NERDGASM ALERT: Detailed Rendering of CG just got infinately better. The polygon is dead

i can imagine how something like this will work for the new planetary engine tech. for example, this guy ]

took this guy's idea (fun planetary engine demo)
www.xwaytoonsoft.com...

and expanded on it.

this guy went even farther


and all of it would be incredible in unliimited detail. you could take a 3cm detail height/land map of the earth, and have insane amounts of detail everywhere. the whole planet would be just one part of the game, since you could also take the height/land maps of mars and the moon and do the same thing.

blink blink. ....blink. sniffle. anybody got a hanky?

Originally posted by Pumper
It will not kill polygons anytime soon. The thing is, that it will take then at least 2 extra years to reach the current visual level (i.e. Rage, Crysis 2 dx11, BF3) - as of now they can only offer unlimited geometry. By the time they sort out all the other major issues like no animation and no physics dx12 or even dx13 will be out with superb interactivity (destructible environments, dynamic tessellation, real time ray tracing and other stuff) and their 'unlimited detail' technology will look like # especially considering that in that time the polygon counts in games will increase at the very least (in about 3 years that will take to finish they software) 2x2x2=8 times (they say it doubles every year) and their unlimited detail will not look any better that the good old polygons.

Nothing to be excited about.

Lol?

No amination? How do they do it now then - there is a structural skeleton that represents how things move, but painting it with pretty polygons is the hard part. I dont see how painting it with this technology instead would be impossible. Remove the polygons and you have the basics of 3d animation, a mesh which requires very little computational power. A frame.

I also don't see polygons exponentially getting better. Not at all, I see it culminating to a set point and then stagnating. Or hardware tehcnology is required t6o go hand in hand, but that means suddenly no one can afford it, negating itself by it's own design.

By saying this however, I don't automatically claim this to be as shiney as it proclaims. But I want to see more of it, that is for sure.

Originally posted by WielderOfTheSwordOfTruth
First thing I thought when I saw this technology... "@#%$@ everything I learned in 3D school is useless!! Damn."

..Do you even know what you're talking about? Anyway, learning 3d isn't something that will ever become obsolete, and even if conventional methods become that way someday, this tech won't be the one to do it. This engine simply converts the detailed models you create via software into its atom-based structure. Amateur/unskilled 3d artist will be unaffected by this, and likewise for those that are skilled/talented; except this will actually help by eliminating polygon budgets. If you knew anything about 3d from your 'school' then you'd realize the the implications of this tech don't make the artist' techniques useless...

Well they seemed to disappear off the web. At least their website.

www.euclideon.com...

Originally posted by Grimbone
Well they seemed to disappear off the web. At least their website.

www.euclideon.com...

There is something odd about this.

First, yes at first glance my JAW DROPPED, but then this is really basically like voxels, just one step further.

The video did in no way explain what the "secret" actually is to get this amount of data and rendering detail "into a normal computer", to use simple speak. Where are the billions of "atoms" stored, what is the break-through technology about this engine which supposedly should make this kind of rendering techniques possible for today's computers??

I mean..with "infinite" powerful hardware it can be done..sure, but there is a reason WHY today polygons are used...and this is limited CPU/GPU power...3D models need to be simplified. Sadly, that demo leaves us totally in the dark when it comes to the technical aspect.

Their site is down also, which is a little odd.

reply to post by flexy123


The account seems suspended. The full link it brings you to suspendedpage.cgi.

www.euclideon.com...

S+F

This looks really cool but I think to be rendered to it's full potential we will have to wait for the next generation of consoles and an upgrade/cost decrease in PC processors and memory. Still that 20 fps video looked amazing. Can't wait.

He he I was like:

But as they said we will not see the technology in use for some time, might be beaten by other solutions.

It is however really awesome and I hope to see something usefull from them soon.

Also the video was very cool.

Sorry guys, but this is just an over-hyped tech demo.

It's not that this approach isn't possible, it's that is not practical. The reason polygons are used is not because nobody has thought of a better way, but to get around hardware limitations. A computer only has a finite amount of processing power, and rendering eats most of that up. By limiting the object to polygons, you cut the processing cycles way down, because instead of having to calculate the lighting and position for every point on the object, you only have to calculate the points of the vertices, and then extrapolate the information to fill the polygon.

Essentially this is the same as rendering polygons, except each point on the object is a 'polygon'. Even with extremely conservative estimates, this would increase the CPU cycles needed by 1000s of times. So instead of calculating lighting for 1000 faces of a traditional model, there would be a near infinite number of 'atoms' to calculate.

At best, what I see this doing is you 'model' using atoms, and then the model is down scaled and rendered traditionally with a limited number of faces. Then as you move close to the model, it would use a larger number of faces. But this is no different than what is currently being done with 3D meshes, where they will display fewer number of polygons depending on how far away you are.

I have worked on the open source Cocos3D project, that uses OpenGL ES to render 3D objects on iOS, and rendering even a simple cube is a very complex process. Sure, we could build the cube out of atoms, and have hundreds of thousands of them make up this one object, or we could build it with 8 vertices and 12 polygons. Which do you think is going to be more efficient?

We are currently limited not by software, but by hardware. There is simply no way you can have a 1000 fold increase in detail without the same increase in processing power. Everything still has to be rendered and computed, there is no way to avoid it.

In short, it's a novel idea, but don't expect anything like this unless its accompanied by new technology for graphics cards. Whether you want to call it a 'polygon' or an 'atom' doesn't really matter, each one has to be computed individually. Anyway you look at it, this 'tech' is not going to be as efficient with current hardware as polygons.

reply to post by SaturnFX


And we will become Gods. I always thought this was possible, hehe, looks like us "Reality is a very intricate computer program" people are starting to look more and more correct about what reality really is. Basically these innovations are going to blur the definition of reality. Holy Cow!!!

reply to post by SaturnFX

a graphic environment could be built from an infinite number of 3-D virtual atoms, much like the physical world. It was a cool idea. Then Dell and his Unlimited Detail graphics system disappeared. ___ Dell describes in perfect exhilarated-Aussie just how awesome this technology could make our video game worlds and other virtual environments.

Unlimited Detail can now pack one million atoms into a single virtual cubic inch, allowing for unprecedented detail. And it could make such environments less virtual, allowing game designers to “scan” in objects from the real world and present them as they look naturally, making video game worlds a kind of hybrid reality with some parts real and some parts engineered by artists.

WOW 3D VR Atoms instead of Polygons

You may know what im thinking Right !

take those Atoms and use this ! Bunch The Tesla Processors or a huge Super Computer = Use Micro Nanotechnology with DNA GATTACA Sequence Mapping, with it put in the Hands of a Young Mad Scientist ! some way to make those Atoms turned into Matter ( if there was a way )

Would that Mean
Then One of my Favorite Movies will become a Reality !



To bad there wasn't anything like this !

Well Gaming will be in the Real Virtual Reality World!
If Someone that has Creation Ideas like a Fictional character named Kevin Flynn !

I have a Bad Feeling about this in the distant future ! if it Could Happen

when you think about using electronic VR Atoms in gaming
and using a laser beam or Something close to disintegration to Electronic 3d VR Atoms(1s & 0s) and reintegration to matter.

new meaning of digitized for that to happen you need a Classified super computer LOL!

Here Give me your finger (Prick) Im now going to make Digitize you through this computer over here !
Computer sees the DNA coeds and Preps to Map it and Creates the Digital You it !

The Horror of that ! just had to say LOL!

Virtuosity Trailer (1995)


TRON: LEGACY Official Trailer






Supercomputing at 1/10th the Cost. The Tesla Processor , page 1
www.abovetopsecret.com...


Thanks Saturn FX for showing ATS

i thought the graphics now are good but those are just amazing. its amazing what they are coming up with now, just awesome

I don't know if this is real or not.
Its a hell of a lot for someone to go through just to say "sike!"
I look forward to seeing them try and prove their claims.

Here's the other side of the story:

Perhaps you’ve seen the videos about some groundbreaking “unlimited detail” rendering technology? If not, check it out here, then get back to this post: www.youtube.com...

Well, it is a scam.

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.

SOURCE

All I want to know is, how much I should start saving up for a machine that will be be able to handle this?

Originally posted by klenker

Originally posted by Pumper
It will not kill polygons anytime soon. The thing is, that it will take then at least 2 extra years to reach the current visual level (i.e. Rage, Crysis 2 dx11, BF3) - as of now they can only offer unlimited geometry. By the time they sort out all the other major issues like no animation and no physics dx12 or even dx13 will be out with superb interactivity (destructible environments, dynamic tessellation, real time ray tracing and other stuff) and their 'unlimited detail' technology will look like # especially considering that in that time the polygon counts in games will increase at the very least (in about 3 years that will take to finish they software) 2x2x2=8 times (they say it doubles every year) and their unlimited detail will not look any better that the good old polygons.

Nothing to be excited about.

Lol?

No amination? How do they do it now then - there is a structural skeleton that represents how things move, but painting it with pretty polygons is the hard part. I dont see how painting it with this technology instead would be impossible. Remove the polygons and you have the basics of 3d animation, a mesh which requires very little computational power. A frame.

I also don't see polygons exponentially getting better. Not at all, I see it culminating to a set point and then stagnating. Or hardware tehcnology is required t6o go hand in hand, but that means suddenly no one can afford it, negating itself by it's own design.

By saying this however, I don't automatically claim this to be as shiney as it proclaims. But I want to see more of it, that is for sure.

Let's take a look at the Forza Motorsports line of games, as an example. Listed is the average number of polygons per car:
Forza 1, released in 2005: ~35k
Forza 2, released in 2007: ~100,000
Forza 3, released in 2009: ~400,000
Forza 4, releasing in 2011: over 1,000,000

See how each year it doubles or triples? Yeah, that's exponential. Forza 2-4 are also all on the same system, so that kind of invalidates your whole premise. Also consider the original Doom, released just over 10 years before The original Forza, had maybe 50 polygons in the whole scene. Now games an have upwards of 100's of millions.

This is just one recent example, but there are dozens of others. Now keep in mind an exponential increase in polygon count doesn't lead to an exponential increase in appearance. Each polygon up to a certain point makes a huge different, like during the PSX/N64 era. But after that, it takes more and more polygons to have a noticeable effect. Things like texturing and lighting play a very large part as well.

Also, I would avoid speaking about matters you have such little experience with, your attempt to grossly simplify things shows your ignorance. Just 'paint' the 'frame' with this new technology? How does that make any sense in your mind? Have you ever worked with a 3D object, not to mention skeletal animation? Traditional 3D objects are hollow, basically a 2D mesh wrapped in 3D space. It is very easy to bend the mesh at its vertices in relation to a skeleton, just like folding a piece of paper along its crease. This is just not possible if the mesh is no longer hollow, and doesn't have vertices to bend at. It would require a completely different set of physics to bend the solid object and visualize it. That alone would make the methods completely incompatible.

Originally posted by klenker


No amination? How do they do it now then - there is a structural skeleton that represents how things move, but painting it with pretty polygons is the hard part. I dont see how painting it with this technology instead would be impossible. Remove the polygons and you have the basics of 3d animation, a mesh which requires very little computational power. A frame.

. . . .

--1. Voxel engines do not animate. For a reasonable voxel engine, trying to follow any skeleton is between eight and twenty terabytes of data per second. This means they're limited to moving large blocks: People move like mannequins or dolls.

2. Voxel engines need to work in high precision modes constantly. A Radeon HD 6850 can get you around 1.4-1.8 Tflops at 32 bit precision. That's around 200 Gflops at double precision (64 bit). It cannot do quads. To avoid popping and stairstepping, a voxel engine needs to work at 96 bit or better precision constantly. It's directly analogous to IMR's z-buffer, and we all know the terrors of a too low z buffer precision. Such high precision modes batter memory bandwidth and throw away every last advantage that GPUs have, their explosive single precision performance and their enormous memory bandwidth.

3. The "sparse" in "sparse voxels"? Means you're not actually using enough of them. Helps out performance, #s all over quality. Change "sparse" to "adequate" and you're back at terabytes per second again. No way around it. You want all those points? You need data to describe them. You need three 96 bit precisions. 32 bits of colour. 40 bytes per voxel, bare minimum, low detail. For a decent scene, you need billions of them. Oops, back at terabytes.--

From the link I posted earlier

Originally posted by Odessy
I don't know if this is real or not.
Its a hell of a lot for someone to go through just to say "sike!"
I look forward to seeing them try and prove their claims.

Here's the other side of the story:

Perhaps you’ve seen the videos about some groundbreaking “unlimited detail” rendering technology? If not, check it out here, then get back to this post: www.youtube.com...

Well, it is a scam.

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.

SOURCE

That is absolutely true, that's what I was trying to get across in my post further up the page. Hardware is already at its limit with polygon calculations, and all further subdividing the object into 'atoms' (aka voxels) is going to do is increase the calculations and memory required exponentially.

So this may be 'real' in the sense that it is a working demo, but not 'real' in the sense we will be seeing anytime soon. That 20FPS could be on a $10k machine, or a distributed CPU set-up for all we know. Though I question the fact that it's even doing what they claim.

For everyone saying how great it looks, you are ill-informed if you think you'd be able to tell the difference between his magical rendering system, and a polygon rendering engine at that resolution and bitrate. Actually, if you don't have to worry about spending CPU cycles on game logic or physics, you could make a small island scene that with millions of polygons where it would be near impossible to make out the individual faces, and it would run on most mid-high end computers.

It reminds me of all the videos I see demonstrating a new 1080p game, and everyone is like 'Oh that looks awesome! HD looks so much better!', even though they are only watching a 240p YouTube video.

reply to post by SaturnFX


Oh man that changes everything, Unlimited workspace, and you can use both fiction and non-fiction animations?
Seriously Minecraft + this type of rendering =

I've already posted this to my facebook, many many thanks for bringing it to the attention of the interested.

reply to post by Akasirus


I think most people can grasp the physical limitations of rendering a crap ton of voxels. What if the developers found a way to only track and render the voxels that you are looking at? Also, at a great distance away you would need less voxels, just scale them up or down based on your distance. Maybe they found a shortcut to reduce the hardware requirements. Maybe they only track a limited number of voxels and predict the rest based on previously rendered voxels?

I personally have no experience in 3d rendering and voxels are new to me.