The "Unlimited Detail" Guys Are Back

I honestly was waiting for a "college humor" image at the end...seems too good to be true.
 
Hmm.. I could see this working if it's as it's described on the site.

The Unlimited Detail engine works out which direction the camera is facing and then searches the data to find only the points it needs to put on the screen it doesn’t touch any unneeded points, all it wants is 1024*768 (if that is our resolution) points, one for each pixel of the screen
Click to expand...

It could store all the rest as background crap or map data and rough forms but only the view of the camera would be converted and you'd only need to render whatever the resolution is. By the sound of it, it's taking the aspects of cynicism and applying it to what a game really is: a 2D world. If that's true, even collisions/explosions should be incredibly easy as it's only going to process and show whatever you're looking at. It won't bother with stuff you're not going to see, it'll make it vague and crappy just like mario brothers on the NES but that doesn't mean it isn't happening or that, if you look at it, it won't render it into point data. Maybe I'm drawing conclusions but it sounds simple and it could work. It's limits and memory usage are really only defined by your pixel count on your monitor. I imagine if you use, say, the screen hovering over the Dallas Cowboys midfield, you might run into problems, but on a comp it shouldn't be hard. It's just perfect cynicism, if you aren't looking directly at it.. it's not there.
 
Wouldn't they have to kill a tiger and a swan to make their picture? Good example fellas...
 
Yossarian22 said:
Could you describe what it is that you did without breaking any NDA you might have signed upon employment? This is the first time of me hearing of such a technology.QUOTE]


It's just as I said. It was a point cloud for an oil compressor you find in a car. The company already had it all assembled with all of the different components, and they needed us to match that particular production run with a new production one we were doing - common parts that they could use between the two of them to reduce costs.

We had to overlay the point clouds with the current Compressor ASMs we had and had to match the lugs and such.
Click to expand...
 
Yossarian22 said:
Could you describe what it is that you did without breaking any NDA you might have signed upon employment? This is the first time of me hearing of such a technology.
It's just as I said. It was a point cloud for an oil compressor you find in a car. The company already had it all assembled with all of the different components, and they needed us to match that particular production run with a new production one we were doing - common parts that they could use between the two of them to reduce costs.

We had to overlay the point clouds with the current Compressor ASMs we had and had to match the lugs and such.

Damn lack of an edit button. Let's see if this looks better.
Click to expand...
Click to expand...
 
I understand what you're saying, but it now brings up questions. If the image is being rendered by pixel and not by mapped polygons, does that mean every time you change the fov, it will seek from the hard drive to figure out what's next? Or will that blue print be stored in the memory awaiting the processor to organize everything so that the gpu can render the image? I have so many questions for this and I honestly don't think this technology will take in gaming for at least 5-10 years IF THAT.
 
Yossarian22 said:
Nah, it was like watching Zero Punctuation, really.



Could you describe what it is that you did without breaking any NDA you might have signed upon employment? This is the first time of me hearing of such a technology.

You could dynamically scale the level of details by decreasing point cloud count on specific types of objects. Within the engine certain objects inherit from base class environment, character, background and sliders in game. Models being defined with point clouds with specific levels of priority/importance (base elements that make up the shape of a rock vs. extra point clouds that give the rock more detail if you will)

Bam. Done. But it might look like shit as someone pointed out with the low count voxel tech.

I don't really care, it's just time we have moved forward.
Click to expand...


Can't speak for what Solidstate is working with, but I work for a civil engineering consultant and we've been working with point clouds of highway schemes (generated from either truck or helicopter mounted laser scanners) in 3DS Max for a while now (we mix existing with 3d models of proposed roads and bridges to produce visulisatons for public enquiry etc). You can pick up a survey grade scanner for 'only' £20k, (or perhaps less, the university I attended snagged an ex-demo Topcon GLS-1000 for £9k)

As mentioned these point clouds (that visualy seem to be much lower definition than the OP) require a hell of a lot computing power. They (UD) handwave away the repetitive nature of their video saying they are not artists, but the chaotic nature of, well... nature means that 'realistic' scenes, without uniform repitition will be significantly harder than their 'unlimited' simulation to accurately reproduce.

That's not to say it can't work, a good artist could probably still compose scenes in such a way to make it less noticable.

It's inevitable that things will get better as time moves on, I'm skeptical that this is all it's cracked up to be, but hopefull none the less.
 
We are getting an interview with this company. Post your questions here that you want asked.
 
How is this different then voxels?

Is there "atoms" within the objects, so you can cut them up and see their innards?

Is all lighting done in realtime? How is it handled?

Will this increase or decrease the amount of time artists need to create graphics?

What are the gameplay advantages of this tech, or is for visual representation only?
 
Question for Interview: It is our understanding (from past interviews/QAs) that you (and your company) achieved most of your work in "isolation". While this certainly has it's benefit (freedom to think outside the box and not be confined by traditional thinking), have you since taken a look at A. other research into this style of rendering (point clouds, voxels, animation etc) and B. the current state of the art in polygon rendering? All to see where your method stands and if there is anything to be learned/improved upon by this information?

(Basically if this is all done "in a vacuum" it's quite impressive and one would think that this original line of thought could be significantly improved by studying what other people have done).

More standard Qs: what is the time/space memory tradeoff for your way of storing the data? Is it fast/slow to access, large/small to store, does it take a long time to pre-compute the datastructure so that realtime changes/movement (eg. animation) becomes prohibitive?

Give us some disadvantages to the technique and your thoughts/ideas on how you have, or plan to overcome/address them.
 
This guy has been pitchin the same shit since 2009, apparently the method is all done by the CPU, without hardware acceleration:

http://www.somedude.net/gamemonkey/forum/viewtopic.php?f=12&t=419


He disappears and comes back and seems to get alot of attention every time. In one of his videos he claims to have taken his tech to a developer and the techies wouldn't even let them go to the board of directors:

At 4:55

http://www.youtube.com/watch?v=Q-ATtrImCx4


It doesn't add up, and the way his tech is presented isn't very professional. 99% sure he's just looking to rip off investors with some well known but little used old tech that's ultimately impractical at a gaming level.
 
How much RAM this will use during gaming?

Can consoles take advantage of this tech as well as PC? (Very important, without console support it will not catch on)

Can this be used in animation?
 
PornoSatan said:
Yeah. Maybe I'm immature, but I f'in lol'd at the top comment, it matches the guys voice perfectly:

"WE NO LONGER MODEL DICKS BY HAND. NO MORE DICK POLYGONS.

WE SCAN THEM IN. EVERY SINGLE PENIS ATOM IS RENDERED.

TRUE DICK TECHNOLOGY."
Click to expand...

Am I the only one who laughed at this?
 
This is a game changer. Providing the hardware requirements to run these graphics isn't ridiculous, the SDK (which they speak of during the video), once available, will allow developers to move away from polygons and over to atoms.

When they compare "their" graphics to graphics seen in current games, the current games look terrible.

The difference appears to be light and day. At present though, I think the atom models are limited to static, non-moving objects.
 
Kaitian said:
Am I the only one who laughed at this?
Click to expand...

no i nearly laughed out loud... literally... which is something i never do especially here at work :)

I'll give the guy the benifit of the doubt for now, is does seem to lack any real detail (beyond the claim that it's "unlimited") but there has been research and development in this field. Anyone play the voxelstein game? Imagine that x100.

I've also heard some stuff carmack has said about the topic, how it might be coming soon:
http://en.wikipedia.org/wiki/Id_Tech_6

+1 for the bad voice-over and video demonstration... and holy hell why not let some artists have a go with it, surely many would do so for free, just pay for them to sit there while you watch them if you are so protective over the tech... without a visually stimulating demo it just won't impress very many folks.
 
The developer behind Minecraft has a few choice words about the people that put out that "Unlimited Detail" video we posted yesterday (here). Wow, you don't get much more to the point than that...and this guys knows what he is talking about.

"They’re hyping this as something new and revolutionary because they want funding. It’s a scam. Don’t get excited." Or, more correctly, get excited about voxels, but not about the snake oil salesmen.
Click to expand...
 
If this technology is nothing new, why do we still see inferior graphics in games released in 2011?
The tech demo where the compared the palm tree, showed just how ridiculously bad today's game engines are compared to the atom-unlimited system.

I'm presuming the problem is computing power, BUT in the video, the voice-over claims that his engine works on today's hardware at decent frame-rates.
 
aggies11 said:
More standard Qs: what is the time/space memory tradeoff for your way of storing the data? Is it fast/slow to access, large/small to store, does it take a long time to pre-compute the datastructure so that realtime changes/movement (eg. animation) becomes prohibitive?
Click to expand...

WorldExclusive said:
How much RAM this will use during gaming?

Can consoles take advantage of this tech as well as PC? (Very important, without console support it will not catch on)

Can this be used in animation?
Click to expand...

sunama said:
If this technology is nothing new, why do we still see inferior graphics in games released in 2011?
The tech demo where the compared the palm tree, showed just how ridiculously bad today's game engines are compared to the atom-unlimited system.

I'm presuming the problem is computing power, BUT in the video, the voice-over claims that his engine works on today's hardware at decent frame-rates.
Click to expand...

I think these are great questions. Yet they all refer back to on big questions: How does it work? I'm not just talking about the rendering of the image, but the system as a whole. Which PC components will receive which tasks during a 3D intensive experience and how much stress will those components be able to cope.

I see this technology as completely changing the way that 3d images are rendered in today's games, but that could possibly give us new bottlenecks as far as hardware. It all comes down to how does it work.

I think this would be a great article to have on hardocp that would surely receive many clicks ;)
 
If the question was asked, "How does this thing actually work?"
Will the makers of tech will give all their theories/algorithms/secrets up?

What we need is for some form of benchmark to be done on today's hardware, with an FPS counter. If the FPS is high enough, then this technology, by right, should obliterate and effectively re-write the rule book when it comes to games (graphics) engines.

They need to release a tech demo, like 3dMark used to release on an almost yearly basis. Joe Public could then download the demo and run it on their own PCs.

Also, my question still stands:

If this technology is old and nothing ground breaking, why are we seeing vastly inferior graphics being used in games being released in 2011?
 
I wouldn't say its a scam. Far from it. It is obviously working, though its pretty obvious that at present, their tech can only be used on static objects.

However, this isn't necessarily a bad thing.

What game developers can do is use the unlimited engine to create the background/static objects and for anything that moves, they can use a conventional polygon based engine. They make no claims about moving/animated objects, so I don't see how the word "scam" can be used.

There is no denying that that tech demo was damn impressive (even if their engine is limited to non animated objects). The palm tree comparison, the elephant statue and the rock put the graphics engines being used by current games, to shame.

The question has to be what are the hardware requirements. If we need ludicrously powerful PCs that are not yet commercially available, then perhaps this graphics engine isn't so great. Everything hinges on the hardware requirements.
 
A quote from the video "we increased it so far, we could abandon polygons all together and move to little atoms and run them in unlimited quantities"

IMO, that statement alone is enough to say they're claiming that they can at least match the ability of polygons.
 
It depends on the hardware they are using though.
If they have unlimited processing power (say, for example), then of course they could abandon polygons.

We need to know what sort of hardware they are using during those tech videos.
 
Earlier in this thread, somebody stated that these guy had been given $XX million, already. Is this true?

If so, this a heck of a lot of money.
 
sunama said:
Also, my question still stands:

If this technology is old and nothing ground breaking, why are we seeing vastly inferior graphics being used in games being released in 2011?
Click to expand...

You can make a really nice tree with their voxel tech. Problem is you can only make one, so in some game your forest will look stupid because every tree will be identical. Either that, or you get to buy a game that will require you to have thousands of hard drives to store the data for the forest. Have fun rendering all that data, too.

There's no shortcut to storing "unlimited detail."
 
Aluisious said:
There's no shortcut to storing "unlimited detail."
Click to expand...

Is this not what these guys are claiming?

If you listen to the video, he says that he has created a huge island, made from atoms. He can also run that huge island, at decent frame-rates.

The problem is that he didnt state exactly what hardware was being used.

The video was running at what appeared to be an acceptable frame rate.
 
I'm not jumping to any conclusions yet...but I wouldn't take a company's word over a potential competing company's word. Everyone wants to be able to roll out a revolutionary new way to create real time graphics. Whos to say the company calling this a "scam" isn't just trying to spread FUD to drive attention or credibility away from this "unlimited detail" project for their own market gain?
 
stop!theradio said:
I'm not jumping to any conclusions yet...but I wouldn't take a company's word over a potential competing company's word. Everyone wants to be able to roll out a revolutionary new way to create real time graphics. Whos to say the company calling this a "scam" isn't just trying to spread FUD to drive attention or credibility away from this "unlimited detail" project for their own market gain?
Click to expand...

Minecraft devs? They're selling a game...

These guys are selling rendering algorithms...

They're not competing at all.
 
They're not competing for video game sales dominance, no. A lot of companies do, however, take pride in developing their games using their own graphics engines and rendering algorithms - stuff they can patent and license out/sell. For a developer to be able to sell/license their engine or keep the proprietary graphics engines of their own games modern and competitive, the best thing they could do is 'scare' consumers away from the competition. In this case we have an engine developer who isn't well known at all and who is making pretty crazy claims. It's a golden opportunity that I wouldn't blame popular game/engine developers for taking advantage of. 'Unlimited' is claiming they've been able to nearly perfect an amazing engine that topples basically anything on the market. Hell for all we know the minecraft devs are constantly hard at work perfecting their own 'unlimited' engine. Maybe they decided to spread FUD in order to scare people away and buy time to bring their own product to the level that 'Unlimited' is at. Of course they would want to shine a negative light on their competition, even though their competition isn't exactly selling games.


And to address a thread-wide derp issue, I don't think some of you are understanding that they are NOT claiming to have invented the actual rendering style their engine is based on. Call it voxels, call it what you like because it doesn't matter. All they've done is (claimed to have) created a powerful engine that can run at good framerates even on 'regular' systems.


Hell, with the way some of you are acting I'm surprised you don't get all butthurt when a dev brings out an awesome engine that looks and runs great but uses polygons and shaders. Are you gonna get all pissed at them and claim they are a thief just because they didn't invent polygons? No. They created a polygon-based engine of their very own. It doesn't mean they are claiming to have invented the polygon. Same goes with Unlimited Detail. They created an engine of their very own that happens to seemingly be voxel based. Doesn't mean the engine isn't their own in every way, and doesn't mean they are claiming to have invented the core idea behind their engine.
 
You must log in or register to reply here.
Back
Top