Slashdot Mirror
Euclideon Teases Photorealistic Voxel-Based Game Engine
MojoKid writes Not many would argue that current console and PC graphics technologies still haven't reached a level of "photo-realism." However, a company by the name of Euclideon is claiming to be preparing to deliver that holy grail based on laser scanning and voxel engine-based technologies. The company has put together a six-minute video clip of its new engine, and its genuinely impressive. There's a supposed-to-be-impressive unveil around the two minute mark where the announcer declares he's showing us computer-generated graphics rather than a digital photo — something you'll probably have figured out long before that point. Euclideon's proprietary design purportedly uses a laser scanner to create a point cloud model of a real-world area. That area can then be translated into a voxel renderer and drawn by a standard GPU. Supposedly this can be done so efficiently and with such speed that there's no need for conventional load screens or enormous amounts of texture memory but rather by simply streaming data off conventional hard drives. Previously, critiques have pointed to animation as one area where the company's technique might struggle. Given the ongoing lack of a demonstrated solution for animation, it's fair to assume this would-be game-changer has some challenges still to solve. That said, some of the renderings are impressive.
Oh wait, use the higher-quality video link at https://www.youtube.com/watch?...
The 360p version of that beats the 1080p version of the original.
This company with it's impressive-looking but completely static scenes shows up every few years. Honestly, I didn't see anything that couldn't be done in that video with any modern engine targeting high-end video hardware. It's a bit of a cheat if you only have to show the terrain. I'll be more impressed when I see a demo with physics, animation, and dynamic lighting, because that's where things tend to get tricky. They mentioned in the video that they do have animation working - I'll be curious to see how it looks in the next video.
This company seems to be trying to solve the problem of how to accurately capture and reproduce the real world, but how many games actually want to capture real-world data? If you're in the business of creating fantasy worlds of any sort - and that's precisely what most games are - there's nothing in the real world for you to scan. There's a reason no one else is working this way, I think. As far as the game industry goes, I'm guessing it will probably remain a very niche product, if it's viable at all. I just don't see them throwing away 15 year's worth of maturing polygon-based tools and technologies anytime soon.
Irony: Agile development has too much intertia to be abandoned now.
"Dead Meshing" a real environment, and then shooting enough JPG compressed digital photographs to texture the mesh from most possible viewing positions is the very WORST way to handle the problem of real-time rendering of open world environments. And BTW, this method has ZERO to do with 'voxels'. But periodically we get this same nonsense sold to fools alongside the cretinous "ray-tracing uber alles".
John Carmack- an individual with intelligence vastly greater than the con-man behind Euclideon- attempted to follow this approach with every state-of-the-art programming method. He created the so-called MEGA-TEXTURE method, which solves the only interesting part of the Euclideon problem. The result was dreadful- utterly hopeless. 'Rage' and 'Wolfenstein- New Order', although both games with merit (especially the later as an old-school FPS) had textures and meshes vastly worse than the best of the current AAA titles using traditional rendering engines.
A back-of-an-envelope calculation quickly demonstrates to anyone what the problem is. To create unique textures, even with feasible compression, for even a modest environment, takes horrific amounts of memory. There is a good reason the Euclideon conman talks about streaming his data from a multi-Gigabyte hard-drive. Even a relatively tiny area (by current open world game standards) needs such astonishing amounts of texture data.
It gets worse. No modern rendering engine relies on 'dead' textures with pre-baked lighting any more. Indeed what most of you think of as the 'texture' is often less than 20% of the REAL texture data- the rest being texture data that defines the 'material' and lighting properties of the surface.
-Programmable shaders with per-pixel lighting
-realtime shadows
-realtime day-and-night light cycles
-complex tessellation
-multi-layer textures combined with alpha-transparency
These methods allow a limited data set of textures to cover a vast open world area with unique visual variation or the textures in each different location.
Pre-baked, dead-mesh and dead-texture methods, as sold to idiots by Euclideon, are from the stone-age, and look 'good' in controlled demos only.
Carmack destroyed iD perusing this same hopeless method, and he understood its limitations from the off. Carmack's mega-texture literally has ZERO advantages over current methods, and extraordinary numbers of downsides- the worst being the disruption of the ordinary artist development pipeline. Modern games on modern engines, by comparison, are crafted very much as Hollywood builds (and lights) real and digital sets for their movies.
Megatexture- the infinitely more intelligent approach to Euclideon methods, is just horrible for artists and designers- increasing development time (with no visible advantage over traditional methods) by a depressing amount. The processing of Megatexture data for a test build takes a cluster of power PCs working for a long time. A modern AAA gaming engine allows REALTIME visualisation of new assets in the actual game world. It is a bad joke to even compare the two methods.
Euclideon's bigest con is the ability of such a limited rendering system to use JPG quality compression of the world textures in near real-time. A gaming environment, on the other hand, cannot use compression anywhere near the quality and efficiency of JPG. Dead meshing using dead textures, JPG and simple streaming can appear to achieve 'miracles' (if you are a credulous fool). No useful real world renderer can rely on these methods.
Here's a 2011 interview
I’ve revisited voxels at least a half dozen times in my career, and they’ve never quite won. I am confident in saying now that ray tracing of some form will eventually win because there are too many things that we’ve suffered with rasterization for, especially for shadows and environment mapping. We live with hacks that ray tracing can let us do much better. For years I was thinking that traditional analytical ray tracing intersecting with an analytic primitive couldn’t possibly be the right solution, and it would have to be something like voxels or metaballs or something. I’m less certain of that now because the analytic tracing is closer than I thought it would be. I think it’s an interesting battle between potentially ray tracing into dense polygonal geometry versus ray tracing into voxels and things like that. The appeal of voxels, like bitmaps, [is that] a lot of things can be done with filtering operations. You can stream more things in and there is still very definitely appeals about that. You start to look at them as little light field transformers rather than hard surfaces that you bounce things off of. I still wouldn’t say that the smart money is on voxels because lots of smart people have been trying it for a long time. It’s possible now with our current, modern generation graphics cards to do incredible full screen voxel rendering into hyper-detailed environments, and especially as we look towards the next generation I’m sure some people would take a stab at it. I think it’s less likely to be something that is a corner stone of a top-of-the-line triple A title. It’s in the mix but not a forgone conclusion right now.
In 1999, he was working with 3d "light maps".