Hacker News Reader

Comment by hypertele-Xii | Hacker News Reader

As an artist and game developer it disappoints me how much resources are wasted in pursuit of "realism".

Consider how art exploded when people realized they could paint something other than, you know, reality.

Reality is boring. Get your realism off my games. Give me fantastical alien worlds to explore rendered in stylized and weird looks.

View on HN

Realism in regards to rendering software has nothing to do with realism in the final product.

Pixar and Disney have been among the most influential and revolutionary companies in terms of rendering techniques. And most of their products are stylized.

Stylized looks come from the assets, not from the rendering engine. The rendering engine gives you physically correct light interactions (which you almost always want, and even then you can turn knobs to make it non physically correct or outright do whatever you want using shaders), things like fur/hair rendering, fluid simulations, volumetrics etc. All these things can be tweaked to non realistic values by the artist.

If you want a movie/game where characters are made of wax, or fire, or smoke, or something even crazier with shaders, you want as a basis a renderer that is physically based and uses mathematical models based on the real world, and then tweak from there.

Then your art will always resemble the physical reality we already know. Computers and mathematics enable us to render literally anything. And all you can think of are things you see every day. I'm looking at wax, fire and smoke right now, there's a candle on my desk. If it's realism I wanted, I could just stare at that (and sometimes do).

Could you give me an example of what you have in mind that current rendering engines can't handle?

An example would be good, but I think hypertele-Xii's point isn't about what engines can handle, but about where engine developers are (and aren't) spending their money and time, and what they're making easy for users to do --- that if they make realism easy, then that's where users will go.

Not OP, but I think rendering scenarios where light doesn't follow straight lines would be problematic for current rendering engines. Also, some hypothetical portal or other dimension scenarios.

Sure.

Here's someone developing an engine capable of rendering infinite, fractal worlds:

https://www.youtube.com/watch?v=V1MjZbGbi4I

Here's someone developing an engine with hyperbolic geometry (though dressed in mundane setting):

https://www.youtube.com/watch?v=VYfWfrk5P7w

Mirror Drop is an abstract art puzzle game that's kind of difficult to define:

https://www.youtube.com/watch?v=e1zHgYUE9HY

Here's a game where the speed of light is simulated and variable:

https://www.youtube.com/watch?v=uu7jA8EHi_0

The primary point of my argument is that we don't have things in mind that current rendering engines can't handle, precisely because our current engines are so hell-bent on replicating reality. Why dream of what your tools can't make?

I really enjoyed games like Red Dead Redemption 2 and CyberPunk 2077. They are not per-se reality as in as you and I live today, but different lifestyles in different times.

I feel like these types of games really benefit of the work that is put in to making humans and environments look realistic.

Heck, even if you're making some weird alien world you're still going benefit from realistic physics engines for example, because that's just how our universe works.

Imagine another universe, that works different from ours.

So go and do that, but unless the way that universe works is defined as "whatever is easy for a computer to simulate and render" then you will still likely benefit from techniques that were originally developed to simulate the real world.

Agree that I tend to prefer stylized non-photorealistic rendering in games.

That being said, I think there’s room for both and luckily, the rendering learnings from one often improve the other.

Nothing wrong with exploring both venues.

That's totally cool. We're all entitled to like and value different things. The awesome thing with video games is that because there are so many of them, one can pick from a spectrum of realism. It's not an all-or-nothing problem.

There are a ton of people who value photorealism, and that's why resources are being put into it. And it's certainly not seen as "wastefulness" by those people. Personally, I would absolutely love to be able to put on a pair of VR goggles and be unable to tell if I'm viewing a real-time render, or a remote stereoscopic camera feed of a real scene. Yet, I don't think that fantastical physically-unrealistic rendering is a "wasteful" use of resources.

(photo)realism is not inherent to gaming. the end result would be a perfect simulation of the physical world.

If we were discussing a video game, I'd agree, but we're discussing a tool for creating video games. A tool that influences creators' understanding of what they can make. If the tool becomes opinionated, it will produce works accordingly. When Unity makes hyperrealistic humans easy to make, that's what we'll be seeing people make with it.

Even if we got to the point where the digital humans are indistinguishable from real footage, all your realism is gonna go out the window when you encounter a bug where a character is half-clipped into a wall. Or when an animation breaks and someone is standing with a backwards bent knee, and no reaction on their face.

The idea that the ultimate endgame for video games is to have them look like a playable movie, or be a real-world simulation just seems silly to me.

Not if your game is themed that way like a Philadelphia Experiment, Fringe, or other scifi where people get cinched between dimensions.

I find that the increasing photo realism makes it even funnier when things break down in the typical way.

Photography replaced painting. I don't know about boring, but reality wins.

Painting is still a thing.

If anything, I would argue the opposite.

Photography gave painting (and visual arts in general) a massive boost by the virtue of stripping the aspiration of going for "ultra-realism" in paintings. After all, that's what photography can accomplish much better and much easier. This made painters look for other directions to push their art into, which freed up artists' bandwidth for a lot of other, previously overlooked and not as popular genres, such as surrealism.

Not saying that photography is responsible for creating those genres, they existed before. But (relatively) cheap and accessible photography definitely diverted attention from ultra-realistic paintings to other painting genres that enjoyed greater growth due to that.

After all, what do you think will happen once it becomes very easy and cheap for anyone to have ultra-realistic 3D models and animation in videogames? It will become much less valued, as it is something that anyone can do very easily. So people doing 3D art will need to look in other directions to distinguish their work. I have already been observing it happen over the decades in gaming.

For present day examples: Nintendo manages to nail the style of visuals in a lot of their first party titles without even attempting photo-realism. Guilty Gear -STRIVE- (made by ArcSystem Works) is killing it with their over-stylized visuals. Atlus and their Persona series keeps pushing more abstract and surreal depictions with each game, and they resonate well with the audience. And I've only listed AAA-tier examples. Once you dive into indies, you can go on forever with lots of well-executed products that pretty much intentionally went against photo-realism.

I think the plethora of filters and their popularity totally disagrees with this. Also, things like HDR processing to make things look eerily unnatural, desaturation f/x, etc are all ways of taking the reality into a new place.

What do you mean by this? I can't think of a part of my life where I use photographs, where someone in a different time would have used paintings.

So in other words, "stop being so creative, you're stifling my creativity!"

What if I told you you can have fantastical alien worlds using stylized and weird rendering that are also physically correct?

Physically correct is the antithesis of fantastically alien. This is the opposite of creativity. This is imitating reality.

As a serial buyer of games, I'd like to disagree (if only to prove that there is at least one). I often buy games just to marvel at the technical aspects and realism of them. I find realism and reality extremely interesting. Now, "resalistic" doesn't need to be "everyday" looking. Take CP2077 for example.

Reality is also hard and that's why people aim to reach it. Making alien world is easy because the viewer has no reference point, so you can go banana with it.

It is the opposite. Reality is easy because it is familiar. Our brains are tuned to it from the core. To make something truly alien is to make something you've never seen, heard, touched, smelled, nor felt before. That is incredibly difficult, has always been, and has the potential to birth whole new genres of art.

There will never be new genres or art or music. Everything has been discovered already. Like colours, they are all already known. Sure you can have a new take something or maybe discover a new subgenre, but the majority is already there. And just because a genre would have a new "name" (just a branding thing anyway) doesn't mean it's not part of some other genre already.

You can't capture psychedelic and or mystic experiences, which are very subjective anyway and yes, rooted in _reality_ or require a reference to _reality_ as they are the so far away from that.

> There will never be new genres or art or music.

Do you believe that there were no genres of music born in the 20th century, or that we conveniently discovered all the genres just a few decades ago? Neither of those seems reasonable to me.

You can't automate art. AAA is in large part about converting money into more money by way of video game sales. That's why it shouldn't be synonymous with anything elite.

Surely you can automate this:

https://en.wikipedia.org/wiki/Black_Square_(painting)

you really couldn't

You could always say "I could have made that", but the important part is: you didn't.

No one cares if you automate the painting of a black square. They care about the intention of the mind that thought it was important enough to do. That's why it's art.

Good luck automating the part where you convince people it's worth something.

Step 1: Add subtle "imperfections."

Step 2: Don't tell anyone that you automated.

Step 3: Publish a personal account on what "experience" this particular piece represents.

You can certainly pour all these resources being spent on rendering 'hyper-realistic digital humans' into developing tools, workflows and processes that give artists power and utility for crafting wonderous worlds.

Eh, keep building this technology and give me the ability to push a script into an AI and it render my world for me with a little tweaking here or there. We're not there yet but it recalls a moment in one of Iain Banks' novels where one of the minds "lies." Since it would be trivial to generate a false flag that had all the data necessary to prove an event "happened" you had to trust that he mind wasn't lying about what it witnessed.

What is truly astonishing to me is we are close, not quite there, but close. I remember when The Spirits Within came out in 2001 and we were so blown away. I look at it now and it feels so amateurish in comparison.

Reply on news.ycombinator.com