What do you think about the future of 3D MODELING?

#1

The Metahuman shared: https://www.youtube.com/watch?v=S3F1vZYpH8c

I think the modeling (especially 3d character design) industry has been taken over by software. Ten years from now, it seems like there will be no need for character designers anymore. Maybe “art” may not die, but all sectors of 3D design creation like these will die.
There are serious problems already, let alone 10 years later. If you were a boss and wanted to create a character, would you use such software cheaply and produce quality characters in a serial manner, or did you hire expensive workers? What is your opinion?

#2

The roles will simply evolve.
Look what happened when 3D animation overtook 2D (hand drawn) Movies in popularity. It may have put a lot of line tracers out of work, but the 3D industry actually helped create more jobs because the movies themselves were more complex.

If it’s no longer financially reasonable to pay for 3D modeling, then modelers will just have to transition into something else. Or maybe it could give people the opportunity to make their own movies/games instead?

If I only have to click a few buttons to generate a Grand Theft Auto-esque world, then why not sell this content on the market instead of just looking at it? And honestly, I feel like it would be for the best.

Hollywood or AAA publishers will no longer have a strangle hold over the industry. Everyone can start producing their own original ideas from the comfort of their basement while retaining the production values of multi-million dollar budgets.

#3

The transition from 2D to 3D is not the same as taking the industry completely under AI.

#4

Can you explain?
3D technology in general is just one big automation. All the old 2D techniques of having to draw everything frame per frame was completely eliminated by computers that could render photoreal scenes in less than a second.

Unless we get to a point where AI literally just spits out movies or video games at a press of a button, human involvement doesn’t disappear. For example, look at gameplay. How do you tell a computer to make a video game that is fun? Sure, it could probably simulate something very simple like Pong or Tennis, but then what if you want gameplay that is much more complex and rewarding? A computer doesn’t have human emotions to compare itself with.

Or in the field of movies, how do you tell a computer to write a story meant to make people cry? Without real human input, a computer generated storyline will come across as generic.

Computers have gotten more powerful every decade but at the end of the day, they are still just tools. We even have technology that could theoretically replace Actors, yet people are still lining up to watch Tom Cruise or Will Smith act out on screen.

#5

I agree with what you said, especially in the art part. Artificial intelligence cannot come to a position to produce art. Maybe years later, but that will take many years. If one day artificial intelligence reaches the level that it can produce artistic works, then spirituality comes into play. This is why handmade workmanship is still expensive. This is the reason why the king of Saudi Arabia bought handmade carpets at an expensive price instead of machine-made carpets. Years from now, people will split the games into two, robot made - human made. Human-made will be more expensive, and robotic construction will remain an option mostly preferred by the poor, especially since it is introduced to the market cheaply due to factors such as cost, spirituality and product abundance (value). Maybe we will not be able to see those days, but if this site remains for many years, they will read our comments in the future and understand how correct our opinion is.

#6

Let’s put this into perspective for you. I remember when Fractal Design first came out with Poser in 1995. Similarly, the various trade magazines of the day heralded it as the beginning end for character artists. The app was deeply flawed, but showed us the promise of a future where we could just point-n-click our way to greatness. Version after version, 26 years later, and the human character artist is still here. Apps like Poser and Daz have certainly made certain artists’ lives easier, but the promise of a point-n-click future never really manifested.

The art community is typically presented with two options. OPTION #1: Fewer clicks, but templated. OPTION #2: Many more clicks, but unique and individualized.

Apps like Poser and Daz fall into the first category. As the apps shield the artists from the more arduous and mundane tasks in the early stages of character creation, they can instead use them as the starting point for larger projects. This is, in many ways, similar to how custom action figure creators might use a stock figure as the basis for something else. The downside to using such apps, however, is that you can almost always spot that “made with” signature. It doesn’t matter how much you customize them, spotting a project where Daz or Poser is involved is pretty easy. That makes sense since the foundation is the same.

With the second option, the artist has to get their hands dirty more and the entire process takes longer. Block. Sculpt. Detail. Retopo. UV. Texture. Rig. ETC and so on. It’s by no means a one-click solution as it demands input from somebody with more traditional art skills. That’s not an insult to artists using Daz or Poser. It’s just an acknowledgement of the fact that modern CG character artists using something like Maya need to know more about the underpinnings of art. The learning curve is higher. On the plus side, no two character pieces look exactly alike. If there’s a “made with” signature then it usually points to the artist and their style instead of the app of choice.

As we move toward a future where AI has its say, don’t expect this two camp approach to change all that much. AI, like the introduction of Poser, will make certain people’s jobs easier and quicker, but many will look at it as a starting point and nothing more. Artists looking to create works that say more about them than their app choice will almost always want more control. You can put a paintbrush into the metal hand of a super smart machine, but it’ll never replace the subtlety or experienced touch of a soft, squishy human.

Ultimately, you’re just going to see app developers provide us with more tool options. They’re going to create tools that are more intuitive, but also ones that allow us to express ourselves as individuals.

Look at ZBrush. It was a true revolution for character artists. Previously, character artists were limited to poly-by-poly or box modeling techniques, both of which can take a very long time. Digital sculpting brings us closer to the traditional world. ZBrush bridges the gap between chisel and stylus. Despite being that much more intuitive, ZBrush really didn’t make our lives easier. If anything, it raises the bar for what was possible and expected.

In truth, app developers like Pixologic aren’t necessarily supposed to make our jobs easier. They’re not supposed to automate the process either. Their one role is to make digital art more accessible and to allow us to apply more traditional, time honored, concepts. Their job isn’t to eliminate the toolbox, but to create better, more intuitive tools. It’s not about providing artists with an assembly line. It’s about providing them a larger, more option filled sandbox to play in.

We may see AI play a greater role in how we make art, but it won’t ever be to the end of replacing artists. The goal of AI, much like that Adobe’s Content Aware features, is just to make the more mundane tasks simpler and get us back to the task of making art. Even then, there will always be that one instance where the artist will say, “Nah. I want this one specific result and I know better.”

These helper apps are still made by humans. Let’s never forget that. Even the best, smartest self-teaching AI will be constrained by the foundations of its own design. Humans are no different. You can be the smartest human being ever to exist, but you’re still going to be subject to human failings. The moment that AI created art transcends human created art, is it really creating art for humans? Hmmmm… :slight_smile: If you want art for humans, humans can’t ever fully exit the process.

In the end, these are all just tools. They’re not here to replace us.

#7

The article you wrote was really nice. I like it.
I want to ask you something. Well, with the increase of these tools (for example, with the widespread use of meta-human), when it comes to the stage of producing quality people, the situation that will occur after the production of side characters such as NPCs by such programs does not mean that artificial intelligence will take over the human profession but it is clear that the boss’ demand for 3D artists will decrease. If you are a boss and want to produce NPC, would you use such programs or would you use people by paying a large amount of money? Look, an X quality product is worth Y dollars. Again, a product of X quality is Y + $ 100. If you choose the same quality Y dollar product, people will be eliminated. In the short term, I expect a decrease in the number of people in the 3D character design industry as the quality of such programs increases. This decrease will drive people to produce niche characters, leaving only a handful of 3D designers who have to do their job perfectly well. Natural selection works not only in nature, but also in every part of the world, in every system.

#8

I don’t disagree with what you’re saying. Companies/Bosses will do anything that could cut costs but reap greater profits. It’s why we see more 3D art assets being outsourced to third world countries.
So if automation comes along and it drives down cost of 3D art to mere pennies, then there’s going to be widespread adoption.

I also agree that it will drive the more professional artists into serving another niche. But ironically, this is also another opportunity for them to become popular again.
For example, I brought up in my first post that 2D Animation has declined but in truth, it’s making a resurgence again. That’s because all those Veteran Disney Artists came together and developed a new technique to make 2D Films popular with the world again. One example was the Klaus movie that came out in 2019.

So I could see something similar happening again with 3D Artists. Even if robots take their jobs, there’s nothing stopping them from innovating in ways computers haven’t caught up to yet.

#9

As a long time Poser user, I want to toss in my two cents here…
I’m one of those people who mostly wanted people in my scenes for scale. I didn’t get into Poser looking to make specific characters, and it was a long time before I seriously tried making a character within Poser (using the included) assets.
There are a number of things that made Poser use so recognizable for a long time:
It had a very simple renderer with very specific known limitations, so anything rendered in Poser looked rendered in Poser, even things brought in from elsewhere.
The included figures showed up everywhere, on the internet, in games, and IRL. People are good at spotting recognizable faces, and those early characters became so ubiquitous people knew they had seen them before even if they did not know they were from Poser.
The figures themselves were (and still generally are) idealized in a specific manner. They are neither super heroes nor fashion models, they are not “average” in the way an average person is average but yet still tend to be plain and generic. They tend to give themselves away because that blandness often carries over into the final created character. A model made to be cute or menacing is simply more likely to be more cute or menacing than a Poser figure “adjusted” to be more cute or more menacing.
And, for the longest time, Poser had the most simplistic rigging that could have been made to work, and lacked soft-body physics or even an easy way of faking the effects of soft-body physics. Weird collar bones and shoulder deformations, gravity defying breasts, calves intersecting thighs/biceps intersecting ribs were all problems a Poser user could expect to see. Those problems were made even worse when the breasts were made stupidly large, or muscles were inflated to fantasy proportions.
And perhaps the biggest tell of Poser was any scene with a well modeled figure mesh(es) but otherwise using mostly simple props and primitives. Such figures were often nude or semi nude, as the “artist” neither had the skill to model their clothing/accessories nor took the time/spent the money to find them elsewhere.
Now, 26 years later, and all of these things are indeed still common with beginner/lazy Poser art. What I suspect has changed, and this what prompted my reply, is that there is likely much more art out there that was created using Poser that you do not immediately recognize as having been made in or with the help of Poser.
The archaic P4 renderer was supplemented by the Firefly REYES renderer in the mid 2Ks. It had some artifacts, or “tells”, of its own but was sufficiently flexible that an artists’ work need not look like every other artists’ work. Poser also incorporated blender’s Superfly renderer a few years ago, so it is now possible to get images out of Poser that rival those now being made in blender.
Every version of Poser has released alongside a set of figures, and while most are still very plain they are nowhere near as recognizable as those included with Poser4. Poser now has a figure set-up room, unimesh skinning and weight mapping/painting. It is much easier to rig a figure than it was 20 years ago, and it is much easier to import a figure from elsewhere than it was 20 years ago. Also, the morph brush makes it easier to sculpt unique characters even from the ancient P4 figures.
In addition to the unimesh skinning and weight mapping, Poser now incorporates bullet physics for soft-body physics.
Unfortunately, no matter what is added to the program, there is nothing stopping someone from just dropping in a figure, adding a few props, and clicking render. And those images will always look like they are from Poser (even if they aren’t). But unlike 20 years ago there is no “wall” a Poser user hits in which their art cannot improve unless they leave Poser for a higher-end program. Poser’s animation tools are woefully out of date, some say they don’t even work as well as they once did. But for illustrators and those making digital comics it is possible to make images in Poser that look as good as those coming from anywhere else (and which can’t be immediately recognized as made in Poser).
Although, ironically, it might not actually be any easier.

#10

Don’t get me wrong. I don’t mean to deride apps like Poser or Daz. In fact, for the first 3 or 4 versions, I was a Poser user. I mainly used it for pose references and previz, opting to hand model my characters instead. I have nothing against such apps. I agree. In capable hands, they can be powerful tools for artists looking to use them as a starting point.

That said, I often find myself browsing ArtStation and noticing some gorgeous work with a set of signature “tells”. Almost invariably, looking at the apps used, I see the Daz logo. Nothing wrong with that. The artists in question used these apps to excellent effect. I just think that artists have a signature style. Even when a supremely talented Poser/Daz artist heavily works over somebody else’s base, you can usually find hint and traces of what came before.

I don’t see these apps as cheats or lazy shortcuts. I’m pretty sure that there are plenty of “made with” pieces that I haven’t seen or noticed. You’re 100% right.

As a side musing, in trying so hard to delete that signature look, I wonder how much time is really saved when compared to just creating your own character; Modern block out techniques for poly modeling & sculpting have really streamlined the process.

I’m not sure if that’s a compliment or a dig. :stuck_out_tongue: Uhmmmm… okay. lol

In what way is it clear? Consumer expectations and developer over-promising are only driving up the demand for more content. Ideally, that would actually INCREASE the demand for more 3D artists. In truth, however, it probably just means that more bosses will work their employees into early graves. More hours. Same pay. One artist sometimes doing the job of three. ETC and so on. History kind of bears this out.

The NPC situation is actually not nearly as complicated as you’re making it out to be. Game artists and film studios like Pixar solved this problem decades ago.

Let’s take the case of Pixar. In 2005’s “The Incredibles,” they realized that there was going to be an inordinate number of non-speaking background characters and bit players. Crafting each one from scratch was surely possible, but not financially feasible or practical given time constraints. Their solution was simple: Universal Man/Woman.

They created these two simple, yet flexible base models in house. These were custom built models that fit the style of that world, but were topologically optimized to such a degree that they could repurpose it to create any one of a hundred new characters. Topped off with custom props, costumes, and wigs, they had a complete chorus of players that required the least possible effort to create. It was, effectively, kit-bashing on a character level. Honestly? How many of us repurpose old props and base models. That’s all this is. Model once. Repurpose forever-ish.

The BG/NPC problem was tackled even earlier than that. Go back to 1999 and “The Phantom Menace” from Lucasfilm. Take a look at that pod racing scene. Love that grand crowd in the stands? Bet they must’ve had a ton of extras. Wait. This is Lucasfilm. The MUST have created that crowd with CG and particle instancing or something, right? Nope on both counts. Their solution is so insanely simple and effective that it’s stupid for me to say. Lucasfilm’s art department actually used colored Q-tips. To make them look like a waving, cheering crowd, all they did was move them around with a hair dryer. Yep. It was that easy. Yet, from afar, you don’t notice.

Right here, in these two real world instances, there is proof that you don’t have to remove the humans from the equation. Complex problems often result in creative solutions.

Your claim that “natural selection” will drive down the number of 3D artists is just crazy and has no historical evidence to back it up. Like a few other people here, I’ve been doing CG for MANY years - some 32 in general and 28 in a professional capacity. I’ve seen the tools get more intuitive and artist friendly. I’ve seen tasks that would once take a week now take only hours. More than that, I’ve seen the desire for content increase and the number of artists employed increase accordingly.

You can have the smartest AI on the planet and you won’t eliminate people from the equation. That’s just silly. It’s like saying that, because we now have robots on assembly lines that human workers will no longer be needed. That hasn’t happened either. In some cases, humans are recast in other job roles. In other cases, we still require people to oversee, repair, and upgrade the machines.

These new tools and AI driven techniques ARE exciting. Let’s just not oversell or over-promise. They may well make our lives easier, but they won’t replace us. They can’t. I think that it’s also reasonable to at least suggest that, as with other game changers from the past, that these new tools will drive up the demand for content with greater scope. There are still going to be a lot of 3D (character) artists in the mix. Our roles will simply change or adapt to the moment. We’ll do less of “x” sort of task, but more of “y” instead.

I would also like to point out another real world example. When ZRemesher for ZBrush was introduced a lot of very enthusiastic (naive) artists proclaimed it the end of manual retopology. After all ZRmesher (and its cousin, QuadRemesher) produce excellent automatic topology. In many cases, it DOES replace the need for manual retopo. However, in as many cases, we still see artists resorting to hand retopo since they can get the flow that they want, at the poly counts they want, and with whatever side optimizations they can dream of. Something (semi-)automatic like ZRemesher was indeed a game changer when trying to make sense of these dense, poorly organized sculpts, but it didn’t end manual retopo. There will always be that one artist who wants it done a very specific manner - one which the programmers of the automated solutions had not intended.

Save this post. Seriously. Come back to it in 5 years and tell me just how accurate your crystal ball is. I suspect that you, and everybody else heralding our new machine overlords, will eat crow by the bucket load.

#11

Lol, that was really a compliment.

We actually say the same things. We have no differences of opinion here. But except for one chapter. In the last sentence of the paragraph, you said “5 years”. I did not say these will happen in 5 years. If, after 10 years, tools such as these come to a position to produce the same product that an employee would produce for the same price, the employee would not be needed. You say these tools will create new job opportunities. If such programs will facilitate the process of making games, the number of people who make games increases and new jobs arise. It sounds good. But this is not the case in the market. If the number of game makers increases, gamer makers will not need employees because they use such programs. It is necessary to understand the vicious circle here. In addition, these are “tools”. We speak for 10 - 20 maybe 30 years from now. If one day these tools cease to be a “tool”, if you become able to produce anything you want (except art), even someone without a talent will be able to produce any character. Poser, Daz, Meta-human. These are tools, I’m not talking about them. Imagine a much more systematic version of these. What will the role of the employee be when a software is developed that can create any object you want, imagine, in the polygon setting you want? This is not like creating a Base Mesh using Poser program and increasing the polygon value of that Base Mesh to make it quality. This is when all the tools come together to become one software where everything is done easily. Artificial intelligence develops rapidly. Technology is at a level we cannot even imagine. Programs such as Poser, Meta-Human will facilitate the work of 3D Character designers since they are “tools”. But one day, when a great software (not a “tool”) is developed which everything is done, that crystal ball here will show that what I said for 10 - 30 years later is true.

#12

Without getting too deep into politics or economic theory, what you’re saying about automated entertainment is true.

That said, even cutting costs is not a guaranteed recipe for success. The market can still absolutely demand quality from their products instead of what just took 5 seconds to make.

Great Theft Auto is an example. Despite costing millions to make, nobody except for Rockstar has successfully cornered the open world market like they did. There have been many attempts by competitors to clone their game but they still fall short of selling 100 million units that their recent entry has accomplished.

In the future, people could press a button and maybe create their own GTA clone. But if it’s considered outdated at release or Rockstar manages to create something even better again, than the impact of robots remains the same as it is today.

In fact, that’s the weakness of robots. They don’t innovate, they only create what already exists.

#13

You are Right and also love all comments it is more benefit for my thinking.

#14

Wrote an article about this on my webpage some time ago (and wrote something similar in the AI thread) , time to summarise it here:

Automation in real life (and not what some people think who read to much scifi ) is not that a machine suddenly COMPLETELY replaces the same worker, especially in the first stages.

Usually its done by partly replacing the easy to automation parts by the machine, and use the worker for tasks wich are not that easy. Due to the free time, this person operates more machines / produces more and qualitative higher output with the help of the machine, and / or is replaced by a lesser skilled (and cheaper) worker in later stages.

The same happend to 2D animation with the introduction of computers. Same people, but overall a lot more output, reducing the cost and need for human labor

AI will do the same, artbreeder is usually the site I send people to when they say “creative” jobs can’t be replaced.
No they might not in the next 10-20 years if things work out like they do… but they will reduce the workload drastically for artist before he does his creative decission. Making a 5 people job a one man band show. Therefore killing jobs.

There already people who feed artbreeder or similar faces into daz face transfer or Character Creators Headshot plugin… sure the automated results still suck and you need human intervention, but you can clearly see where things are heading and obviously there is still room for automation of a lot of trivial tasks.

And last but not least, the pressure to automate these jobs due to labor costs is relative high in that department.

And if you observe how many tasks have been made easier / more accessible in the last few years its not a matter of “future” but we are already part of the process.

So in short: Yes human artists will still be needed, but a lot less.

#15

Technologies that shape the future of 3d modeling are AI, including photogrammetry and automated 3D generation.

#16

3D modeling has been used for decades to help create software, art, and technological improvements within their respective fields. Due to all of this, the 3D modeling industry is expected to achieve up to 16.6 Billion USD within the year.

#17

Technologies that shape the future of 3d modeling are AI, including photogrammetry and automated 3D generation.

#18

Basically, we’re at this point where a lot people’s minds are about to break.

Historically, “Art” unbeknownst to a lot of people is about concept/expression and what emotional response it evokes in a viewer based on third party real life experience, which is unique but relatable.

Most people believe art is about technical proficiency, difficulty and tools.

Once you’re forced to accept that you were wrong, depending on how pig headed you are your mind will break hard or quickly and without much damage.

Keep in mind there will always be jealous snobbish professionals within the field that will degrade another artist for even using a Metahuman in the background of what most people (especially non-artists) would consider an amazing image/piece of art.

Could a computer create this with the command “Make Green Wolfish Bust”? :

Could a computer create this with “Make woman bust with Blond braid”?