Unlimited Detail, octree traversals

[Alucardx23]

Is this a technology that shows promise for this and next-gen consoles? It seems too good to be true, so what’s the catch.

http://www.youtube.com/watch?v=Q-ATtrImCx4&feature=player_embedded
http://unlimiteddetailtechnology.com/videos.html

 

[draconian]

fascinating...

Not to be egotistical, but I always felt there were lots of computational waste in 3d graphics since there is only so many pixels on the screen.

 

[Alucardx23]

Not to be egotistical, but I always felt there were lots of computational waste in 3d graphics since there is only so many pixels on the screen.

The video is made by a small company, but it shows how even games like Crysis are polygon limited, Crysis looks impressive in pictures but its really noticeable when you start moving and all of the objects start loading the Higher polygon model.

 

[draconian]

2008 article of unlimited detail

http://www.tkarena.com/Articles/tab...icleId/38/Death-of-the-GPU-as-we-Know-It.aspx

 

[idsn6]

http://forum.beyond3d.com/showthread.php?t=47405

 

[London Geezer]

Unlimited Detail is a software algorithm that gives unlimited geometry. When we say “unlimited geometry” we really do mean it. It really is Unlimited, Infinite, endless power, for 3D graphics.

I stopped reading there.

 

[Shifty Geezer]

That's not really that difficult, as long as you don't sweat adding 3D detail and just keep subdividing...

 

[Rolf N]

That's not really that difficult, as long as you don't sweat adding 3D detail and just keep subdividing...

They could pick a fractal and displace the mesh at whatever resolution desired. Infinite detail indeed. Still not worth paying attention to though.

 

[Simon F]

Is this a technology that shows promise for this and next-gen consoles? It seems too good to be true, so what’s the catch.
http://www.youtube.com/watch?v=Q-ATtrImCx4&feature=player_embedded

Quote: @4:58 "We sent it off to a certain company..."

I would imagine that whoever was at that certain company would have been more impressed if the demos did not just consist of a fly through of a static scene consisting of replicated models. It gives the impression that there is a fundamental limitation to the technology <shrug>

 

[Simon F]

FWIW, Bruce Dell , who I believe is presenting that video, was a member on B3D.

 

[eloic]

FWIW, Bruce Dell , who I believe is presenting that video, was a member on B3D.

Indeed... but he registered because of this very subject (I suppose that somebody invited him) and he's not an active user (he only posted 3 times):
http://forum.beyond3d.com/showpost.php?p=1142124&postcount=16

That thread was interesting, but it died...

Let the 'voxel hate' begin!

 

[Mobius1aic]

I'd hate to think what the computation requirements for manipulating this data would be in real time, where you have moving objects as well as needing data to move across space. That's alot of little points to take into account as they are moving about, along with all the other crap the CPU has to orchestrate. I can only imagine needing a super computer to run let's say an ocean simulation. Rendering and output to a monitor makes sense but everything else really does not.

 

[Dungeonscaper]

Or we can just move towards tesselation more.

 

[Laa-Yosh]

Yeah, you have to wonder why movie VFX has not transitioned completely to voxels if they're the better solution for unlimited scene complexity.

Voxels are cool for procedurally generated smoke, fire and other fluid-related effects, and for medical imaging. But I'm not sure about game worlds, particularly if you want dynamics in them. Voxel trees won't move in the wind, voxel level geometry with the lighting baked into it won't be destructible, and so on.
We'll see if Carmack decides to push for it, though. He's still a force enough to push the industry into this direction if there can be obvious advantages in practice.

 

[DeadlyNinja]

I was just about to post this.

I heard that this technology also has problems with animation and physics.

Or we can just move towards tesselation more.

I've been meaning to ask this. If we use tesselation with displacement mapping, does that mean we'll finally have bumps that don't flatten at grazing angles? This is the biggest flaw with all forms of bump mapping today, but I've never seen displacement mapping running on my PC.

 

[Acert93]

Yeah, you have to wonder why movie VFX has not transitioned completely to voxels if they're the better solution for unlimited scene complexity

Devils advocate: Established tools, filters, and plug ins and an army of people already trained in said tools? And how much of an advantage is unlimited geometry in offline rendering? Diminishing returns and the cost of rendering relating to lighting and post processing being pretty huge.

 

[Alucardx23]

Yeah, you have to wonder why movie VFX has not transitioned completely to voxels if they're the better solution for unlimited scene complexity.

Voxels are cool for procedurally generated smoke, fire and other fluid-related effects, and for medical imaging. But I'm not sure about game worlds, particularly if you want dynamics in them. Voxel trees won't move in the wind, voxel level geometry with the lighting baked into it won't be destructible, and so on.
We'll see if Carmack decides to push for it, though. He's still a force enough to push the industry into this direction if there can be obvious advantages in practice.

A game like Final Fantasy XIII would be a good candidate then; almost all of the environments are static and non-interactive. One question, is it possible to combine a voxel based engine with a polygon one? I mean use voxel for the static parts of the world and polygons for the dynamic parts, again using Final XIII as and example where most of the backgrounds you can only see but not interact with.

 

[jlippo]

A game like Final Fantasy XIII would be a good candidate then; almost all of the environments are static and non-interactive. One question, is it possible to combine a voxel based engine with a polygon one? I mean use voxel for the static parts of the world and polygons for the dynamic parts, again using Final XIII as and example where most of the backgrounds you can only see but not interact with.

You combine them with Z-buffer and if you use deferred rendering lighting would be unified.
Voxels can be used for collision and thus could work nicely with particles, perhaps for things like flowing water on pebble road?

 

[MfA]

Yeah, you have to wonder why movie VFX has not transitioned completely to voxels if they're the better solution for unlimited scene complexity.

Generally they already use either raytracing or tesselation, so they are less input sensitive to start with. For the scan conversion renderers Influence of overdraw on framerate isn't a problem either, in most scenes it won't be big enough to really worry about and if a couple of frames are slow to render it doesn't matter since it's not interactive.

 

[PixFarmer]

I've been meaning to ask this. If we use tesselation with displacement mapping, does that mean we'll finally have bumps that don't flatten at grazing angles? This is the biggest flaw with all forms of bump mapping today, but I've never seen displacement mapping running on my PC.

Yes - tessellation overcomes the grazing angle problem.
* Parallax displacement
* Tessellation displacement