In August last year I posted the last of a few articles on promising new graphics technology called Unlimited Detail. As I posted there, the team were going to ground to work on getting the technology to a stage where they have something even more substantive to show off.
That may be a little while off yet, but xbigygames.com has an interesting piece on how Euclideon are doing. A snippet:
As mentioned when Euclideon was first revealed, this technology is something they plan to utilise not only for video games but also scientific research. Supposedly there will be “some Euclideon products released in non-games related industries over the next few months”. “There turned out to be a lot of demand for our capabilities across quite a few industries, so we have tried to put that demand in order and address each area one at a time. As soon as we have revenue coming in, we can expand our team into different departments to deal with each industry,” Dell tells us.
“I think it’s fair to say that people are starting to accept that the future of 3D graphics is atomic,” he finally points out. “Polygons will still be around a bit longer as an editing tool, but I don’t know how much longer they will remain for visualisation. So many games today have polygons that are so small that they are only a few pixels in size. When polygons become smaller than the 3 corner points that make them, there is no point in treating them like triangles anymore and it makes sense to use atoms instead.”
On the question, when we will get our next look at Euclideon powered gaming, all Dell responds is, “Well there is soooooo much I’d love to say about that, but I’m afraid that I’m sworn to silence at this point in time. My apologies, but I think you’ll find it worth the wait.”
So things are still progressing and we should start to see some implementations of the tech before the end of the year by the sound of it.
Thanks to Phillip Street for the heads-up!
ok
It would seem to me that if by referring to a vertex as ‘Atom’ that they are implying that the points are applied as color obviously, but imagine if you would a method for making a set of images that were similar to the unwrapping of the UV’s associated with a model just so to where you can guarantee that that there are a fixed number of pixels or points per inch for the surface volume and rendered much like a normal map. In that sense I deduce that the method most likely acts like a special image filter that selects only certain pixels of specific color in reference to the xyz location and rotation of the camera. That would be the only feasible way that I can think of to have the shaders in the previous videos. If you were to do it by image selection with a method essentially similar to the magic wand in photoshop, it would be much simpler to calculate reflection without a ray. To compress and speed things up, they are most likely breaking pi with key and tumbler style sine waves propagation as doing so would grant you the gift of saying something like….
EG what large number when repeatedly added to 1 and the subsequent results when counting from 1 to 10 will give you in perfect order :1, 8 ,4, 7, 9, 10, 4, and 2. You see by inverting our inherently negative counting system to positive you can break pi and thus hold infinite amounts of data as our number system uses infinitely empty or zero to describe numbers getting closer to infinity.
EG difference between 1, 10, 100, 1000, 10000. This is a inherently negative system whose consequence is infinite pi. The other way removes time from pi and thusly gives you the exact locations in infinite pi for which all existence and in particular that piece of data is located, without having to first calculate pi in order to get there. By storing so many digits of correct pi you can use a series of 10 sine waves (being of each magnitude and amplitude from 1 to 10) averaged together would contain in base 10 all numbers that exist in base 10 in all combinations and lengths and orders, if they were to use say base 64 the computation would be extremely fast in assembly through C and allow them to store this data very easily. This method would also allow the possibility to pre-store in advance every single possibility that could occur on frame period. The above example method using walk numbers on each of those sine waves could produce key and tumbler style compression and thus be infinite. The method could allow the ability to pull data out of the future as well, but for fear of my life I can’t reveal that information.
Holy moly, just what has ^ guy been smoking?
That’s some very seriously waked-out bullshit.
Infinite pi into the future? this guy’s obviously smoked a HUGE sack of meth, licked a toad and taken a 5 tabs.