http://notch.tumblr.com/post/8386977075/its-a-scam Voxel technology IS impressive and important, but if that's the amount of data they're shifting, that demo is kind of bollocks, it also isn't anything new, the technology has been around for ages, just limited by rendering speed, which I don't imagine they've been able to solve. Even if they have, it's being massively oversold, calling it 'atoms' rather than voxels seems to be some silly marketing ploy to make it sound new, which it really isn't, they were doing emergent depth voxel engines half a decade ago, the only difference seems to be that this one is rasterised rather than procedurally generated, which means a shitton of data to store and shift, to the point where it totally isn't commercially viable. The theory is good, and will be coming into prominence as optimisation technology for this rendering approach gets better and hard drives and transfer speeds get bigger, but eleventy billion jiggabytes of scene data in this day and age is bullshit. It isn't going to work.
it won't work practically and it's not some great revolutionary thing the fact that they claim so makes me feel a nice bit of schadenfreude
As the video progresses the guy shouts louder and louder, but i'll have to say, I don't think any of the big engine names (Unreal, Id, etc.) will be using this stuff for a long time. The guy said that the new generation will make fun of our games the way we make fun of the atari 2600, but they'll probably see this tech demo and laugh at how someone would think it possible. Like who is going to go prone and look at the ground instead of shooting people, who's going to pay attention to the rocks, which actually are the perfect objects to be done with polygons. and 65 voxels per cubic millimetre, are games going to come out where people pick up clumps of dirt and look at them under microscopes? in summary: who the fuck is going to make a supercomputer out of playstation 2s just to play a game like that.
Um.. PlayStation 3's are mini supercomputers. Just crappy graphics. From what I gather tho... its technically possible to do this with games, just incredibly infeasible.
The main appeal of voxels is unified development, rather than making polygons and sticking textures over them, instead you carve out of voxels and then paint said voxels to produce surface colouration. Essentially, we currently stick raster images onto vector polygons, with voxels, we could use raster voxels to replace both, unifying textures and geometry, all of the cool stuff you can do with textures you could do with geometry, for example as you can change pixel colours with pixel shaders, so you could change things about voxels with the same technology, creating bulletholes that actually dig into objects and go through them, causing three dimensional burn effects and suchlike. It would be a major leap for destructable environments, destroying the floor would be no harder than destroying a building and both would have incredible degrees of fidelity. LOD meshes are replaced with voxel-mipmapping, simple concatenation of voxel groups into one larger voxel just as you do with textures in the distance, and it can be done entirely procedurally because there is uniform voxel density, unlike with polygons where density is all over the place. It would massively, MASSIVELY cut development times and allow incredible amounts of detail and moreover, incredible feats of artistic flair to be put into games with far more ease than we currently can. There are major challenges, as stated the animation problem is still unsolved as far as I know, but the future almost certainly will use voxels, it will just take a long time. 10-15 years ago people were saying 'why would anyone want 3d, all these new 3d games are crap, we want the good 2d games back' and now 3d has given rise to entire new genres. Expanding technology leads to entirely new types of games.
I just noticed your join dates, I always thought mayama was around a lot longer then Trickster, weird. Also about the topic: Gief SDK to play with and then we'll talk, until then I don't really care what they are doing. But yeah they haven't shown anything that is fundamentally better then the last video from a year or two ago. They have figured out they can scan real world objects and make static light maps. Great, nothing that resembles a game even close. I have no idea why they are trying to market that for games, they could make a neat biomedical visualization tool and get billions, but no they are trying to force technology upon gaming that just won't work without some big research breakthroughs.
Generally getting it to work with the scanning software. And not marketing it by saying how much better it is than polygon videogame graphics.
Nothing in this video is fake. They just dont talk about the problems... they make it look like the messiah of video gaming. ID software mentioned in an interview that the next ID engine might be something like this.
That is the great part about technology....It don't have to work. All you need is the idea. Yes what they claim might be bullshit but I am sure the CEO Sony Nintendo, Nvide and so on are all looking into this IDEA. So even if these people can't make this idea work like they claim...some one will or they will come up with something better.
So if the two main drawbacks are lack of animation and lack of ability to create many unique voxels, why not just use these for scenery only (trees, statues, things that will appear over and over again) and polygons for rigging and animation?
