Tahir Khan

Read Full Story Here.
_________________________

A new graphics start up has shown off their new 3D software engine, that they claim, does away with polygons

Judas

awsm

I laugh at the fact that they mention that they couldn't even get into the board room due to them having a fit there are no polygons.

__________________

LeanWolf

clever, I like it!
"one's called ATi and the other's called nVidia, and they don't seem to like each other very much"

__________________

Gastrian

Even the plain explanation is straight over my head. If its searching for these points where are they stored and how are they stored?

__________________
Current PC
Athlon XP 64 3200+ (S939)
2GB DDR RAM
128MB GeForce 6600GT
Windows 7 Beta/Win XP Pro

Judas

X/Y/X per point i would assume..

__________________

IvanV

Interesting. If that thing really can work with a reasonable performance, it's what Intel needs to do away with ATi and Nvidia.

I'm still not sure how it works. If I understood correctly, the rendering process is similar to the current one, except that object models comprise of just a huge number of coloured dots? Still, it would seem to me that they would need to fetch and take into account more than just one dot to accurately present the colour of a pixel - if the scene consists of billions of dots, then there are thousands of dots which fall into one square in the pixel grid, though I suppose that they could get away with taking only a handful of them and blending them easily enough.

But it seems to me that there's one huge hole in the plan. As someone said, there is no animation to be seen anywhere in the video. A moving scene would require geometric transformations. In the traditional approach, when doing such transformations, the coordinates of vertices are calculated and vertices are - dots! So, in essence, the object models in both techniques are made of dots and when doing transformations with the new technique, new coordinates would have to be calculated for the dots making the models the same way it's now being done with the vertices. The problem is that in their scenes there are more dots than we now have vertices by several orders of magnitude.

__________________
Feed the porkers and don't touch the buttons!

BlueMak

I hope it is real, but, something doesn't feel right.
I get the point and how it works, but....just...

__________________
The people who are regarded as moral luminaries are those who forego ordinary pleasures themselves and find compensation in interfering with the pleasures of others(Bertrand Russell)"You go into Afghanistan, you got guys who slap women around for five years because they didn't wear a veil,You know, guys like that ain't got no manhood left anyway. So it's a hell of a lot of fun to shoot them." - Lt. Gen. James N. Mattis
This is slavery, not to speak one's thought. [Euripides-The Phoenician Women (c.411-409 B.C.)] http://www.macedonia.info/FALLACIESANDFACTS.htm Sic semper tyrannis.

EfrainMan

Doesn't tessellation make blocky shapes rounder and smooth out lo-poly to hi-poly transitions when zooming in/out? Without coming up with a new 3D modeling technique? I do like that it can give you some very detailed models without slapping together an endless amount of triangles (which might still look blocky), but wouldn't it be more efficient to reserve this for models that should be round and/or detailed? No need for such a complicated technique on a side of a building or something, a pair of triangles will do just fine.

And I think I know what doesn't "feel right." Everything seems to be made out of a fluid. Nothing seems solid. Mebbe with more dots (MORE DOTS MORE DOTS OK STOP DOTS ) things would look solid, but from the videos everything seemed like it it was about to fall apart into a gelatinous goo.

__________________
My modest video card history: S3 Virge, 3DLabs Permedia 2, 3dfx Voodoo3 2000, 3dfx Voodoo5 5500, GeForce 3 Ti500, GeForce 4 Ti4600, Radeon 9700 Pro, Radeon X800XT, Radeon X1900XT+Crossfire, GeForce 8800GTX, Radeon HD4870x2, Radeon HD5870

Judas

Yes tessellation is something ATI had developed long ago but required special engine support..

it makes rough edged objects appear much rounder and more realistic.. however it's neither amazing or perfect at all.

Where as point per pixel rendering would solve all this... as demonstrated in the video.

if all you have to draw on the screen and caculate is what is to be shown on the screen, that reduces a craptone of workload.

Original 3D was also unanimated, so baby steps aside, it's not at all hard to conceive that not far off from now we could see another video detailing and showing off amazing animations that aren't possible at all today or tomorrow on typically accepted rendering methods.

__________________