Euclideon


bzar

A Commando
Joined
Sep 22, 2008
Messages
4,500
Location
Finland
Website
Visit site
HardOCP article


Couldn't find any threads on this in gp32x nor here. Maybe my google-fu is just inferior, but I'd like to hear what people here think of this. People have been having very mixed feelings about if this is legit or not, and what are the boundaries of this technology. Games industry celebrities like like John Carmack and Markus Persson have commented on this and seem to have quite differing opinions. We can all take our best guess at what we're seeing here (besides the explanation they offer), but they seem to be very confident that the technology is not as limited as people make it seem in their comments.


My first impression was a lot like notch's, because voxel stuff like this has been done before and is known to have severe problems with animation and transformations. The fact that the terrain is highly repetitious doesn't help here, because it makes the entire structure seem like a rigid BSP tree with references to repeating subtrees (like the elephant statue) that's just searched for the first intersecting "atom" (as they call them) for each pixel of the screen. They claim however that stuff like skeletal animation is possible with their engine, which makes me think there's more to it than meets the eye. If the structure is dynamic and modifiable in real time with sub-partition movement (not just from one bsp cell to the next, but with actual coordinates) and scales well, they may be on to something.


Or they could be just scamming for money :p


Anyway, just want to hear the opinions people here have :D
 
I think that for static objects (landscapes, environments etc) this will be an absolute boon. If it removes a lot of the burden from the GPU, then more power can be concentrated on making the things that really matter - such as animated models and lighting effects - much more detailed. Higher poly counts are not desirable, but they are the quickest route to higher quality graphics given current technology and this may help achieve that.


I cannot logically see how this will be of benefit in anything but a static scene.


Another thing to note is that this tech might be licensed and if so I sincerely hope that the guy who narrates Euclideon's videos doesn't demand a voice-acting role in games that use this - sales would plummet.


D.
 
it looks cool. i don't really get why everyone is so up in arms, but then again it is the internet so...


if it works out and is successful then that's just marvellous, otherwise hey its good to know people are trying something and even if 10% of their ideas leads to improvements in gaming graphics then that's still a good thing for the industry.
 
I've been following this for quite sometime now, even when they dropped off the face of this planet with no news of it whatsoever, so really all we can do is wait out, I guess.


Time will tell us what becomes of this all.
 
Last edited by a moderator:
Seems like every time they come out to show something nothing has changed much. Its always massive repeated environment.


Given the memory overhead of point cloud data and the repetitive of many of the same objects that there might be a high memory consumption.


Also these recent demos do not show any lighting/shadows, animation. Although I wonder if they could have a hybrid of point cloud for static objects and normal polygon shapes for animated objects.


But we will see if any of those questions are resolved in another year.
 
I think these concerns over transformation and animation were said of pixels, as well. It will simply take time. Voxels are still relatively new compared to polygons, so there _will_ be growing pains and hurdles. No good technology gets developed overnight.


Though I am skeptical at the "infinite" detail.
 
Back
Top