid content) I think this is a terrible idea, I think it's going to attract a lot of marketplace consumers away from the stuff that actually makes money for DAZ. chasing $4/month subscriptions lol

vectorinus

We will learn from AI how to render beautifully and become better. In fact, there is nothing impossible about this. For a long time this was called post-processing rendering. We should take care of this so as not to be left behind.

Masterstroke

vectorinus said:

We will learn from AI how to render beautifully and become better. In fact, there is nothing impossible about this. For a long time this was called post-processing rendering. We should take care of this so as not to be left behind.

i am already sick and tired of this AI-look.
(Video has NOT killed the radio star) 

plasma_ring

vectorinus said:

We will learn from AI how to render beautifully and become better. In fact, there is nothing impossible about this. For a long time this was called post-processing rendering. We should take care of this so as not to be left behind.

I actually probably owe generative AI a debt of gratitude on some level, because it made me realize that I'm not happy unless I have complete control over my work and the skills I need to depict exactly what I mean to. That's why I enjoyed using renders as a base for paintovers to begin with, because I could make those artistic choices the AI is sort of haphazardly simulating. But it's going to be like any kind of artistic Photoshop filter; you can only make sure your renders are composed, lit, and narratively interesting like paintings by learning how to paint. You can only make sure you're depicting exactly what you want to by learning how to model.

I recognized in some AI users the exact creative bargaining I did when I first started 3D; what I did in Daz was better than anything I thought I could do on my own, so it was good enough even if it wasn't what I would have done if I had the skills. That look in the mirror dropped me from being content with improving and personalizing what I was doing over time straight down to being completely uninterested in cutting corners until I'm skilled enough in fundamentals that I don't have to.

I do kind of wonder how many people will make a picture with gen AI for the first time, follow a winding road to learn to manually edit the things they're dissatisfied with, and then realize five years later that they did about three times the work of someone who tackled it head on, but with less of their own output to show for it.

wolf359

As a storyteller, I start with some environments and characters, which need to interact on progressive shots. That's the very basic. Now first I don't see how it is possible to AI generate a full environment then reuse it in different shots, with different camera angles. 

 

 

Right now the closest you could get to this is AI generated 360 degree “skyboxes”
similar to the HDRI’s we already have for DS or used in game engines.
obviously very limited for animation ,of course, but at least you could rotate your camera around on the X and Y axis and get different shots

Padone

wolf359 said:

As a storyteller, I start with some environments and characters, which need to interact on progressive shots. That's the very basic. Now first I don't see how it is possible to AI generate a full environment then reuse it in different shots, with different camera angles. 

Right now the closest you could get to this is AI generated 360 degree “skyboxes”
similar to the HDRI’s we already have for DS or used in game engines.
obviously very limited for animation ,of course, but at least you could rotate your camera around on the X and Y axis and get different shots

Thank you, how do you tell the AI to do that, is it enough to put in the description "generate a HDRI map" ?

Then of course I wasn't thinking of animation, but more of comics thus single images. But it seems to me that AI just can't do that, becuse it can't "remember" a definition and reuse it in another contest. As per my original question, in comics we need characters interacting among them and with the environment.

wolf359

Thank you, how do you tell the AI to do that, is it enough to put in the description "generate a HDRI map" ?

Then of course I wasn't thinking of animation, but more of comics thus single images. But it seems to me that AI just can't do that, becuse it can't "remember" a definition and reuse it in another contest

@Padone
You need  a specific type of  AI generation engine to make 360 Skyboxes
I made a 50 minute feature film last summer in which all of the backgrounds were the 360 degree Sky boxes from this site
https://www.blockadelabs.com/
I hoarded over 150 of them BEFORE they put up a paywall
and yes they work in Blender Cycles as well.
They are best suited for “Stylized”or  NPR types of Art & animation IMHO


Here is a 4 minute clip from the film

 

 

Certainly this would work for still comic panels I just created this with the iclone toon render engine,and some
of my 360 degree backgrounds and cobbled up. some quick comic art with "Comiclife" in about 15 minutes.

If Daz could work out a way to generate their own 360 skyboxes ,with AI, they could even sell them as "quick environments" for Iray

“some people” already are.
https://www.renderhub.com/anabran/sci-fi-hdris-for-daz-studio-iray

junk

To me AI has really come a long way and while it's cool and exciting it has also devalued so many of the things us regular humans can do.  That devaluing has made me feel "what's the use" in trying to create something I'm proud of and want to show others.  Because now we're talking a comparison between what the lamen see with AI (often in seconds) and what you might have done with the help of Daz3D (often over days of work) in this case. 

With the foundation of what I just wrote, I'm trying to say that me as an end user/artist have lost so much of the wind that would drive my sail.  I know.. waa, waa.  This though translates into I haven't bought anything since Daz3D has introduced "Daz AI Studio" nine days ago.  That's not normal for me who was buying something every single day as an average.

So yes... there are people not buying as much with the embracing of AI on Daz3D.com and I'm not alone.

Played and uploaded images back in 2022 of Stable Diffusion altering my Daz imagery but then decided to not use it at all.  So maybe my feelings will change on AI and will eventually fully embrace it as it is inevitable. 

JasmineSkunk

I haven't been around in a few days so I'm just catching up with everyone. But I'm surprised to see and hear some reactions. I don't know if my words will be helpful or... something other, but I hope they're seen as truthful. Or at least, my perspective of it. The truth is that digital art has changed with AI, and there's no going back, nor standing still. To those saying, "I'll still buy assets", I say, perhaps you will, but - what was once a lucrative skill set, has now been turned into a zero (or at least, minimal) sum hobby, sorry to say... And that is because what once took months, or even just days, or hours - now takes only mere minutes... And those who were "buying" those skill sets from artists, are now able to create something that's better than just "good enough" with a few sentences themselves with no art skill - and no budget - whatsoever. They're not going to hire artists to make ads or other marketing when they can do it for free themselves. Why would they? It's nothing like NFT's (which was an attempt to place "real" value on intangible digital art. It was an effort to make digital art more like "gallery" art that could be traded and sold as a unique entity. This is just not that. This is a tool. It replaces old tools. Like the automobile replaced horse and buggy and lightbulbs replaced oil lamps and candles. Sure! There are still people who own horses and people who still love candles, but they're no longer the "industry standard" and have become luxury hobbies or interests. So too will be the way digital and 3D art and copy have been created until now. The simple reality is this is the "baby" version of AI and it's only going to get better from here forward. Now... To those who say, "yeah but cameras didn't stop painters from creating art", I should mention that while this is "true", the vast amount of people buying "art" for their homes are also - not buying it from independent artists, right? Many, maybe even *most* get their home furnishings from Ikea or Target or wherever. Artists have the majority of their own collections and the average human decorating their personal space turns to mass markets more often than independent artists in today's consumer driven economy. And that's just true. Now, I'm not suggesting you stop creating art, or using tools you enjoy in a medium your comfortable. I'm certainly going to continue to do it - simply because I enjoy it - but things have changed. Absolutely they have.

wolf359

The cost of creating  usable visual art is now much lower for everyone.

Eeyore

Having read some of the comments earlier today I can see where this can be harmful to some vendors, but I can also see where this might open the door for some with a willingness to be creative, detailed, and willing to broaden/narrow their scope of offerings.

I firmly believe CG art is that which someone creates uising a computer, not someone telling a computer what to create.  I suspect I am far from alone.  And, yes, I dabbled with the AI studio quite a while today.  There's some bugs that need to be worked out, too.  I can't tell you how many images I had with characters have been short fingers, having extra fingers, arms and hands/feet and legs intersecting eachother or blending into objects.  At least one character had two different eye colors, and another was cross-eyed too.

Anyway, the one thing I noticed is how the AI generator seem to stereotype some of the characters.  For example, I tried to create a native American woman to try and visualize what my grandmothers may have looked like.  Well, every one had to be decked out in headdressed, jewelry, and feathers.  Even when I tried to change some of these things, something was always still present.  You cannot make a realistic image when even AI shows an unrealistic stereotypical bias.  However, what I noticed is how beautiful the outfits, jewlery, and feathers were.  Then I ventured into other realms for other characters, and the details were there too.  What a fantastic opportunity to vendors I thought.

Sure, some people may opt for AI over using skill to create images, but the opportunity to create more realistic content, historical or contemporary could be something people value.  I mean, there's a lot offered now, but also a lot that is not truly realistic contemporary clothing.  Sure, there's some, but not a tremendous amount of it. Possibly some of this comes about from limitations and the requirements to create and render it all; but does this sound like something that may help the vendors that are questioning the future of the marketplace?   I say this because there's people like myself that like to create our images the old fashioned way, with our own vision, and I want available content to continue doing that.

Padone

Honest, can't understand what all the fuss is about. It is my understanding that the majority of daz users either make comics or animations. Then AI can't do either. AI is only good for single images, not related to a common contest. AI can't do comics. AI can't do animations. So the daz shop and vendors are pretty safe in my opinion.

wolf359

AI is only good for single images, not related to a common contest. AI can't do comics. AI can't do animations. 

@Padone

This is not correct ,you are a bit behind the times, on AI it seems 
Generative AI has had the ability to repeat Characters with your customized training data for over a year now
you can even  repose them with “open pose”.

The New Daz AI image system  effectively does repeat characters with M9 &V9 by default

Still the Artist must layout the comic pages & panels as always
unless they want to “roll the dice” and use this FREE AI comic page generator ( with or without dialogue boxes)
https://huggingface.co/spaces/jbilcke-hf/ai-comic-factory
This comic page below is based on a prompt, I asked chat GPT to write for me, about humans arriving on an alien planet.

But yes AI “animation” (even the new sora) still lacks enough control for actual directed filmaking …yet.

But AI can create motion data and export as FBX  but in this case you still need access to  software that can retarget the Data to your Character rig ecosystem such as Iclone or 
MotionBuilder or Blender(with addons).
So the market for canned  Character Motion in Daz ,.duf format is not really affected  by AI yet. 

So the daz shop and vendors are pretty safe in my opinion.

Except that there is already a  RAPIDLY growing demographic of AI art users who are creating amazing still art and illustration with Zero Need for investment in a powerful PC to run NVIDIA Iray or buy  Content from the Daz store

bohemian3

Like the introduction of all new technological innovations, yes, some jobs will disappear but new ones will emerge. I'm still new to this whole world of AI, but here are a few examples of new opportunites, and I'm sure there are many more.  

• We will need still need targeted, unique and distinctive genre specific content for AI generators and solutions for overcoming the shortcomings of AI (deformed hands, consistent imagery, generic or incorrent imagery, case study: man playing spinet)
• An up and coming fashion designer can take photos of a new garment from 20 different angles and work with a professional AI Content Creator to produce a LoRA (used in AI Engines), allowing us to use intricate, unique couture garments with perfect draping, overcoming some of the current limitations of cloth simulation
• Poses will still be a thing and we will need unique high quality curated reference art  - Stable Diffusion has cool functionality ControlNet allowing us to accurately control the pose of a figure and I look forward to see if DAZ AI Studio will offer more granular control of poses.
• AI still needs more accurate and reliable placement and control of camera angle, and we'll see what technologies emerge from artists to fill that void.
• Better control of shots containing multiple character.

Yes, a lot of those things are free already on the internet, but it's a Wild West out there, requires a certain amount of technical savvy, and we'll need reliable vendors to fill a gap for more directed and higher quality options in an accessible interface.  Like all freebies, you don't pay your money but you take your chances.  

At this point, the train has left the station.  I don't want AI to do all the work for me, only assist where my skill set and experience lacks, the goal of all technological innovation.

There's a huge difference between getting what you want, and wanting what you get.  While AI is amazing, its a deal with the devil. I'm no where near getting exactly what I want, but for now I will move forward with my projects living with what I get.

Padone

wolf359 said:

This is not correct ,you are a bit behind the times, on AI it seems 
Generative AI has had the ability to repeat Characters with your customized training data for over a year now
you can even  repose them with “open pose”.

Yes I am a complete newbie in AI, but when I refer to AI I'm talking about DAZ AI Studio that's what this forum is about. I didn't see those features you're tallking of in DAZ, or if they're possible then I wasn't able to use them in the free version. Or that's what DAZ AI Studio is missing.

p.s. With DAZ I even have issues getting basic anatomy, as the AI often comes up with extra fingers or missing limbs or a reversed head, go figure to keep the same character or make specific poses. So may be I'm not the only one being "behind the times" LOL.

Green Tiger Tora

I don't believe AI will completely replace the 3D industry in the near future. Both AI and human creativity have their own advantages and disadvantages – nothing is perfect at everything. It's similar to 2D animation, 3D animation, and other art forms. New, modern styles and improved quality constantly emerge, but they never eliminate the use of older styles. We still see frame-by-frame animation, 2D puppet animation, pixel art, drawing, and painting – each with its own set of benefits. Creators have more control over specific aspects in different mediums (A or B). Even with advancements, AI will likely remain somewhat limited. It typically requires a vast amount of data to generate new things, unlike humans who have greater flexibility and control in their artistic endeavors.

wolf359

I didn't see those features you're tallking of in DAZ, or if they're possible then I wasn't able to use them in the free version. Or that's what DAZ AI Studio is missing.

@Padone
Well this is the first iteration so hopefully Daz will add the ability to generate 360 degree panoramas.
it would drive alot of traffic to the Daz ecosystem now that.

https://www.blockadelabs.com/
 
is behind a subscription paywall.

thorsten_370e9b1e2e

I already had $200 worth of items in my shopping cart, but when I saw the AI ​​nonsense, I immediately deleted all the items because I suddenly didn't feel like it anymore.
DAZ3d helps destroy creativity and the fun of rendering as a hobby!
I really can't understand why they support this!

creed2003

I'm not going to stop buying from the Daz store just due to AI. AI to me is a tool, an enhancement to renders created from store content. It's great for making backgrounds and environments but I've found it lacking in character creation. Too many mangled hands and other issues. To be fair I've only used DAZ AI Studio so I'm not sure if others are the same.