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
© 2026 Daz Productions Inc. All Rights Reserved.
Comments
Always interesting to see how people are using img2img AI with their DS renders as input!
Personally, I view post processing as a second stage of rendering, with DS/Iray being the first. This has always included (and still does) traditional 2D image editing like more dynamic, perceptual tone-mapping of 32-bit HDR renders, scripted actions in pShop, etc. And now, for me, img2img AI processing has been integrated as another tool in the process.
The first example shows how an Iray render can be enhanced. In this case, the scene was lit with just sun/sky and a couple of light blocking planes. Also, a backdrop was used as opposed to a full environment.
In this case, the original render was 800x1040 and only cooked for 400 iterations. This took 79 seconds to render on a 5060ti. The Post Denoiser was used as it would have been very noisy after only 400 iterations without it. The output from the AI was 1264x1648 (resized to 800x1040 for the comparison).
As somebody pointed out, AI can be used to improve lighting, but it can be used to do so much more than that. In addition to enhancing the lighting, the resolution was also increased. Perhaps most importantly, the softness introduced by the post denoiser in DS was counteracted. Check out how much more detail the skin has, and how much better the hair looks.
So more pixels, more detail and better lighting. All this happened locally on the same machine running a 5060ti in 25 seconds. Original DS render on top:
I've spent literally years messing with this stuff designing workflows, running batches upon batches of images every night only to rinse & repeat again. It's always a competition between enhancing the image vs. staying true to the original source. I think in this case the resulting image has better lighting and more detail - all while remaining true to the original composition. It was much faster than cooking a higher res Iray render for thousands of iterations, too.
Img2img AI can also be used to turn renders into digital paintings, ink drawing, watercolors, etc. The same trade-offs and considerations exist. I fear I've written way too much already, so I won't bore you all with the details this time. Original render on top, then a digital painting style, followed by some black & white linework, and finally a watercolor-ish storybook style:
Sorry for the book! lol
Please keep all the img2img examples coming!
- Greg
I like both the originals and AI results
Very nice
Great looking images, @algovincian
Thanks. I feel that DS and img2img AI are a useful combo. IMHO, the things that make DS fun for users (composition, framing, character/clothing/hair design choices, posing, etc.) are still reflected in the end result and require the same skill to create.
For existing DS users, img2img AI processing speeds up the render generation process, enhances the output, and increases resolution.
For those coming from an AI image generation perspective, DS provides a way to exercise much greater control.
- Greg
I animated the cover of my Daz-powered take on the Marvel Spider-Man character and story.
It's the Dynamic Spider-Man Plus
Cool animation! The flares add a lot.
- Greg
3DIO said:
Just as an amusing point of reference, for over a decade that's exactly what "real 3D artists and modelers" at other sites said about people using DAZ 3D and DAZ assets.
Yep I remember people saying that it wasn't art if you used Poser or Daz to create it
the irony
I laugh at people who are openly copying other peoples intellectual property who say AI is stealing
not denying the latter but the self unawareness is staggering
just because you skillfully steal with your artistic abilities and hard work, doesn't make it not stealing
I understand Fan Art but honestly if you are selling it, it's no different to using AI legally
That, i can cope with, i think. But when someone comes into my home and takes something i can not replace, that had sentimental value, no amount of insurance could replace it, as though an insurance adjuster would pay me for the lock of my mother's hair when she eight years old. Damn movers said they would walk off if i tried to retrieve it from the garbage. But the books i didn't want and was not interested in reading, that they would move. What a nightmare!
i don't mind sharing, but to deprive me ....
Anyway, i let ai learn from my photos on DeviantArt in hopes it will be a better engine.
I take photos of my plates and dishes, my meals i cook or bake. It animates those into the movie clips. i take photos of me and it does a better job than the Daz texture map of me on a G2 let's say. See attached cartoon of me.
What this means is, i can take a photo of my dead brother and put him back to life, narrating and acting in a book about himself, about his life and likes, about his illness for five years with cancer, and his death in 1973 at the age of 18. i can clone his voice from audio cassette recording he made.
That came out super fresh. There's subtle reflections on the carbody and windshield.
I had a similar effect in one of the music videos- where I went to the extreme and spent some time working that idea of the windshield as a canvas.
----------------
I also think some of the issue is spending too much time around "people that make things". and then making things becomes the point and there tends to be a huge focus on the "rules of making things".
You know, what you can use to make stuff...
How you can make stuff...
Even what you're allowed to make...
as opposed to making stuff for the people that are meant to enjoy it.
Your peers versus the people.
I've succumbed. Ended up using AI to generate a 5s clip using reference images to show an electronic gate sliding open. No matter what I did manually, it always looked jerky.
You can see how it would be hypocritical, coming from anyone...
Long as you accept this as your own personal judgment, based on your attitude and feelings.
Every artist sets their own bar/criteria/rules of art.
In a pool of peers, it's about respect (for your work/method)
Daz Studio (and others that use similar software) tend to consider every other type of artist some kind of peer and value their opinion as if they were actual peers.
--------------------
You're hinting at an Artistic Rabbit Hole (of purity) that one can never get to the bottom of- as soon as you name you approved method, someone else can come along and explain why your system is indeed 'Cheating'.
You can see how it would be hypocritical, coming from anyone dependent upon raytracing software to produce every image they ever rendered, to then complain about an AI tool doing the rendering instead
Long as you accept this as your own personal judgment, based on your attitude and feelings.
The full quote changes nothing. You are making a statement based on your own personal judgement. Which is fine.
And to be blunt- the two are not equivalent. Using Daz is nothing like using AI, unless you mean Daz as a tool and AI as a tool.
We all know the most denigrated aspect of AI use is prompts leading to generated material. That's what the fuss is about.
------------------------------
But with AI, there's no end to what a dishonest person could post and get away with as their own work.
Did you know that there were people posting Daz Promo art and claiming they made it?
Did you know that there are numerous people posting art in art groups that they did not make - and many are using it to get commissions?
I just think it'd be nice if we stopped saying AI this and AI that and start talking about the people using it.
People are doing stuff, not AI.
If AI isn't a tool, then what on earth are you using to create your videos?
Well, accoring to those that do not use AI, I simply typed a sentence in ChatGPT and viola, it made the entire thing.
Same issue with Daz. People that have no idea what it does or how it works- have attitudes about its use.
For every tool, there's low level use and high level use.
When people goof on Daz Studio, they usually make up an example of someone just randomly throwing things together and hitting the render button.
They do that with AI too. Just type in a random sentence and see what happens. lol.
You hit a good point about transparency - most artists hide the fact of AI use to avoid being pre-judged.
Once you involve AI, it becomes unclear what you did and what the AI did.
It's hard to know what kind of credit to give unless you are experienced in the tools and techniques.
Most of us can evaluate a Daz Render and see the work that was done in posing, lighting, set design, camera settings, morphs, textures, post-work etc..
You know where the user kicks in after Daz Studio reaches its limits.
To a non-Daz user, you don't know the struggles. lol
This topic on this page (well i haven't gone back far enough as i just recently joined and still viewing) i am surprised if it wasn't brought up earlier.
It reminds me of an episode of Mad Men where a kid is accused of cheating because he used tracing paper in the 1950s or 60s. And Don Draper the creative director at thecad agency says so what, we regularly use tracing paper.
In my universities where i modeled in the States and Canada, the professors said to the students to use tracing paper and cut out photos, to use them as reference. Even going so far as to use a projector on canvas of an art work, as seen in The Moderns for example.
What can i add?
Human = Director (Me)
Ai = Director of Photography (Gemini, ChatGPT, Grok, etc.)
Engine = Camera Operator (Nano Banana or other)
And this is the point that I think a lot of people miss while being angry or dismissive.
Over the last 35+ years dealing with the evolving tech with regards to the arts I've seen the same pattern repeat itself over and over.
1. A new technological tool is introduced that allows new (and unskilled or untalented) users to create something they couldn't before the tool was introduced.
2. The world is flooded with creations by both the skilled and unskilled, talented and untalented users and because the skilled subset is smaller, the world sees a lot of horrible "art".
3. Eventually the skilled and talented users show how the tool can be used to improve their already excellent output, while the untalened or unskilled users either learn to use the tech and actually become better artists and the others? They either lose interest or become lower and lower in number that their influence is no longer all that noticable.
In the 80s, when home computers became able to produce professional-level typesetting and page layouts the world was flooded with people putting a dozen different typefaces on the same page and producing layouts that were fast and cheap, but ugly. Now? You rarely see that work anywhere.
In the late 90s, early 2000s you saw a mass of pretty low-quality 3D characters and scenes being rushed out; looking back, for every one or two really impressive renders (often due to postwork) that came out, there were a dozen or more crudely-placed-and-posed Poser/DAZ people in a scene that are still a little cringeworthy when viewed today, but again, you rarely see that work anywhere anymore.
In the 2020s we have a whole generation who could type a quick prompt and generate AI art (complete with three arms, seven fingers, and so on) and who shared their creations with the world. As the tools get better and the people learn, we're seeing that less and less.
In every case, we see New Tech, Floods of Slop, Refinement and Acceptance.
In every case, you saw people in the industries rant about how the New Tech is destroying their industry and who point at the swarm of low-quality being produced by the unskilled and untalented.
In every case, the talented artists who take up the new tools become the leaders in the industry, and the masses who jumped on the bandwagon either learn and get better or slowly die out until there's another new tool for them to play with.
I agree. My hope is that before & after comparisons, along with an explanation about what was done and why, may be useful.
In the following example, a sun/sky based lighting setup was used. It includes the ability to better control shadows, rim lighting, etc. It also adds the ability to render figures with dappled lighting effects while using sun/sky.
From top to bottom:
1. base sun/sky (DS render)
2. dappled lighting added (DS render)
3. better shadows & rim lighting (DS render)
4. img2img AI processing of #3
The DS renders were 800x1040, and each took < 1 min on my system (5060ti). The AI output was 1264x1648 (resized to 800x1040 for comparison) and took 25 seconds to process locally:
IMHO, in addition to the increased resolution, decreasing render time and better detail in general, the img2img AI processing of the DS render specifically helped by:
1. adding some atmosphere without requiring the rendering of volumetrics in DS
2. improving the hair
3. improving the eyes
4. improving the default tonemapping for Iray in DS
5. removing the artifacts on the right edge caused by the way DS wraps the backdrop
BTW, does anybody know how to stop the wrapping from happening when rendering backdrops in DS? It's most prominent in the lower part of the right edge where it's dark. The bright sun on the wood from the left edge is obviously wrapped to the right edge.
- Greg
Also as soon as i heard synthesizers i heard in my mind Tears For Fears - Head Over Heels (1985)
I was listening to synthesizer music in the 70's as a teen and loving it
Switched on Bach was a much played LP on my primitive little turntable that folded into a case
surprised anyone could think it wasn't real music, but then again, I did have a music teacher in school really into experimental stuff and we did learn about things like the Fairlight computer watching documentaries
and I don't question for a second if any AI is art or not, only if it's the prompters art
like the chimpanzee who took an award winning photograph
Synth always makes me think of Jean Michel Jarre and Vangelis.
I still have my cassette tapes of both
Just watching TaurusTarkus play on YouTube is so beautiful.
I did have a shed load of cassette tapes, until a replacement player was going to cost more than a deck to plug into the PC and convert them all to mp3.
Ha! You're right! Should have paid more attention to the pose and camera angle as the way the rear paws line up with the front does make it look that way. It wouldn't be the first time AI messed up fingers/toes, but the blame goes on me this time!
I was too focused on testing the light setup/fur. The scene was originally setup with NPR in mind (as opposed to photo-realism). Now that I've actually processed it with the storybook style, the poor little guy looks even more like he has way too many digits:
lol . . . fail!
- Greg
who's Florian?
and no, was nothing that fancy
was sort of a white with black houndstooth pattern vinyl covered case, very 60's
I do watch a lot of videos about musical instruments, particulary keyboards as played myself
not so much now
this looks like it