Euclideon & Unlimited Detail - Bruce Dell Interview @ [H]

Hellbinder said:
There are a couple posters in this thread that telling them what a couple of ignoramuses they are would be almost worth getting banned over.

I mean seriously... you two are apparently completely blind, and about as intentionally uninformed about what someone has said (in a freaking video no less) as i have ever seen. :rolleyes:


Nice job on the interview and followup on this interesting development in graphics!
Click to expand...

Maybe some of those people have more knowledge on the subject than you do, and possibly work in the industry?

Just some food for thought.
 
Well guys I just have to say I am totally amazed at how many 3D graphics coding experts we have on this forum.

Its impressive. Or is it just the volume of hot air I'm feeling?

I look forward to what these guys in Australia have to deliver.
 
Any one complaining about the size the thing would have to be, think about it, if you say that the thing is 100GB, (which we have been told that it fits on a DVD so obviously not), that elephant would be 4-5 gigs, the ground a few more, pedestal a few more, that would be higher then the amount of ram, 8 gigs, so you think he is fetching this from the HDD? If he could render this straight from the HDD it would be a feat in itself.

Then we are told that this runs on 1 core, that in itself is amazing, think of what your gpu renders in a game, then ask yourself how fast it would run on the cpu only, you would be lucky to get a few fps.

Even if this would only work on static objects, think of the scenery this could render, only on the cpu then you could do all sorts of things with the gpu, you could essential focus the whole gpu on only animated objects. You would be able to make really good Characters and NPC because that's all it would be rendering.

If they can deliver this then it will be the greatest leap since going from 2D to 3D, because the level of detail done by their engine is amazing. Hopefully in 5 years this will be able to come out and amaze us.
 
Jaer44 said:
In the original video posted, he did explain the repetitive objects, he simply stated that they are not artists and he also apologized for the reuse of items on the island, stating that once actual artists get the engine in their hands, they'll see what it can really do (or something to that measure).

I personally don't know what to think of the technology, I honestly hope that it can be implemented in future game engines etc, but only way to know is to wait and see.
Click to expand...

Yes, that's how he explained it. However, if you look at all the older demos in the interview video, they *all* seem to rely on repetition. The old video with "lots and lots and lots and lots of objects (sic)" also contained only a few unique objects, repeated over and over. If every single demo relies on repetition, you naturally begin to wonder whether its a limitation of the graphics engine. He did say it only takes 15 minutes to laser-scan an object into their engine. There are also tools that can generate elevation meshes so the island didn't have to consist of very detailed Lego blocks.
 
Still a lot of hate going around for these guys. I give them major points for trying to do things different. I hope it is successful.
 
SoulHunter said:
Any one complaining about the size the thing would have to be, think about it, if you say that the thing is 100GB, (which we have been told that it fits on a DVD so obviously not), that elephant would be 4-5 gigs, the ground a few more, pedestal a few more, that would be higher then the amount of ram, 8 gigs, so you think he is fetching this from the HDD? If he could render this straight from the HDD it would be a feat in itself.

Then we are told that this runs on 1 core, that in itself is amazing, think of what your gpu renders in a game, then ask yourself how fast it would run on the cpu only, you would be lucky to get a few fps.

Even if this would only work on static objects, think of the scenery this could render, only on the cpu then you could do all sorts of things with the gpu, you could essential focus the whole gpu on only animated objects. You would be able to make really good Characters and NPC because that's all it would be rendering.

If they can deliver this then it will be the greatest leap since going from 2D to 3D, because the level of detail done by their engine is amazing. Hopefully in 5 years this will be able to come out and amaze us.
Click to expand...

Seeing is believing brother and as I said in another post, I wish I could show you all what I got to see and tell you all everything I know. But I understand why Euclideon need to protect their product for now. I was a beta tester for years and it all ways amazes me when they get to play with or just see an unfinished product they cry like little kids. Sit back people the time will come when you can play with the new toy if you hate it then say what you will. :)
 
Last edited:
SoulHunter said:
.../QUOTE]

You are underestimating the needs. Let's do some math. Imagine a precision of 1mm for one point. You need to represent the point in space, that menas 3 variables with long type (X, Y, Z). That is 3x4 bytes. Then we should have it's color and alpha too, that is another 4 bytes (R,G,B,A). This means we need at least 16 bytes to represent objects up to 4 294 967 295mm = 4294.9km (unfortunately short integer which is 2 bytes is too small for these needs, it can represent only objects up to 65m).

16 bytes per point. Let's create a plane from these points, with that 1mm precision, for example with size of 100x100m. That is 100k mm x 100k mm = 10 000 000 000 points x 16 bytes per point = 160 000 000 000 bytes = 149.01GB for one, simple 100m^2 plane with no terrain deformations.

Small objects like that elephant are doable this way, for example the typical person has around 2m^2 skin surface area. That is 2000 mm x 2000 mm = 4 000 000 points x 16 bytes per point = 61.03MB per person.

So now how to represent surface areas - because it is clear that point clouds are not good for this. Of course we can cheat there and represent the rough surface area by the good old polygons, place the repetitive items on this polygon surface area. But then the floor will behave differently than rest of the engine, imagine for example explosions which create a hole and throw the dirt around. You can save space by using many models of grass (each will be small, but if you combine hundreds of grass models, it shouldn't look repetitive) and of course many, many tree models.

It is not the small models which are the problem with low enough resolution - it is the surface area which in any meaningfull game is much, much bigger than 100m2. You can repeat models for small world items like grass, trees. What you can't do effectively is the topography of the area.
Click to expand...
 
JohnnyGatt said:
Sit back people the time will come when you can play with the new toy if you hate it then say what you will. :)
Click to expand...
You know us gamers, IT peeps, computer enthusiasts are some of the most impatient people on the planet. :D
 
I'll say this as cleanly as possible since I'm in a basic forum and not in a Genmay type setting.

The concept is there, the premise sounds VERY promising. Ultimately the things causing the most controversy is the fact that the company makes such bold claims that challenge everything we were taught to know and then disappears and then leaves us all wanting.

It's like finally getting the girl of your dreams to agree to a night of well yknow and then only letting you put the tip in, then stopping you and then saying. We can do the rest in a years time. LOL :D

I think Euclideon is giving gamers Blue brains if you get what I mean. That is sure to cause some angry tech elite and gamers :)
 
Looks cool and shows promise. I can't wait until I can use this technology in a game and evaluate it for myself.
 
SoulHunter said:
Then we are told that this runs on 1 core, that in itself is amazing, think of what your gpu renders in a game, then ask yourself how fast it would run on the cpu only, you would be lucky to get a few fps.
Click to expand...
You can find out yourself using SwiftShader.

Note that Unlimited Detail's demo does not appear to be using any advanced shading, while SwiftShader can run Crysis at up to 20 FPS on the latest mainstream CPUs. So in that respect, the UD demo isn't extremely impressive. Modern CPUs pack a lot more power than many people seem to realize. It's been over a decade since games were all software rendered, and CPUs have gotten a lot faster since!

And they're about to take another massive leap in performance. AVX2 will add support for 'gather' and 'fused multiply-add' instructions, which both used to set the GPU apart from the CPU.

So hopefully Euclideon will continue to wander off the beaten path and use the superior flexibility of the CPU to create a unique graphics experience.
 
Random thoughts as I read the last few pages:

If you go through the history of technological progress, youll find many people who created something new and radical from being segregated from the intellectual base in that field. This guy claims he is one of them so his way of thinking may be completely different then say John Carmack.

Bill Gates famously said a LONG time ago, that the PC will never need more then 640kb of memory...never say never, especially when it comes to technology.

There are a lot of very educated people in these forums (me excluded :p ), but some of these educated people are showcasing a very nasty human flaw that comes with age most often but can be exhibited any time in life: stubbornness. It's alright to be pessimistic but to close your mind to the idea even when your being given something that is demanded, thats just ignorant and really, shows that you should not be involved in the tech industry. This is a place of ever-changing ideas and notions on how things can be done. If you were to go back in time and confront ANY inventor and show them the current state of their invention, they would say its impossible.

Loose example: The 2.9L engine of the Ford Model T (yes i know Henry Ford did not invent the car/internal combustion engine) created 20hp and had a fuel economy of around 20 miles to the gallon and a top speed of around 45mph. The Toyota Prius of today has a 1.8L engine that creates about 98hp and has an mpg of about 48. These numbers do not involve the electric engine btw...so please dont argue that.

I'm sure if someone went to Ford or one of his engineers of the day and said, I want you to create an engine that is, roughly, 5 times faster but uses about 2.5x less fuel oh and it has to be a full liter less in size....that engineer would tell you its impossible...you'd never be able to do it because X or X and X.
 
Great video. I think this technology is awesome and hope to see it further developed soon, so we can PLAY AWESOME GAMES. The moniker of "unlimited" has yet to be seen though.

Very impressive tech considering they are rendering only in software. When they get things going with video card hardware we will hopefully see a lot more improvement in the graphics,

But gameplay still rules, no matter what the graphics look like of course.
 
USMCGrunt said:
Random thoughts as I read the last few pages:

If you go through the history of technological progress, youll find many people who created something new and radical from being segregated from the intellectual base in that field. This guy claims he is one of them so his way of thinking may be completely different then say John Carmack.

Bill Gates famously said a LONG time ago, that the PC will never need more then 640kb of memory...never say never, especially when it comes to technology.

There are a lot of very educated people in these forums (me excluded :p ), but some of these educated people are showcasing a very nasty human flaw that comes with age most often but can be exhibited any time in life: stubbornness. It's alright to be pessimistic but to close your mind to the idea even when your being given something that is demanded, thats just ignorant and really, shows that you should not be involved in the tech industry. This is a place of ever-changing ideas and notions on how things can be done. If you were to go back in time and confront ANY inventor and show them the current state of their invention, they would say its impossible.

Loose example: The 2.9L engine of the Ford Model T (yes i know Henry Ford did not invent the car/internal combustion engine) created 20hp and had a fuel economy of around 20 miles to the gallon and a top speed of around 45mph. The Toyota Prius of today has a 1.8L engine that creates about 98hp and has an mpg of about 48.

I'm sure if someone went to Ford or one of his engineers of the day and said, I want you to create an engine that is, roughly, 5 times faster but uses about 2.5x less fuel oh and it has to be a full liter less in size....that engineer would tell you its impossible...you'd never be able to do it because X or X and X.
Click to expand...

You're right. Always expect unexpected things to come out. Tech is progressing very rapidly in our time, and I always look at these things as a positive. Even if this company fails (I doubt) someone else will pick up the slack.

Remember guys, it wasn't TOO long ago we were playing 2D NES or Virtua Fighter! Also Moore's Law still applies to some degree.
 
I just wanted to post this video in case anyone missed it:

http://www.youtube.com/watch?v=tkn6ubbp1SE&feature=player_embedded

download the source code... hmmm, maybe that is what euclideon did recently? :)

To me, the explanation made by Carmack as to why we do not yet use voxels (or whatever) makes the most sense. Even if that environment had more unique objects and looked a bit better, you have to add the rest of the game on top of that, and it is not hard to imagine we are not quite at the tipping point. What is obvious though, is that we are close.

Anyone remember the angst that occurred when 3d started becoming useful? No? Well, the internet was not quite as big back then... I remember it was pretty exciting. Games used true 3D way before Quake, but Carmack's ability to time out when it would become truly useful made that a good tipping point reference. We have also seen voxels used in games for a while. I think the first time I noticed it was in Blood (which was around Quake release, time-wise) and it was easy to imagine that being the future.

When will that tipping point occur? If it is in five years like Carmack says that will be cool to witness. If it is sooner, then great. Later, like 10 years? I will be alive then probably so no biggie. What I am most excited about is cutting up people and seeing the layers upon layers of innards... Think Sim Butcher... I have thought about this type of game since I was 12 I think... This was WAY before I even heard the term voxel, I somehow knew it was the future of graphics, just imagining bitmaps but in 3d.

Seriously, if it is so easy for a 12 year old to imagine, and I am hardly that smart in the large scheme of things, then it probably will become reality. I am 28 now btw... and very peaceful (in case anyone was sworried)
 
Last edited:
After all the complaining and moaning brought forward by a few users in this thread, I feel like letting the [H] staff know, that a video like this nothing but a great addition to everything else being offered here (for free).

Not only is it technology news, it´s relevant, t´s an exclusive interview, it´s thought provoking and interesting for people who enjoy and consume real time graphics one way or another.
 
I liked the demo videos and it looks impressive. And the Unlimited detail seems to be an appropriate name for the technology when looking at the statues and other 3d scanned objects as they move in and out without any noticeable LOD popup, but their technology seems to be very limited in variety and that is where I hope they can improve and show in their next demo.

It doesn't matter if you have an extremely detailed small area or building block in games. I certainly would hate to see what I saw in the videos used in a RPG game, all the levels would be exactly the same. That was one of my biggest problems with Dragon Age 2. They repeatedly used every level over and over and over and over again. It looked great once, then after the 3rd time of going to a supposedly new room/area, you are greeted, yet again, with an area you finished fighting just a little while ago. That gets old very fast.
 
ianken said:
Show it to me when all the trees don't look the same. Or the blocks. Or everything else. Point clouds. I'm sorry "atoms" suck memory like mad, and the fact that their demo is copypasta in the extreme shows that.

Later in the video they show a 2.5 year old demo and he states "there are lots and lots of objects" No. There are not. There are at most a dozen being replicated over and over again. It's a misleading claim.

They have a search that pulls an "atom" from the scene and assigns it to a pixel on the display. OK. Where are the atoms stored? in magic "unlimited" memory?

TANSTAAFL.

I'll belive their claims when I can download and run the demo myself.
Click to expand...

I think their logo island is parametrically generated, which I don't fault them for. They wanted to make it huge so they could show it rendering vast distances, you just aren't going to do that without a lot of repetition, and preferably let a computer do it based on a bitmap of your logo.

But there IS a solution to the memory problem that does satisfy the definition of "unlimited". Basically, if you scan your physical object, you get a point cloud. You can convert your point cloud data to a nurbs object, and the search can take place on the nurbs object. Doing that leaves you the ability to in theory calculate any point on the curve at any resolution, while at the same time using up only a small amount of storage.

I suspect that would be pretty computationally intensive, but it gets around your gripe on storing information. texturing it would be a PITA though.

But I can also think of other methods. Like store a sparse point cloud of your object. You choose your level of detail contained in the point cloud, keeping it sparse so it is reasonable to store. Then you just have to have a smart strategy for smoothing over the gaps in your sparse point cloud. Like fitting a bunch of pre-defined NURBS meshes to the points nearest your chosen "atom", repeating it with a selection of nearest neighbor points until you get your smoothest set of patches, and then using your chosen patch to retrieve your interpolated point from your point cloud. You could also replace the NURBS meshes with sparse pre-calculated meshes that stand in for a very dense pre-calclulated meshes that serve as your look up table for the point once you have calculated the correct mesh shape.

For large terrain, you just calculate a polygon that yo don't ever render that you scatter objects over to get a surface, you can use displacement maps while calculating the mesh to do easy complex terrain. You can use similar bitmap like object to map the types of objects you scatter.

Textures? You still have your texture library, which you use to calculate the color of any point. For anything in between two points, you use them to determine the bounding reagion for the texture map, and use your hero mesh I described above to calculate that points color.

HEck, you could even apply bump mapping/displacement mapping/bitmap controlled tesselation like effects by having a displacement map you can apply to your point cloud to determine offsets from the master point cloud. It'd just work like the texture plan above but cause you to jog your point cloud points.

The beauty is that you would only be doing the hairy math for things you got really close to. So the number of objects on the screen would be by definition very limited.

Not saying that is what they are doing, but jsut some ideas. You can always swap computation for storage in things like this, or compromise someplace in the middle.
 
OK, there must be more to this than meets the eye.

AA: This is going on somehow. When you zoom out far enough many "atoms" will make up one pixel.

Shading: you must do raytracing of somesort, and drawing all the atoms to do that for each light source would be hard.

Search algorithm: Derp. I have no idea how it is possible to compress Petabytes of data and index atoms so easily.

This looks real, but I can't understand it.
 
Wow, just got done moderating the YouTube thread. It reminded me of just how well our community is behaved. Took me an hour to get rid of most of the name calling and ridiculous and offensive statements. Anyway, that is a big THANK YOU to you guys for acting like adults.
 
sirmonkey1985 said:
/me scratches head.. did you actually watch the whole video or just skim through it?:D
Click to expand...

I did watch the whole thing. When he says "trillions of polygons" my bullshit meter pegs into the red.

Words have meaning. "Unlimited" has meaning. They picked that word, and they repeatedly use it. My guess is their definition of "unlimited" has more in common with the definition used by cell phone carrier data plans than how a person would normally interpret it.
 
Very cool tech, I think the best part about it is the idea that moving from platform to platform can be almost seamless. Heck, would it be past a game company to even make a version for a retro system, just as a bonus item?
 
ianken said:
I did watch the whole thing. When he says "trillions of polygons" my bullshit meter pegs into the red.

Words have meaning. "Unlimited" has meaning. They picked that word, and they repeatedly use it. My guess is their definition of "unlimited" has more in common with the definition used by cell phone carrier data plans than how a person would normally interpret it.
Click to expand...

While their defination of unlimited probably doesnt mean unlimited by exact defintion, I believe they mean more like...its unlimited to the point that you'll never hit the limit like you do with polygons.
 
USMCGrunt said:
While their defination of unlimited probably doesnt mean unlimited by exact defintion, I believe they mean more like...its unlimited to the point that you'll never hit the limit like you do with polygons.
Click to expand...

The engine scalability is technically unlimited (from what they are showing). The limit is your monitor resolution and your system.

The way they are using the term is accurate.
 
It is damn funny to see adults act like 10 year olds arguing if superman is better than batman. You guys need to get a fucking grip.

Wether this story is BS or not I enjoyed watching it and it was interesting. HardOCP is here to make money and getting an interview of a company like this gets a ton of hits. It is just good business, would you guys rather just see another page of the same shit every week. Honestly, how many power supply, mobo, and video card reviews do you want to read?

Do I question wether their engine will work? Absolutly.

Have any of you people read the short story "The Lottery" by Shirley Jackson? The point is you don't have to keep doing things the same way just because this is how it has always been done. It is possible this guy and his small crew have or are close to something new, but rather than wish the guy good luck on making it happen, people woudl rather talk shit to the guy.

Anyway, stories like this give us something worthwhile to discuss. It is nice to see something fresh on the front page. Some of you dudes are acting like Kyle exposed you to a video of 3 90 year old men jacking off, that is something you can never "un-see." This story you can easily forget if you think it's BS.
 
They compare the term "unlimited" to how colors are now 32 bit and that is a practical point at which no more colors are needed.

I'm not sure that's a great analogy, but he stresses it's in a "practical" sense. He also chuckles when the word is brought into question, suggesting to me he's aware it's a marketing term and not a technical one. Of course there are limits to hardware and software, anyone here should assume that much, right?

His comparison to google's search engine is probably a better analogy. There isn't "unlimited" things to search for, it just seems that way. He explains his engine will produce detail that "seems unlimited"...

Anyone else agree with the above?
 
hpglow said:
It is damn funny to see adults act like 10 year olds arguing if superman is better than batman. You guys need to get a fucking grip.

Wether this story is BS or not I enjoyed watching it and it was interesting. HardOCP is here to make money and getting an interview of a company like this gets a ton of hits. It is just good business, would you guys rather just see another page of the same shit every week. Honestly, how many power supply, mobo, and video card reviews do you want to read?

Do I question wether their engine will work? Absolutly.

Have any of you people read the short story "The Lottery" by Shirley Jackson? The point is you don't have to keep doing things the same way just because this is how it has always been done. It is possible this guy and his small crew have or are close to something new, but rather than wish the guy good luck on making it happen, people woudl rather talk shit to the guy.

Anyway, stories like this give us something worthwhile to discuss. It is nice to see something fresh on the front page. Some of you dudes are acting like Kyle exposed you to a video of 3 90 year old men jacking off, that is something you can never "un-see." This story you can easily forget if you think it's BS.
Click to expand...

Nothing to do with strictly wanting to shit talk the guy. Either he has something novel, in which case he should be patenting it (and thus describing the methodology used); he has something which is an extension or implementation of known methods, and he should be citing his sources (and thus... describing the methodology used); or he is bullshitting, and he should be relentlessly mocked. I'm not seeing much of a third option here.
 
USMCGrunt said:
Bill Gates famously said a LONG time ago, that the PC will never need more then 640kb of memory...never say never, especially when it comes to technology.
Click to expand...
Bill Gates never said that.

Dunno what to think of this tech, but if it's legit then awesome, 3D gaming has become pretty stale.
 
ianken said:
I did watch the whole thing. When he says "trillions of polygons" my bullshit meter pegs into the red.
Click to expand...

I'm pretty sure he said 'trillions of atom things' as in the 'voxel' points, not polygons.

Regardless, the ultimate idea, since this is using occlusion by design and because the level of detail on distant objects is seamless, you can have a rendered object be nothing more than one pixel rendered per object for each available pixel at the given resolution.

So most of us can only have 2,073,600 objects rendered at one time regardless.

On the discussion of repetitive objects and memory usage is completely moot. It's not even a proper tech demo we saw and nothing close to a working game engine. It's a proof of concept and considering how it was being run, I think it's been proven very well. The intent is to display as much stuff on screen with as much detail as possible as quickly as possible. Memory concerns or storage are part of a game engine, not a rendering method.

Aside from that, as it was mentioned a number of times, the cloud point data is highly compressible and would be easier to compress en masse than a number of UV/Bump/Displacement maps for a thousand different rocks, pebbles and cacti anyway. With a setup like this, any lighting would and should be handled in real-time by the GPU with shaders . Color and luminosity values (I expect) would be stored in the point data in a way that texture maps of any kind would become completely superfluous. This would even allow, eventually, per-'atom' control of values for a physics engine to use in collision, rebound, softness, etc.

This is, after all, much closer to how things work in the real world instead of various intersecting planes with paint on them.
 
§kynet;1037619418 said:
Bill Gates never said that.
Click to expand...
For those who want a somewhat reliable source on this:
http://www.wired.com/politics/law/news/1997/01/1484

Dunno what to think of this tech, but if it's legit then awesome, 3D gaming has become pretty stale.
Click to expand...
While I'm excited about more detail in games... it doesn't answer to making games more fun. But, hopefully this brings in some new and worthwhile gameplay elements.
 
I thank HardOCP for this interesting interview, but it really did not assuage any of my skepticism regarding this company's claims. Until I see proof, I am not a "believer". I'll eat my words if some small startup of less than 10 guys will turn computer graphics on its head.
 
folklore said:
Is this an April fools joke???
Click to expand...
I don't think its a joke but if I had to side I guess I would be a naysayer. Not smart enough to technicaly oppose it but the indian in me tells me that he's got an idea that he can't any farther and by the looks of it, neither does the industry.
 
It needs alot of polish and work, but appears to definately have some potential to me.

Port to run on the GPU, alot more item variety, 1920x1200 resolutions (high def tv and pc resolutions), and a bit smoother frames, and you could really have a beautiful looking engine.

From my understanding, items, objects are all scanned in, and the picture on screen just pulls pixels from a database as needed using search engine algorithms? Plus some lighting applied I would guess.

Need to also see something dynamic happening, such as a flame burning, candle flickering, and the lighting and shadows from that. Once they can add that in, I bet it looks beautiful.
 
DiabloRojo said:
On the discussion of repetitive objects and memory usage is completely moot. It's not even a proper tech demo we saw and nothing close to a working game engine. It's a proof of concept and considering how it was being run, I think it's been proven very well. The intent is to display as much stuff on screen with as much detail as possible as quickly as possible. Memory concerns or storage are part of a game engine, not a rendering method.
Click to expand...

Sorry but that's completely wrong. A game engine doesn't exist as a solid entity. It is a composite of more dedicated technologies, with the rendering engine included in that list.

The rendering is the rendering. If your rendering takes a ton of memory there's nothing the "engine" can do about that. If you believe otherwise then tell me why Epic doesn't just make a standalone raytraced renderer for UE3, then shove it down the pipe to the "engine" guys and say "hey your problem now, make it perform well!"

It simply doesn't work that way.

Algorithms and rendering methods have properties by the nature of how they accomplish the rendering, and the engine will inherit those properties.
 
1305081038HxNDdMp38s_9_1.jpg


This technology has shwick's seal of approval.
 
If it happens, it happens. The End.

If it does, everyone knows its a good thing, why would anyone care other than that?
 
You must log in or register to reply here.
Back
Top