Unlimited Detail Real-Time Rendering Technology

[Guest Retarded Llama]

Tweeted by a Rockstar North developer:

I hope this is possible.

Embedded?

[DuPz0r 5362]

Tweeted by a Rockstar North developer:

I hope this is possible.

Embedded?

Yeah this is possible to an extent. I've read into this in the past. I watched this recently too. How they're implementing it i don't know. But good news for me is that they made a plug-in to use it with 3DS Max, so i can export all my high poly models as points of detail .

I'll be watching this as it could be the future of gaming.

On the other hand, this could make games far too expensive to make. The more detail means more work involved to bring it all together... Keep sceptical about it until they have a better tech demo.

ok i just found an interesting view on this tech:

They made a voxel renderer, probably based on sparse voxel octrees. That’s cool and all, but.. To quote the video, the island in the video is one km^2. Let’s assume a modest island height of just eight meters, and we end up with 0.008 km^3. At 64 atoms per cubic millimeter (four per millimeter), that is a total of 512 000 000 000 000 000 atoms. If each voxel is made up of one byte of data, that is a total of 512 petabytes of information, or about 170 000 three-terrabyte harddrives full of information. In reality, you will need way more than just one byte of data per voxel to do colors and lighting, and the island is probably way taller than just eight meters, so that estimate is very optimistic.

So obviously, it’s not made up of that many unique voxels.

In the video, you can make up loads of repeated structured, all roughly the same size. Sparse voxel octrees work great for this, as you don’t need to have unique data in each leaf node, but can reference the same data repeatedly (at fixed intervals) with great speed and memory efficiency. This explains how they can have that much data, but it also shows one of the biggest weaknesses of their engine.

Another weakness is that voxels are horrible for doing animation, because there is no current fast algorithms for deforming a voxel cloud based on a skeletal mesh, and if you do keyframe animation, you end up with a LOT of data. It’s possible to rotate, scale and translate individual chunks of voxel data to do simple animation (imagine one chunk for the upper arm, one for the lower, one for the torso, and so on), but it’s not going to look as nice as polygon based animated characters do.

Source: http://notch.tumblr.com/post/8386977075/its-a-scam

[Guest Retarded Llama]

Hmm may be used in GTA 12 by the sounds of it.

[Psycho 2]

Sounds like a bunch of fake crap to me.

If you have a polygon converter hows does that help you are still basing it off the original polygon all you have done is just change it into points in a polygon shape, the only way to actually do this properly would be to make achieves of very detailed, high quality structures of every object, with multiple variation within them. for instance with sand you would have 10-1000 different sand crystals and then you would colour them individually and chuck them into your environment. If you can have infinite detail (which you can't) then no game will have better or worse graphics so there will be no need to improve graphics so once the item is made above the perceivable resolution of the eye then down graded to the resolution of the monitor you could keep the models for life, in the case of inanimate objects as well as cars, landmarks and pretty much anything that isn't original to a game. Another problem with this whole idea is that most objects interiors are never seen and therefore the interior design of the structure is rather irrelevant, of course this could lead to very dynamic destructible environment however you then still have to code how the object will break apart (wood splits, rock crumble, metal melts/bends and twists,which would also take time, so plenty of jobs in the games industry for material scientists then, and mathematicians trying to create the correct real world effects of different objects and their relative effects on the world once user interactions occur on them.

When you think about that it would get kind of boring, different games having different styles make them more interesting, however the real question is "Are lifelike graphics really perfect graphics?"

As for Unlimited detail, I have created protein structures using point cloud data from Xray Crystal structures and I will tell you now, a top of the range Imac won't run more than 3-4, 300 residue (~3000 atom) atomic structures smoothly at the same time and takes 5-10 seconds to render them in different styles with only one open.

So though entirely possible to do as a theory, no computer could run it, not to mention not needing too due to the relatively low resolution of screens. If you notice they never mention any type of equipment they are running it on or even how they are creating or rendering it.

But to be honest anyone who believed this was real was a bit of a fool anyway.

[DuPz0r 5362]

I don't think anyone would believe "unlimited detail". But voxels could be a good way to go for creating static environment objects with a lot of detail. At current, we use normal maps which give us the illusion of the high-poly detail which we bake from the high-poly models. The only problem with that is it doesn't look good from extreme angles as the illusion breaks. and the models look flat. Voxels however could give us a lot more detail, and could be good for important set pieces or props.

[DuffMan 266]

As long as they only have it running in software we're no closer to getting it into games than we were when they last demoed it.

I thought they previewed it well before 2010...