Games company claims their graphics are 100,000x better
ausgamers.com
ausgamers.com
The second problem is that 10^4x improvement in level of detail does not mean 10^4x aesthetically pleasing (or in fact, more aesthetically pleasing at all). Ray tracing gets very expensive the moment you start adding multiple lights, specular materials, partially translucent materials, etc. It is very, very difficult to do that in real-time even with standard geometry, let alone with 10^4x more polygons. This is why their level doesn't look nearly as good as modern games despite higher polygon count (compare it to the unreal demo: http://www.youtube.com/watch?v=ttx959sUORY) They only use diffuse lighting and few lights. In terms of aesthetic appeal of a rendered image, lighting and textures are everything.
Furthermore, one of the biggest impacts on how aesthetically pleasing a rendered images looks is made by global illumination. That's also something that's extremely difficult to do in real time with raytracing, but is possible with gpu hardware with tricks. The trouble is, these tricks look much better than raw polygons.
Again, I love what they're doing. Real-time ray-tracing is without a doubt the future of graphics, but it would be nice if they were a little less sensational about the technology, and more open about the limitations and open issues.
Then there is memory. The elephant is looking great, no doubt about that. But I think you will need a lot of space for it. On a PC this might work, but I'm not sure it can be used on consoles.
They might have to resort to polygonal collision models in the same way that polygonal games end up using low-poly collision models (with the same pitfalls such as moonwalking, blocked projectiles or mystery-bouncing).
But overall it's a great demo. And you see more and more games using voxels for surroundings. So ofcourse you could combine this technique with polys.
Here's to at least getting a hyper-real Myst someday :)
I do feel that this author is a bit further along at something ship-able than the Euclideon folks are though.
Realistic movement is a more important factor than detail, you can have 8 cubes move fluidly enough to immerse in. 1,0000,0000,0000 polygons seems un-important when you have pixel shader's and normal mapping.
Maybe they should of made duke nukem forever with this engine, as it seems like its from a similar era.
The new frostbite engine 2 a million times amazing, because of the dynamic lighting and radiosity. Plus the amazing fluidity of the animations.
Their technique is based on point cloud rendering. My supervisor has proposed already in 2004 how this can be done on a standard PC. Look at his PhD-Thesis [1].
This technique is usable for static objects as well as for dynamic objects; however, using it for dynamic scenes additional acceleration structures are required (besides of the Octrees) in order to dynamically change according to the deformation of the object. To clarify on several comments made here:
- this is a rasterization technique
- they used acceleration data-structures like Octrees, kd-trees, BVH, ... (which exactly they don't tell)
- the graphic "aesthetics" depends actually only on the artists and not on the technology itself, so not a good point
[1] M. Wand: Point-Based Multi-Resolution Rendering. PhD Thesis, Wilhelm Schickard Institute for Computer Science, Graphical-Interactive Systems (WSI/GRIS), University of Tübingen, 2004.
+1. I'd actually trust the company more if they had someone who spoke in measured tones and laid off the hyperbole.
A thousand times this!
Are you sure about that? He mentions atoms, but then the video also mentions that they're using procedurally generated graphics, which is something entirely different. Also, voxels would not work on the scale they are demonstrating, you'd need way too much memory.
"Re Euclideon, no chance of a game on current gen systems, but maybe several years from now. Production issues will be challenging."
https://twitter.com/#!/ID_AA_Carmack/statuses/98127398683422...
I don't know enough to say if it's the same approach, but it's certainly the same field, so you'd expect him to know a fair bit about it.
That said, I also view his perspective in this field as tending towards unquestionable. He's always seemed to have had a pretty sparkling reputation as well as the obvious talent.
> @Foggen insufficient information to say if it is tracing or splatting.
They state their method is based on a well-known technique used in engineering and medical visualization, and splatting seems to be used there. I'm still not 100% clear on what splatting is, but there's a bunch of papers applying that term to voxels, e.g. http://graphics.cs.cmu.edu/projects/objewa/
[0] not to be taken literally
Their claims of being 100,000 times better than current technology is also "frustrating", as it is repeatedly mentioned as if you've for some strange reason forgot it in the last 30 seconds. It is also a lie if they cannot use this in a game: I would suspect it's not impossible to increase the level of polygons to 10-100 times the amount - if not even more - if there's no animations or dynamic light sources around.
Each light has a shadow map (which can be thought of as "the depth of the scene, from the point of view of the light").
During the final rasterization pass, the shadow map is sampled. If the sampled depth is less than or equal to the current fragment's depth, then the fragment is in light; otherwise it's in shadow.
So, nothing has fundamentally changed here. When a particle (voxel) is rasterized, it outputs a depth, just like a triangle outputs a depth.
tl;dr: Dynamic shadows work fine.
So... where were they in the video?
There's a difference between "some voxel engines can have dynamic lighting" and "dynamic lighting works with this particular engine".
It's not like this is some obscure or useless feature that nobody has ever heard of. In fact for all the "detail" one can't help but notice the number of other things missing from the video.
I meant to imply that Euclideon are lazy and haven't implemented dynamic shadows yet.
There's no fundamental reason why dynamic shadows wouldn't work in voxel engines. But whether you believe me or not is of course up to you.
See his papers page: http://artis.imag.fr/Membres/Cyril.Crassin/
I am curious what methods Euclidean are using, however.
Sometimes you have to take a step backward before you leap forward.
Granted, both of these features are somewhat hacked into the static BSP level structure, but this only shows that it's "necessary" to have dynamic lighting and moving stuff to offer a good experience.
If anything you could create a pretty interesting art style with hyper-realistic backgrounds and cel-shaded characters, or similar. Even as a replacement for pre-rendered background scenery or out-of-level elements it would be useful.
And (thinking of the Game of Thrones titles story) whatever mojo they've got must surely be able to be applied to other industries, eg. CGI - if Weta or Pixar could improve the poly count of their static backdrops without just buying more rendering farms, that's gotta count for something.
I am pretty sure that their tech depends on a few types of repeatable data, which they are able to cache effectively based on rotation -- in other words, they have come up with an efficient way of querying the front-facing voxels in a large set of data based on the resulting view matrix. Where this falls flat is if the data is not procedural, or it is not diverse - as you can see in the video, there is a bunch of the same data copied over and over. However you compress it, such detail is not free and I am guessing there is a lot of data that depends on a good deal of memory/storage to work properly.
I am not so worried about animation or dynamic lights or textures as everyone else is. If they can render it to a buffer and get the normals/depth/UV coordinates, the rest of the rendering can be done in screen-space, including SSAO, deferred lighting, and similar rasterization tricks. Animation can also be rendered on top of the scene, and intersected with the former depth buffer. The only thing I am worried about is the size of the data set and ability to create more diverse landscapes.
It is not likely to be the exact the same technique, but I am guessing they are using similar methods.
So much work left to do.
* demos show nothing new from a technological perspective
* the presenter sounds like a door-to-door salesperson
* as it seems to me, the only purpose of the demo is to raise a hype and somehow (I still don't understand) they succeeded
Euclideon got financed by Australian government.
I really hope the board took a critical approach and relied on at least /some/ technical expertise to grant these people A$ 2m. If they made this decision based just on a demo - I'm moving to Australia at once, where I will invent a technology you have never seen in your whole life before. Ever.
Terms in that program are you have to match the government funding 1:1 so they must have (or raise within 2 years) $2mil to put against the government's. They also have to pay it back "on success" (5% of revenue once that reaches 100k total) and are monitored closely for "fast failure" (they are expected to succeed and repay within 2 years, they have to repay even if they fail after 5 years). Euclideon claims to have had a 2010 funding round so maybe that's how they got into this program. It also says this program is not to be used to "Prove to the applicant that a certain technological problem can be overcome (R&D projects)" so they must have shown it as a viable product that just needs to be packaged up for sale.
What strikes me most is anything under a Commonwealth funding agreement has to have the words "Funded by Australian Government through the XYZ Program. An Australian Government Initiative" in all their promotional material. Yet the shining star of Commercialisation Australia's portfolio forgot. Ouch.
But it's a much higher quality demo than Euclideon IMO, from a technical standpoint.
I could see it being great for a hybrid-RTS game like Ground Control, or some kind of god-game.
Their polygon count is impressive and the object detail looks awesome, but, as other people commented here, I wonder how well it will hold up when dynamic objects, animation and dynamic lighting are added.
Most obvious comparison comes in right at the 4:00 mark, where you can see the "polygonal" grass waving gently in the wind, then at 4:22 their "unlimited detail" grass appears to be carved out of some sort of grass-colored rock. Also observe how at the end they are excited about their improvement in "the lighting", by which they mean, the lighting that you will get.
Why can't we take the good from both and get better results?
IMO they're doing the graphics industry a disservice by putting out demos of a half-finished technology in their rush to "be the first". The technology can work, but Euclideon's showmanship leaves much to be desired.
Check out:
http://graphics.ucsd.edu/~henrik/images/subsurf.html
So, unless they can do all this AND ray trace it at the same time, it really won't make my game experience 100,000 times better.
HL2 has some pretty amazingly convincing faces, and that's obviously not because it has the best rendering engine.
(I don't mean convincing in a "mistake-for-real" way, but in a "lets-me-suspend-disbelief" way.)
Still though, the artists and actors did a phenomenal job with this game.
One problem where in my opinion the advances aren't exactly exponential is body language. If I'm looking at the NPC talking to me, it's not just his facial motions, teeth etc. that I'm aware of, it's also the movements of his body - shoulders, arms, etc. This is still pretty bad. Most of the time they're just flailing around in a pretty uncoordinated manner. It gets better for highly "scripted" scenes, but the usual "shrug / fist pump / scratch yourself" animations are bad. They've gotten pretty good at the purely physical parts, i.e. what muscles and body parts have to move if something connected to it moves (skeletal and muscular animation), but there needs to be a better "body language AI".
To many bold claims & deception in that vid. Colour me skeptical.
In short, the tech sounds great, but will it result in better games or just more expensive games (and in turn, fewer games with more innovative but risky design)?
What would my mom say? This game looks a bit better than this other game? Or this game looks 100,000x better than this other one :)
Improving the rate by 100,000 times is only a six step improvement and that's if they haven't surpassed the limits of the human eye. I much prefer 72 frames per second to 36, but giving me a thousand frames per second is a waste since my optic nerve can't process most of them.
Open source implementation: http://code.google.com/p/efficient-sparse-voxel-octrees/
Yes, they might implement some kind of a hybrid approach but then they will lose all the things they hype about.
For sure their graphics are not 100,000x better.
If the claims from this video are legit, is this level of innovation deserving of a patent? (Yes, a software patent.)
I hope they get this out to market sooner rather then later.
You don't want to compute physics against your art representation, because it's typically over-described.
Of course, you can dump physics onto the GPU, but that will cost you polygons. I guess that could, in theory (if you really had unlimited atoms), give you more cycles for physics.
Fuck this shit. Back to Minecraft.