Graphical Revolution

whatever happened to this bullshit anyways

oh well look at this

It was pre-kikestarter so it was forced to fade away instead of making $1B. But no one wants a game where a palm tree takes up 100TB and can't sway or break.
I hope it hurts.

That's cool. Would try that sword fighting game.

Damn, it's the same guy even. He's still talking about fucking atoms (2:01)

point clouds

most of the best demos (demoscene) lately are not using polygons/shaders anymore

based on some of the 1st places in the last 2-3 years it seems like we're maybe 2-3 years away from photorealistic 3d games using point clouds

I honestly can't think of anything really.

Probably not until rigs become so beastly that graphics is essentially trillions of small, microscopic dots that are layered differently.

...

Euclideon is still around, they've had funding from the Aussie government for a while now. Same guys behind .

Before they move onto that they should try making games actually fun again or something. For example, CRYSIS was a shit game.

The next revolution is not in graphics, but AI.

Unfortunately, AI is hard.

i love this meme EVERYTIME

shit I remember years back seeing a video where lighting was as lifelike as possible or some shit. does anyone know what I'm talking about?

how is this possible?

It'd be nice to see pre rendered backgrounds come back

...

UNLIMITED DETAIL

Photorealistic graphics and true biotech based VR are the only two remaining frontiers.(fuck current VR, having mini LED screens in front of your eyes most likely increases the risk of macular degeneration) And frankly the idea of connecting my nervous system to a computer is frightening, who knows what kind of horrors await if a severe enough malfunction occurs.

So… we're going to progress from polygons to LOTS of polygons?

Great. Just great. I can't wait for a game to come out where literally the only selling point is that the player character has a model with 18 quintillion polygons.

Supposedly the advancement in graphical primitives are as follows;

Pixels -> Polygons -> Voxels -> Ray Traced primitives (muh unlimited detail!)

Technically its a ray tracing engine. And its not ready for use in games, its being used currently to make very detailed geometric scans of real world objects, an emerging field in computer graphics, that may be used for other things as hardware becomes better

Simply increasing the polygon count will NOT automatically equate to photorealism. It requires artists who are willing to take the time to make sure the lighting, geometry, and detail of fucking EVERYTHING in the game they are trying create looks identical to real life. It doesn't just require raw processing power, it requires a LOT of effort on the graphic artists part. PIXAR has been able to make short films that look very photorealistic but even those use only a few models and still took months to complete

We can already use polygons to render nearly photorealistic graphics, there's nothing to "upgrade" to.

Any "infinite detail" bullshit is inevitably going to run into memory and/or data storage limitations, there's no getting around that.

Voxels also aren't graphics, it's a volume system. You can render voxels by using polygons.

Is there much of a need to, why completely change the system that would require mass retraining when it works perfectly fine, fuck off Chad Warren.

It was hyped on the angle it was slicing, the half truths it was peddling and the general lack of knowledge of reddit gamers.

No Man's Sky of computer graphics, except the customers were experts in computer graphics the complete opposite of the customers of Nu Man's Sky.

Point clouds, like what 3D scanners output.
The world is made of dots, it's only logical that the last (not sure if next) step is making game assets into dots

they were lying about everything and just profiting off of marketing.
point clouds are much more complex than polygons and even if they weren't current graphic cards would become useless and we'd all have to buy new ones since what we have is build around polygons

If you're asking if we will upgrade from 3d to something else, then no, 3d is the ceiling of something more closer to reality.
But you can upgrade from polygons with faces to voxel objects composed by a near infinite numbers of points.

In the rendering field we could pass from normal rendering to cycle rendering generating a much better lighting system, with real reflections and diffraction.

Generally from the fake 3d of doom to the real 3d of quake and to the color lighting of quake 2, then going to normal mapping and other ways to render better lowpoly models, etc a number on incredibly revolutions have already happened.

When we will be able to completely eschew texturing because our models will become so complex to not require any texture anymore, but just materials definitions, then another revolution will happens.

has ne of dem advancements been significant or meaningful tho monica since lyk 02?

so wut i b lookin fo is shit w/ good graphix that actually aid a nigga in playin a game n da only thing in 3d that really aids that is sumn lyk dat draw distance fammo and in sum cases nicer animations to be more visually clarifying.

What I want to know is, is it possible to create a graphical style that resembles 2d animation put into a 3d environment without resorting to cell shading or using that god awful Live2d weeb bullshit. Ni no Kuni came a bit close, but still looked very off.

what do you need infinite drawing distance for if this isn't even in the real word due to the earth spherical form? You need enough drawing distance to not make things popping out of nowhere.

And nicer animations are indeed achieved with more polygon complexity, but more complexity = more rigging, more work on animations, more time spent.

Notch wrote a blogpost calling out their bullshit as detailed in , back when he was still in charge of Minecraft.

id b mo interested in what could b acheived with a concerted effort in modern low poly n it would remove a lot of the issues that b plaguin modern game in fa first place w/ shit lyk pop-in n hardware requirements and possibly even monica bein mo n mo confined to launch on dem on consoles n windows n sheieet.

jus a pipe dream i spose tho

lyk pic related almost look real

Live2D Euclid's looking to allow for fully "3d" 2d models. It generally only looks weebish because most character art used happens to be weeb-based.

Holy fuck he sounds so smug like he sniffs his own farts for his own satisfaction.

They died with fixed cameras.

Why? They're all the aesthetic drawbacks of 3D cg, without the one and only 3d selling point of of free 3d movement of the viewpoint. Pre-rendered cg is the worst of both worlds, good riddance I say.

I agree. If you ask me, sure, the graphics on things you're currently seeing up close in your typical high-end game are good enough. But as long as the stuff that's far away is clearly simplified and showing pop-in effects as you get closer, graphics aren't good enough. Pop-in really bugs me.

On a similar note, current technology has the muscle to make a few characters at once look great, like say, four to ten. But, there's still clearly downgrades required to have dozens of characters on-screen at once. As long as that's the case, graphics need more power.

And finally, and maybe most importantly, real-time lighting still kinda sucks. If the lighting is truly good looking, it still has to be pre-rendered and baked-in. Which means it's still a choice between good looking lighting, and things moving. Heck, you still see games where they selectively have to choose what things can simply cast shadows or not, and it is really jarring to me at least, when some of the furniture in a room isn't casting shadows, or the trunks of trees cast shadows but the grass growing out of the ground doesn't. Or the light behind a character is showing up inside their mouths because objects can't cast shadows on themselves, and so many other corners that have to be cut with real-time lighting. Like with the other two things, as long as graphics have those problems, they aren't good enough.

So basically, graphics are plenty good enough when it's a few characters in small areas with static lighting. But not good enough for lots of characters in big places with live lighting effects. What we don't need is making individual things better, so much as having more things at once be equally good.

I absolutely love the 2D hand drawn look and all, but I get the feeling that all this software can really do is your standard generic anime human face. Which, sure, has its place, don't get me wrong, but what if you want non-humans in your game or whatever too? What if I want an animu game where it's people fighting monsters of all shapes and sizes, like I dunno, dragons and such. Would this be just as good for taking my drawing of a dragon's vastly different facial structure and turning that 3D while looking relatively hand-drawn? I doubt it. Which is a problem, because while I share Japan's love of animu style in general, I absolutely have no interest in a cast of nothing but mundane highschool students doing mundane things. It's not what I want as a consumer and not what I'd want to make as a creator.

I think the tech for objects casting shadows on themselves is there, we are anyway still far from photorealistic effects like a cycle rendering render can achieve with prebaked rendering.

I'm assuming it follows a general principle of meshes, rigs, and blend shapes, so chances are it'd be possible.
In all honesty though, the more I think about it, the more I feel Euclid's workflow is gonna be more akin to constructing a model in reverse, if that makes any sense.

It is there, no question there are games with self-shading objects. But graphics aren't yet to the point where it's pretty much guaranteed, you still see games where characters don't self-shadow for one reason or another, most likely because of performance issues. As long as we don't have the graphical muscle for self-shading to be standard for everything, there's still need for better hardware.

From the videos I've seen, the whole selling point is basically automating the creation of the face model with just a drawing to start with. It doesn't do anything you couldn't already do the old-fashioned way, and the end product itself is still pretty much the same thing character modelers already make. It just gets that face model made and animated in a way that's more automated and thus quick and easy.

As such, I have a hard time seeing it working equally well on drastically different non-human face shapes, it seems built around those standard anime faces first and foremost, and they're sure the only thing I've ever seen them show it used for. In which case, I could imagine an art direction problem in projects with stuff other than standard anime girls, in which the humans made with their thing look different from the characters/monsters/etc that weren't made with their technology.