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
Okay so locking Michael 9 definitely generates only M9 stuff and it looks (to me) that unlocking uses it also for some prompts but maybe mixes it with other stuff. At least the body shapes are very similar when you use keywords like skinny, swole, or fat. The face selection that also pop ups also seems based on just a few models like if I didn't specify anything I got this "east asian" face on all variations but if I added keywords like scandinavian or character names like nicholas there was a bit more variation.
It seems to have the pretty normal limitation of being only good in pinup images. Try to add complex actions into scene (i.e. adventurers fighting a dragon) or something very specific (elf archer, dwarf wizard and human paladin fighting a dragon) and you get very mixed results. It does pretty good results when you use keywords like "dwarf wizard casting a fireball" though it's arguable if I ended up with a dwarf wizard and in one instance it looks like the wizard miscast his fireball...(self immolation is bad, mkay).
This for the free version anyway.
That would be me, so thanks for posting this announcement in Commons, otherwise I would have missed it.
Have bookmarked the AI forum for future reference.
There aren't as many of us as there used to be or as would be expected, and if I could have merely told my Arp or Moog: "Play Art Tatum doing variations on Chopin," and it did, then I'm sure there would be even less.
Real people and an artist's rendition of people aren't the same thing. I also never said someone wasn't an artist for using Daz. Nor AI.
The text to 3D tool is still being worked on, I am not sure if this is a first step on the way or unrelated.
Hypertaf’s Winter Trendy outfit is indeed on 3DShards now.
Please, do not blame me here. I just responded and explained how already created DAZ content can still be a resource for great art.
This is the official press release, and it makes it clear it is indeed using Stable Diffusion XL as a base. Then they use Loras built by Daz on top to give us this Daz AI.
https://www.prnewswire.com/news-releases/daz-3d-unveils-daz-ai-studio-with-partner-stability-ai-302106757.html
The biggest issue I see is how poorly it handles things not close to the camera, a problem with all AI generators. Just looks at all the pics shown so far. The best looking ones are all portraits of the face, where the face is closer to the camera. That is all well and good, but the moment you try to make a full body shot it all goes downhill. The person no longer looks like Victoria or even an actual human for that matter.
It comes down to pixel counts. The more pixels dedicated to an element, the better the results gets. Just like the denoiser in Daz. The more pixels your image, the better the denoiser performs. When you have a person farther from the camera, there are fewer pixels dedicated to their features, thus the AI has less to work with.
One way to work around it is to create your intended image, then crop the image you want to fix, and use 'image to image' on the cropped shot. Render the cropped shot at full resolution, and that should create a better pic. Take the output and resize it back into the original image. That takes some work, but that is one way to get it done. You can do this with the hands and face, and whatever needs fixing.
I'm only leaving this reply short and sweet to stop a rant.
I wish Tafi hadn't done this and it isn't something I will use.
I have no idea why people are so hung up on AI, the thing can't seem to learn properly, I mean seriously, after more then 2 years and all the learning from all the images around the internet and input from people, it still doesn't know how many fingers are on the human hand, or the difference between a left foot and right foot. I question it's intelligence and ability to learn, they are lacking.
I like it for the one image I generated. I can't afford to yet but I would pay the current sub fee.
Questions:
a) I generated an image with a prompt. I liked it (the hand was messed up not surprise) but when I went to look at it again on the DAZ AI webpage it is gone. Where did it go? I have the 'free' sub for now.
b) Is there going to be a mode where I tell it to use my DAZ 3D product library as shown in my DAZ 3D account?
If you refresh the page, you lose everything on the page. It does not save anything. So if you want to save your pics, you need to save them from the browser before you close the tab or do any refresh. If you want to save a prompt, maybe whip out the old Note Pad.
Daz would need to add training data for a crazy number of products. With well over 20,000 products, it just isn't possible. Even limiting it to G9 items and more recent environments/props/vehicles ect would be a challenge. It takes time to train a Lora. And in this situation, they need to also render all the images first before doing any training. That's a pretty serious project.
There are people on CivitAI who somehow upload multiple Loras every week, but they are dedicating all their time to do that...making Loras has become their full time job (they typically have something like Patreon to support them).
But that raises some questions. While Daz stated this is being developed by a different group of people outside of Daz, I would assume these people get paid to do this work. Thus one would conclude these are resources that could perhaps be used to help improve Daz Studio by adding more developers to the team. I would think getting Daz Studio 5 across the finish line would be a bigger priority than chasing after AI generation. Daz AI is extremely limited, and even if they add a bunch of stuff it is always going to have some guardrails attached. Meanwhile people can go to CivitAI and use any number of custom checkpoints and loras, not only can they can download locally, they can also generate images online through CivitAI itself. That gives people far more options and freedom. Maybe they do not have Victoria 9, but they certainly have loras that specialize in redheaded women with freckles. Yes, you can get consistent characters using already available loras, plus many more. So Daz is not offering anything that is truly new here.
I will add that CivitAI is priced higher. Like DazAI it has a free plan, but you are limited by a currency called "buzz". You use buzz points to generate an image. There are of course paid plans, but they start at $10 and go all the way up to $50 PER MONTH, while DazAI is $4 (for now). So there is that, but as I said, you can download these to use locally without any restrictions. If you have a PC that is decent for Daz Iray, chances are it will also be decent for Stable Diffusion. You can even use GPUs with just 4GB of VRAM. To use SDXL, which is what DAZAI is based on, it is better to have 8GB or more. But it is possible to use small capacity cards and you even use AMD GPUs. Though I must warn you, AMD is not great for rendering AI, but the option is there, they will function, just slowly. Even the lowly 4060ti can run AI very close to the fastest AMD GPU available (7900XTX,) which is kind of sad. But at least it is possible to use AMD.
All that said, AI still has a lot of problems. It is pretty much only good at basic poses. Anything besides "standing straight and still for a portrait" starts to get funky. You want an action shot, good luck with that. You can use image to image as a pose template, which can be very cool when it works. In fact, in theory you can even draw a stick figure and use that as a guide pose. But it doesn't always work (rather it kind of rarely seems to work as well as you hope). If you really want consistency, you want true 3D models. But AI can be fun to play with. It can scratch that loot box/gambling itch, because sometimes you never know what you are going to get next. You might get something really cool and unexpected, or you might get another mutant freak to haunt your nightmares. That element of not knowing can be part of the fun for some people.
I don't want to be misunderstood. I am not entirely against AI, I am just against generative AI.
some things in DAZ studio renders are hard to achieve, like realistic hair.
It would have been great to have an AI filter to give the hair the last mile of realism, the same with skin. what is neede is the last touch up of realism. That's where AI is useful in DS.
The king discipline would be joint bending. As far as I know, it exists already in the UnrealEngine and MetaHumans.
THIS would be the path to go.
maybe they can let people contribute captioned renders for training
some of us would be willing to do that in turn for free goes at the advanced tools ot extra generators over the 25 per month
even if just on a month to month basis
obviously we couldn't include content by PAs or pther sites in those renders who did not wish to participate but if it's all DAZ originals only that's fairly easy to do
or alterniatively in return for tokens, coupons, freebies etc if cloud GPU usage cost an issue
some could even submit their own captioned photography, plenty of us take photos of local scenery, flowers, wildlife etc
Right now the DAZ gallery doesn't seem to remember by settings, for I have set it to traditional.
I wondered where Blackwood Manor went
Progress bar goes til the end but then just get a loading wheel spinning forever. Am I not doing something right?
Start with really simple prompts first, just a word or two.
This is interesting... mind you I say that regardless of whether I find an interesting fact or see something cool or cut myself so badly I need stitches.
But yeah...
First NFTs, then ads, now AI generated images from a place that sells 3D content...
Cool.
Has anyone here considered Tulips for the next big idea?
I know, I know... you're all thinking: "but... Holland... 1630s... why?"...
And I'm saying "Tulips, but virtual"... eh?... see?... Virtual Tulips can't fail because instead of bulbs to trade, you get holograms of virtual tulips generated by which you can buy with NFTs... the beauty is... actually, I'm still working on that, but it's got all the stuff everyone is doing now... AI, NFTs and Augmented Reality... it's a trifecta of bandwagons conveniently merged into a single bandwagon that's sure to be... something.
My point is... I think we just had an earthquake so I'm going*...
*Not a joke
Virtual Tulips a la Daz AI studio.... stay safe re the earhtquake
The generation history seems to be a per-session thing, so our images aren't automatically saved (unlike with Midjourney).
So if you guys want to keep anything, even to compare with future versions or prompts, make sure to download your images manually.
I'll ignore this AI thing the same way I ignored NFT and "reality shows."
Short version: how do I get to the advanced features like control of image size and ratio?
Long version: I discovered this last night ... have played around with a bunch of prompts. It seems hit or miss - the hits are great and the misses are hilarious. I decided to take a bit of my Daz budget and play with this for a bit even just for amusement ... but I still have the same limited controls that the free version had.
Oh ... and here was an attempt at a more cartoony look:
As an artist but also a content creator, I feel so conflicted about this.
Will this stay a standalone product, or are there plans to try to integrate this with Daz 5?
That is basically being an Art Director which is also a creative job.
OK, thanks for the answer.
I will save future DAZ AI prompts & images locally then.
I don't see what's preventing DAZ 3D/their hired developer vendor from allowing the DAZ AI users from uploading their own DAZ Studio renders made with the products the users own as a "Learning Appendix" to the DAZ AI superset that is currently available to us all. That is how I made, or rather, the AI at Creator.NightCafe.Studio made the current avatar I am using on the DAZ forums. I suppose I could upload my DAZ renders at NightCafe.Studio as a new training subset and in that way I can create subject matter trained subsets.
I have installed a version of Stable Diffision and used it a few times on my local PC, LOL, the hardest part was finding the images' and videos' location after I generated them. Even though I have only an RTX 4070 the generation was impressively fast. For now my interest in having it local is eventually to get around to building and linking it locally. I prefer the ease & speed of NightCafe.Studio though as I get 10 easy credits a day (more is possible but I'm too lazy to do the extra work needed) and add 100 more credits via a $15 a quarter Pro subscription. I'm mostly just hoarding credits now as I'm sure the AI will get better, faster, cheaper, and give the 'prompter' more finely honed control over the image, eventually. Currently the AI produces mostly somewhat random pathways to combine it's stored data when you get beyond a certain set of basic requirements, maybe it needs to be trained in motion and a lot of other things, which can be entertaining in itself to see what it comes up with.
Now I've lost my respect for art directors, thank you.
My workflow has been use Gen 8 or 9, do a pose and render in Iray. I then bring my render to Leonardo Ai using the img to img (set to 90% image guidance) and a simple prompt and my original 3d render looks photoreal. It really does keep a very similar look to my original character but the eyes, hair, skin etc look like a photo. This is the direction I wish DAZ would follow where you can use photoreal or other render styles. I actually like this better than outright prompt to image. Only problem is if you want to get a bit of NSFW than you are out of luck on most web based AI apps.