• Hey, guest user. Hope you're enjoying NeoGAF! Have you considered registering for an account? Come join us and add your take to the daily discourse.

Euclideon (aka "unlimited detail" voxels company) presents Geoverse

Euclideon specialise in providing agile enterprise solutions that improve productivity and the usability of large volumes of point cloud data with our ground-breaking Geoverse™ solution. Read more on Euclideon.com

Video: http://www.youtube.com/watch?v=t8rsEJoh6mQ&feature=youtu.be

Still watching, but it does not seem to target the video game industry anymore.

And Carmack comments:

@notch and @ID_AA_Carmack Hey guys? The "geniuses" from Euclideon are back for more: http://youtu.be/t8rsEJoh6mQ

@xDeaD_MaNx @notch that looks like a much better business target for them than game engines
 

Haunted

Member
Am I the only one who thinks this is wholly inappropriate music for a business pitch in the geospatial industry? >_>
 

E-Cat

Member
Am I the only one who thinks this is wholly inappropriate music for a business pitch in the geospatial industry? >_>
You have to admit, this is pretty epic. :D

I want to model my hometown using lasers and explore it in the Oculus Rift.
 

drexplora

Member
he has claimed in the past that it runs on pretty standard laptop hardware without taking advantage of any GPU resources, totally in software.

I would wage that his 3d virtualization app could be put on the 360 or ps3.
 

Durante

Member
Carmack is spot on, as usual. The issue with any of these voxel technologies for gaming was always animation and deformation.
 

KKRT00

Member
They've just got very rich. Great technology for such applications and google maps in several years will be amazing.

I'm only curious how big files are, but for web datacenters it will be amazing.
 

Alienous

Member
I imagined that it would be useful for skyboxes, and 'out-of-reach' buildings, in a game. An polygon/voxel mixture.

Anyway, this is a better application.
 
Carmack is spot on, as usual. The issue with any of these voxel technologies for gaming was always animation and deformation.

Maybe with the use of the GPU they can use it for animation? I have never done animation or even tried so don't quote me on this.
 

Vesper73

Member
Good on them.

This seems like the best solution for static voxel geometry that exists today. They deserve whatever success they achieve.
 

FoxSpirit

Junior Member
Why do they film a Laptop and release this as an official presentation?
Because a Laptop is not that powerful, no matter how strong.

And yeah, good on them... and Carmack. Amazing tech, too bad unless someone solves the animation/deformation problem we will never see this in games.
 

bobbytkc

ADD New Gen Gamer
Maybe with the use of the GPU they can use it for animation? I have never done animation or even tried so don't quote me on this.

I don't think it can do very sophisticated animations. The way it works right now is like a 3D picture. The have a point cloud which they basically organize so that they can search for the points you need very efficiently, depending on the angle you are looking. You basically need to calculate and store a different point cloud for each frame of animation. Sounds unfeasible to me. It is however, extremely practical for what they are selling it for, 3D visual walkthroughs.
 
Video Is back
http://www.youtube.com/watch?v=Irf-HJ4fBls

I think the Geo thing might be a gateway to other industries. They'll just show off their technology in that industry and it will be taken to other industries.

Euclideon Technology Demonstration for the Geospatial Industry.

As many critics pointed out, if you have the ability to run terabytes of point cloud data, how do you store that much? And how would you load it?

This is a problem inherent to laser scanning, which generates a lot of data. Geoverse is our Geospatial product for running laser scanned data.

We have spent the last year building technology that overcomes all the loading and memory problems that happen when you use terabytes in places where people would normally use megabytes. If you are in the geospatial industry then we hope that our work can be of help to you.

I'm afraid the video isn't very interesting from a gaming point of view and we are aware many have expected a game by now. This does not mean we have abandoned games. If you had the ability to laser scan in the real world and run it, would you abandon games?

You will hear from us again soon.


Kindest Regards
The Euclideon Team.

ok
 

Zaptruder

Banned
Man... this is pretty incredible stuff. Funny that I missed it the first time it was posted.

Although the direct suitability for gaming seems low... is it possible to create a composited hybrid engine that can combine animation with static voxelized point cloud formations?

I mean... the issue is data manipulation - if they can retrieve the data as fast as they claim, then the trick would be to manipulate the data and then retrieve the manipulated data state - which should in turn show an animation.

I wonder how much such latency such a function would create; the time to transform the data, then create a point cloud from it, and then update the data, while at the same time reading the data on a per frame per pixel basis.
 

Zen

Banned
Carmack sounds incredibly salty for someone that hasn't introduced a real engine innovation in years that hasn't been a dead end. I guess Mega Texturing is actually used in like 3 games ever, took years and years, and has been rendered pretty much irreverent with the improved hardware in the industry that developers can work with without any of the compromises that come from MegaTexturing.

Euclideon always seems to be more like a research and solutions company than people making a 'Duke Nukem Forever' as it were.
 
Oh boo I thought this would be a more recent update. Since their first video I have been curious to see how far they can go, especially with video games. It's a real shame that it's been way over a year (or maybe two) and still nothing on the games front.
 

Ether_Snake

安安安安安安安安安安安安安安安
That video is basically showing what zbrush already does.
 
ZBrush actually does things somewhat differently, especially with the renderer. Rending a whole scene at almost real-life-level detail with actual laser-scanned objects is leagues different from a sculpting-based modelling program that uses voxels.


But, at any rate, as someone who has been keeping track of the Unlimited Detail project over the past couple of years, and someone who has long decided that the project couldn't be any further from being a 'scam', Euclideon's behavior over that time has been fairly interesting. Some notes:


  • The focus on Geoverse isn't really that unexpected if you've been keeping tabs on what Bruce Dell has been saying - he's mentioned that there's been interest from industries other than the gaming industry for this tech. Geospacial is pretty much the first obvious choice, and the most applicable for how the project was going during its 2011 demo.
  • There's been plenty of assurances that the company is still aiming for a gaming-ready product for developers. Geoverse is simply easier to implement and get out there to actual clients (and thus has been the main focus of development until its release), and all indicators say that Geoverse is ready and has been distributed, especially considering that there's an news update on the offical website for Geoverse 1.2 detailing improvements suggested by clients.
  • Ever since Euclideon has revived its website and facebook pages, both have been updated semi-regularly - until June last year. The fact that there's only been one update on the website since June and absolutely none on the facebook page since then potentially indicates that the company has stopped focusing on Geoverse (primarily, anyway, likely a small team dedicated to updates) and focused on other applications, most likely gaming, and they really don't have anything ready to show until then that hasn't already been shown. This likely includes the next big gaming-based demonstration that was said by Bruce Dell to include assets made by ex-THQ artists.
  • There's also the possibility that at least one gaming studio is using and testing Unlimited Detail as we speak. Euclideon, being mainly a tech development firm, would need to have the technology tested with an actual game to make sure it's ready for the masses. My first guess is Valve, considering Source 2, and I sincerely doubt Valve would want to release an engine that's more or less like everything else on the market - a fully-functioning Unlimited Detail in Source 2 would pretty much cement it as the top engine in the new generation, there would be literally no way to top it. It would be the reveal of the original Source engine all over again.
  • From what we know, the main challenges towards a gaming-ready engine since 2011 included lighting and animation. The former has apparently likely been already been solved by now, as a real-time lighting demonstration was released back at Gamescom around that time (I think it was Gamescom, but I'm not quite sure, but Crytek's boss sung its praises at the same event), and the progress of the latter is unknown, though existing videos show it can be done, it was at a somewhat primitive state at that time. Other people have been working on voxel animation before with varying results, so it is possible.

As a supporter of this tech (partly because there's more than enough evidence that it'll eventually come out, and admittedly a little bit of patriotism on the side, as Euclideon is an Australian company), I'm cautiously optimistic about it, and I'm hoping it'll be a game-changer in the industry. Its list of features are far more exciting than anything else in the industry, tech-wise, and could potentially take away the focus on polygons and shaders and all that crap and focus on technologies that actually have an impact on the gameplay.

A slight refresher on what UD can do:


  • Extremely efficient voxel renderer, finding every visible voxel in view and assigning it to a pixel on the screen - no more, no less. The 2011 demo was running entirely in software at 30fps. Yeah, that's pretty hardcore. Enables unprecedented rendering for far less computational work, and renders techniques such level of detail models and billboarding completely obsolete.
  • The only real 'limit' to the detail that can be rendered, no matter how much stuff is in a scene, is screen resolution. Bruce Dell has mentioned that Euclideon has talked with graphics card manufacturers, and the possibility of special graphics card features to specially render 4k and 8k resolutions is certainly interesting for both sides.
  • Point cloud format that's apparently relatively size-efficient and can be converted over from polygon formats.
  • Capability for laser-scanning. Everything in the 2011 demo was laser-scanned right from the real world. Yep. This in itself is extremely exciting.

Still, until we see something concrete for the gaming side of things, it's going to be something of a wait-and-see sort of thing. But, hopefully, as long as they don't give up on the project, we might see something within the new year, a new gaming-based demonstration (hopefully a downloadable demo) would be fantastic, at least.
 
Carmack sounds incredibly salty for someone that hasn't introduced a real engine innovation in years that hasn't been a dead end. I guess Mega Texturing is actually used in like 3 games ever, took years and years, and has been rendered pretty much irreverent with the improved hardware in the industry that developers can work with without any of the compromises that come from MegaTexturing.

Euclideon always seems to be more like a research and solutions company than people making a 'Duke Nukem Forever' as it were.

Funny but that's just not true...

MegaTexturing as a technology has made int's way into hardware as a result of Carmacks work.

All modern GPU architectures support Partially Resident Textures, which is really what Carmack was trying to achieve but in software, on hardware that didn't really lend itself to it.

His influence on the industry today is still there, just alot more understated than in the Doom days.
 

Zaptruder

Banned
ZBrush actually does things somewhat differently, especially with the renderer. Rending a whole scene at almost real-life-level detail with actual laser-scanned objects is leagues different from a sculpting-based modelling program that uses voxels.


But, at any rate, as someone who has been keeping track of the Unlimited Detail project over the past couple of years, and someone who has long decided that the project couldn't be any further from being a 'scam', Euclideon's behavior over that time has been fairly interesting. Some notes:


  • The focus on Geoverse isn't really that unexpected if you've been keeping tabs on what Bruce Dell has been saying - he's mentioned that there's been interest from industries other than the gaming industry for this tech. Geospacial is pretty much the first obvious choice, and the most applicable for how the project was going during its 2011 demo.
  • There's been plenty of assurances that the company is still aiming for a gaming-ready product for developers. Geoverse is simply easier to implement and get out there to actual clients (and thus has been the main focus of development until its release), and all indicators say that Geoverse is ready and has been distributed, especially considering that there's an news update on the offical website for Geoverse 1.2 detailing improvements suggested by clients.
  • Ever since Euclideon has revived its website and facebook pages, both have been updated semi-regularly - until June last year. The fact that there's only been one update on the website since June and absolutely none on the facebook page since then potentially indicates that the company has stopped focusing on Geoverse (primarily, anyway, likely a small team dedicated to updates) and focused on other applications, most likely gaming, and they really don't have anything ready to show until then that hasn't already been shown. This likely includes the next big gaming-based demonstration that was said by Bruce Dell to include assets made by ex-THQ artists.
  • There's also the possibility that at least one gaming studio is using and testing Unlimited Detail as we speak. Euclideon, being mainly a tech development firm, would need to have the technology tested with an actual game to make sure it's ready for the masses. My first guess is Valve, considering Source 2, and I sincerely doubt Valve would want to release an engine that's more or less like everything else on the market - a fully-functioning Unlimited Detail in Source 2 would pretty much cement it as the top engine in the new generation, there would be literally no way to top it. It would be the reveal of the original Source engine all over again.
  • From what we know, the main challenges towards a gaming-ready engine since 2011 included lighting and animation. The former has apparently likely been already been solved by now, as a real-time lighting demonstration was released back at Gamescom around that time (I think it was Gamescom, but I'm not quite sure, but Crytek's boss sung its praises at the same event), and the progress of the latter is unknown, though existing videos show it can be done, it was at a somewhat primitive state at that time. Other people have been working on voxel animation before with varying results, so it is possible.

As a supporter of this tech (partly because there's more than enough evidence that it'll eventually come out, and admittedly a little bit of patriotism on the side, as Euclideon is an Australian company), I'm cautiously optimistic about it, and I'm hoping it'll be a game-changer in the industry. Its list of features are far more exciting than anything else in the industry, tech-wise, and could potentially take away the focus on polygons and shaders and all that crap and focus on technologies that actually have an impact on the gameplay.

A slight refresher on what UD can do:


  • Extremely efficient voxel renderer, finding every visible voxel in view and assigning it to a pixel on the screen - no more, no less. The 2011 demo was running entirely in software at 30fps. Yeah, that's pretty hardcore. Enables unprecedented rendering for far less computational work, and renders techniques such level of detail models and billboarding completely obsolete.
  • The only real 'limit' to the detail that can be rendered, no matter how much stuff is in a scene, is screen resolution. Bruce Dell has mentioned that Euclideon has talked with graphics card manufacturers, and the possibility of special graphics card features to specially render 4k and 8k resolutions is certainly interesting for both sides.
  • Point cloud format that's apparently relatively size-efficient and can be converted over from polygon formats.
  • Capability for laser-scanning. Everything in the 2011 demo was laser-scanned right from the real world. Yep. This in itself is extremely exciting.

Still, until we see something concrete for the gaming side of things, it's going to be something of a wait-and-see sort of thing. But, hopefully, as long as they don't give up on the project, we might see something within the new year, a new gaming-based demonstration (hopefully a downloadable demo) would be fantastic, at least.

I'm confident that this technology will be fully feature ready and incorporated into a robust game engine by 2025... in time for a fully functioning metaverse style VR-MMO.
 
Top Bottom