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

Originally posted by Uncinus

Originally posted by Glyph_D
It is claimed that "64 atoms per cubic millimeter (four per millimeter)". This the maximum this technique can zoom in on an object. Surely this is overkill when considering game design. The level of detail would be cut at a virtual distance of an average of 6-12 inches(as per the status quo of current gen games)

How would it render hair, which is 0.07 mm thick?

Here's an in-game weapon from Rage, are you sure you could model this at .25mm?

[atsimg]http://files.abovetopsecret.com/images/member/7be5193e463d.jpg[/atsimg]

Remember, ever little speck on scratch and there has to be made from voxels.

edit on 3-8-2011 by Uncinus because: (no reason given)

Well I guess you could.process the model as normal and then introduce this into your engine.
Example your model is 10x10x10 units You process this in high detail but the instance in the world occupies 1x1x1 unit?
So essentially your space is split into subspaces where the definition of unit is not consistent across them...

EDIT
Essentially you would change the point density of the processing tool..

Limbo



EDIT :-

www.euclideon.com...
As quoted by Notch/Eucludeon 64 atoms per cubic milimeter is the cap. So they do have a resolution cap and my theory about changing the sampling rate is wrong.

Originally posted by Glyph_D
reply to post by Glyph_D

 

He's just describing how voxel rendering with ray casting using sparse voxel octrees works. It's nothing new. He did not invent it. There's loads of other examples of the exact same thing.

>>>
i agree it is nothing new. The only new thing i think he has accomplished is optimization of that already present tech.
>>>

its all ABOUT optimization! No one denied its sort of voxel rendering...thats not the point.
The point is they CLAIM to have found the "magic algorithm" which makes voxel/point rendering REASONABLE on normal PCs. That's all what this is about.

Except...that their "explanation" and technical explanation is "wishy washy" and fuzzy...and since about three years that SINGLE person there is only giving such retarded "google search engine" comparisons instead of a REAL technical description of their engine and how it works.

There IS something fishy about this, i am afraid. You see, zillions of other graphics developers are not idiots either.

Originally posted by flexy123


its all ABOUT optimization! No one denied its sort of voxel rendering...thats not the point.
The point is they CLAIM to have found the "magic algorithm" which makes voxel/point rendering REASONABLE on normal PCs. That's all what this is about.

The difference with this new version and the already present version is its render "count" i believe.

Current voxel rendering renders geometrically every point in real time regardless of resolution. the optimized version only renders what the resolution can handle.

Originally posted by Glyph_D

Originally posted by flexy123


its all ABOUT optimization! No one denied its sort of voxel rendering...thats not the point.
The point is they CLAIM to have found the "magic algorithm" which makes voxel/point rendering REASONABLE on normal PCs. That's all what this is about.

The difference with this new version and the already present version is its render "count" i believe.

Current voxel rendering renders geometrically every point in real time regardless of resolution. the optimized version only renders what the resolution can handle.

Strictly spoken, what he said in VERY simplifying words, its not the "resolution" but that the engine only renders the actual visible pixels. So..in other words..instead of a zillions of points it would render MAX. as many
points as the monitor's resolution. (This concept, of course is also not new BY A LONG SHOT)

(Using some "magical" alogorithm which can detect the pixels to render from a zillion of pixels in a 3D scene)


But this still does not explain a LOT....memory management....animations etc..what happens if a scene is DYNAMIC?

To give a good example..because he used the example of how Google indexed the web and can quickly find results....but now imagine that the web would significantly change EACH second.

How would that graphics engine "index" a dynamic scene with constantly moving objects/observer etc.

This is incredible.....wonder what we will have 10 years down the road? I'm thinking full virtual 3D worlds for people in interact in. I would absolutely love to be able to fly and feel like I am through a real world that I live in!!!! A simulator of our whole world. I would probably never go outside, grow big and fat, and die a happy virtual life

Originally posted by KJV1611
I would probably never go outside, grow big and fat, and die a happy virtual life

It's called "World of Warcraft"

Originally posted by Limbo

Originally posted by Uncinus

Originally posted by Glyph_D
It is claimed that "64 atoms per cubic millimeter (four per millimeter)". This the maximum this technique can zoom in on an object. Surely this is overkill when considering game design. The level of detail would be cut at a virtual distance of an average of 6-12 inches(as per the status quo of current gen games)

How would it render hair, which is 0.07 mm thick?

Here's an in-game weapon from Rage, are you sure you could model this at .25mm?

[atsimg]http://files.abovetopsecret.com/images/member/7be5193e463d.jpg[/atsimg]

Remember, ever little speck on scratch and there has to be made from voxels.

edit on 3-8-2011 by Uncinus because: (no reason given)

Well I guess you could.process the model as normal and then introduce this into your engine.
Example your model is 10x10x10 units You process this in high detail but the instance in the world occupies 1x1x1 unit?
So essentially your space is split into subspaces where the definition of unit is not consistent across them...

EDIT
Essentially you would change the point density of the processing tool..

Limbo

edit on 3-8-2011 by Limbo because: (no reason given)

It doesn't matter what scale you use, you'd still require the same amount of voxel data to model the surface detail desired. And for what? A comparatively lower resolution normal mapped mesh like the one above achieves the same result at a fraction of the cost with none of the drawbacks.

Originally posted by flexy123


(Using some "magical" alogorithm which can detect the pixels to render from a zillion of pixels in a 3D scene)


But this still does not explain a LOT....memory management....animations etc..what happens if a scene is DYNAMIC?

IMHO i proposed earlier the idea that the algorithm was connected directly to the resolution, where as it would by reference of the resolution know exactly how many pixels can be ignored at a given frame. The only pixels remaining would be the most central pixels in any given cluster.


I agree entirely with your second statement.

Originally posted by Glyph_D
Current voxel rendering renders geometrically every point in real time regardless of resolution. the optimized version only renders what the resolution can handle.

No, all voxel renderers above high school level will render based on the resolution. Either they use ray casting, or they use "splatting" which traverses the entire tree down to the resolution needed. Nobody is going to render every point.

He's using ray casting, which is a decades old technique.

Again, everything he's saying simply describes existing technology. He's just spinning it.

Sorry, but this is not new, i barely remember reading this, it might have been in new scientist a few years ago i dont rememeber. I've done alot in 3D terrain rendering in the past(long time obsession and UNI final year project) and it looked dead great from my point of view, but I looked into this and some of the other video demos, there were issues with graphics glitches, reflections etc etc, these were serious issues that you can expect a modern day graphics card to do (I know that technically they cant do all the 'detail' this is claiming but using this rendering method meant you LOST effects that in my book more than made up and exceeded what extra benifits this offered), it had a long way to go.
It is promising maybe, but that video is the same video I saw at least a year ago, prob more and there has been no other progress as far as i've noticed.

Originally posted by Glyph_D
reply to post by Uncinus

 

"We made a search algorithm, but it's a search algorithm that that finds points, so it can quickly grab just one atom for every point on the screen." According to Bruce Dell, it's all about efficiency. "So think about the difference," he says. "If you had all of the points you are seeing on the screen, like in our demo, it's going to take forever. You'll be waiting for a long time. But if you're grabbing only one for every pixel on the screen, then you don't have a trillion dots, you have… well, pick a resolution and do the maths!

"That's the difference. In layman's terms that's how we're doing what we're doing. The workload is so small that at the moment we're running software just fine with real time demonstrations and we're still optimising, because we keep finding more efficient ways to do this."

link

heres a convo with the front man of the development.

as above he states that the work load is relative to the resolution.

Too bad Notch decided to run a smear campaign on these guys before actually knowing who they are or what they do. The Australian government is funding this thing and the big players in the industry are already talking to them.

This is not a hoax, if it was a hoax he would be screwed in the end. The Australian government would lock him up and his investors would sue his ass.

sooo what does this mean for Diablo III????

if nothing then i dont really care atm (:

Originally posted by BIGPoJo

Originally posted by Glyph_D
reply to post by Uncinus

 

"We made a search algorithm, but it's a search algorithm that that finds points, so it can quickly grab just one atom for every point on the screen." According to Bruce Dell, it's all about efficiency. "So think about the difference," he says. "If you had all of the points you are seeing on the screen, like in our demo, it's going to take forever. You'll be waiting for a long time. But if you're grabbing only one for every pixel on the screen, then you don't have a trillion dots, you have… well, pick a resolution and do the maths!

"That's the difference. In layman's terms that's how we're doing what we're doing. The workload is so small that at the moment we're running software just fine with real time demonstrations and we're still optimising, because we keep finding more efficient ways to do this."

link

heres a convo with the front man of the development.

as above he states that the work load is relative to the resolution.

Too bad Notch decided to run a smear campaign on these guys before actually knowing who they are or what they do. The Australian government is funding this thing and the big players in the industry are already talking to them.

This is not a hoax, if it was a hoax he would be screwed in the end. The Australian government would lock him up and his investors would sue his ass.

You can't say "If it were a hoax, he would be screwed. Therefore, it's not a hoax". Whether or not he would get in trouble doesn't validate or invalidate his claims. Him getting in trouble in the future doesn't prevent him from making claims in the past.

Besides, most of us are not saying it's a hoax. It is a solid concept, and has been actively researched for years by many people. But I do not believe their technology is to the point they state, it just doesn't make sense to me. The idea for the funding is to see if they can refine or advance the technology to the point were it will actually be useable. If they were already to the point, they wouldn't need funding in the first place.

This appeal to authority is an error in thinking. Just because the government partially funded it doesn't mean we should just say 'Ok, well it must be true then'. And please, they would lock him up? Thousands of projects funded by all sorts of investors and governments fizzle out every day, simply because they couldn't live up to their overhyped claims, or some other detail made impractical. I haven't seen one example of anyone getting locked up for that. Our government has funded projects ranging from developing a machine that could create it's own jokes, to a holodeck like in Star Trek, but those all went bust too.

You certainly could raycast from the camera to determine visible points and render them. This is currently done in polygon engines, to reduce rendering overhead. But whether or not you render these atoms, they have to be readily accessible in memory to run the algorithm to find them. The same model stored in atoms instead of polygons would take up so much more space in memory as to negate any potential benefits. It doesn't matter how fast you can find the voxels that will be displayed on screen if you can only store 5 different objects in memory at a time.

I completely agree that this could look good on a very small scale demo, on a higher end machine. But increase the size and complexity of the world and the amount of data increases exponentially. There's not enough RAM in my computer lab to store the amount of voxels that would need to be searched through in a large level.

Simply because polygons will always require less data than the same voxel model, and any 'search algorithms' or optimizations could apply to both rendering methods, polygons will always have the advantage of less data to search through and process. You can create a polygon model as detailed as possible, and it would still take up less space than storing it as 'atoms'. So until we get to the point with polygons where we simply can't go any further with them, where graphics card's capacity is going unused, voxels will be less efficient. We are not to that point yet, and polygon counts are still doubling every few years.

Notch says it's a Scam.



I believe Notch.



His reasoning behind it makes 100% sense too.



Sorry to dissapoint you OP - i was just as dissapointed as you're about to be once you're finished reading this


p.s for the people saying notch is running a smear campaing against these guy's? He's not.

He watched the video - He took the information they gave him - and he did some simple math. The main problem is - either they have 170 thousand HDD's strapped together on a supercomputer to show us that pretty video - Or they're lying about the rendering/processing power of their engine.


Etiher one being true? Bad.


notch.tumblr.com...

This is truly amazing, if true.

And if not, like the skeptics think, then that sure was a LOT of work for a scam.

I personally think this is true, and that it will not only revolutionize gaming but also film.

Furthermore, we're all made of atoms and some of us tend to believe we could be a simulation of some kind as well. This is a good step towards proof of such theories, because if this is real then imagine how THIS technology will advance over a hundred or a thousand years! THEN, we'd be making all sorts of simulations and such to learn and to teach what happened in the past, or what might happen in the future.

Does that then mean that we are a simulation? Could be...

Maybe we are becoming self aware of the "matrix" that's all around us.

reply to post by TigaHawk


If you believe whatever you hear from anyone without looking any deeper, or trying to find the truth out for yourself, then you are going to have a rough life.

Notch has every reason to be jealous and upset; this is Minecraft killing, and Minecraft already has some of the dumpiest graphics ever for the very reason that this new technology claims to solve.

Now I don't know any more than you do whether this is really what they presented, however, I also don't blindly assert my belief in something just because someone else said so. Don't be a follower, lead yourself to your own satisfaction of the truth.

if you are unsure if this is real, you know that time will tell. Wait for more information to come out before you jump to conclusions, because technology does nothing BUT improve constantly, at a tremendous rate. The idea that the days of polygons are behind us is very real, just as the days a 8-bit and 16-bit systems can to an end. People said the same thing then, and stubbornly clung to those bits. They bought the 32X enhancement device for Sega Genesis to cling a little longer. Then they finally came around and realized, the bit is dead. Now it appears the polygon is dead as well.

"I’ve been getting a bunch of feedback that my last blog post is wrong for various reasons, and I’d just like to say that I would absolutely LOVE to be proven wrong. Being wrong is awesome, that’s how you learn." - Notch

Time will tell.

Notch was only wrong in the calculation of the amount of memory it would take, because he assumed it was stored the same way as in Minecraft, with every possible voxel taking up memory.

In reality, high resolution voxel environments are compressed with a sparse voxel octree, and the mostly empty space takes up very little memory. Which this guy did not invent, and it's not new. Nothing appears to be new.

In every other respect Notch is correct. There's nothing impressive here, it's just a voxel renderer, with repeated graphics so it looks like it's rendering a lot.

The polygon *sadly* is not dead, UNLESS that person proves it being so.

A tech demo proves nothing, do you know how many amazing tech demos i already had...with each new video gard comes some amazing tech demo.

The point here whether this technology is even remotely feasable to replace current 3d rendering tech (which is based on polygons)..using TODAY's computers. This is the basic claim that MAN makes, he claims he has an algorithm which would blow everything out of the water - but failed to show evidence for already 3 years.

Give me 3 nerds and coders, they make some tech demo similar to theirs..woopdeedoo. The PROBLEM is that the claims he makes do not make sense for many people who KNOW stuff about 3d programming...not even in theory.

So..let's say i scan in one single stone or tree with those details (which allegedly has almost infinite detail)...with 4x4x4 "atoms" per mm....now FOR A START...i would like to know how the "almost infinite detail" is STORED.

Because you cannot "pull infinite detail" out of thin air (except if we assume that this so called infinite detail is nothing else than the same, repeated patterns)..but then its not (almost) infinite detail.

There is simply no way to scan (or design, for that matter) a tree or whatever in such detail that it would be a 1:1 copy of nature...we are talking TERABYTE of data for one single object.

Originally posted by Turq1
So instead of polygons...things are made of little atoms, or spheres...which are polygons.

And how would this not take up massive computing power? If it was "unlimited power" then why did he say they ran it at 20fps? Doesn't seem to make a lot of sense.

Realistic graphics don't make a game good either. Art > graphics.

Originally posted by SaturnFX

Originally posted by countduckula24
I'm fairly techno-illiterate when it comes to the concept of graphics. However this sounds really cool, so thank you very much for sharing. I am reading the article now and will watch the video later, as I am at work at the moment. From what you've indicated it sounds like video games, and hopefully other applications, are about to get completely revolutionized.

More or less, yes.

If your a gamer, this is going to knock your socks off X 1000

Hmm...no. Tessalation, yes.

edit on 2-8-2011 by Turq1 because: (no reason given)

I was thinking about this too. The graphics cards we ALL use today render polygons. They did say that they
were using a software render that got 20 fps and that they are testing and building a hardware render-er. I guess it means we would all need to get new graphics cards. And I doubt that ATI or Nvidia are going to make something just to accommodate this...

PS: sounds to good to be true is my gut feeling. I'll believe it when they have a REAL TIME demo. Anyone can make a high res render and say anything they want to. I didnt exactly see a camera that looked realtime, did you ?

Another PS: How could they get a software render to do global illumination like that... all the math involved and then push it to the graphics card to show on screen ?

This thread is a great example of "know it alls" who don't really know it all..and the one's who flame me are probably guilty , but the truth is we don't truly know if this is or isn't possible, because if they have manged to of found a way to utilize limitless atoms well we wouldn't begin to understand how it works.

HMMMM I wonder if the reality that I am experiencing is a rendered simulation @ the quantum level........ HMMMM