Support NeoGAF

[FoxSpirit]

here's a detailed presentation pdf about what John Carmack is looking into.

It's similar tech to these guys
http://artis.imag.fr/Publications/2009/CNLSE09/GigaVoxels_Siggraph09_Slides.pdf

Full pipeline to render infinite resolution voxelobjects/scenes

That sounds VERY similar. Also, thanks to infinite resolution Aliasing is out of the picture.
And note that they implemeted it with CUDA, which UD Tech could also do. But jey, this would probably require a specialised programmer and those cost money. Just like real graphics artist

 

[Chittagong]

If they would combine this technology with native Gold45Revolver API support, we could see some really killer games.

 

[140.85]

It seems pretty sound. Not sure how they're going to store all that data.

First thing they gotta do to sell this is hire a designer and start importing converted models instead of generating all these ugly alien-like environments. And whoever is pushing them to release these demos before smoothing out the shadow bugs and other visual artifacts should be slapped as well.

 

[ckohler]

Intriguing tech hampered by very poor presentation. The narrator is too informal and rather condescending and their demos, while technically impressive, are full of shadowing glitches frame-rate stuttering and a high lack of polish. In the second video he admits that their tech demos were made by programmers and they aren't good artists but that's just a bad excuse. Why not hire an artist to fix that before releasing videos? Anyway, I'm still intrigued.

 

[MickeyKnox]

Narrator sounds like that one aspergers kid who makes those insane videos with old games, wtf was his name?

 

[Sloegr]

Narrator sounds like that one aspergers kid who makes those insane videos with old games, wtf was his name?

ulillillia

 

[tacotime77]

Lol this calls to mind the halo reach trailer where they were talking about how many more polygons they could use now than before. Now polygons aren't even cool anymore :lol

 

[Firestorm]

The second video is much more impressive than the first.

Also, this guy loves the word "Unlimited."

 

[I_D]

After looking at that demonstration, games just aren't the same. :lol

Since this runs on CPU exclusively, it shouldn't be that hard to devote a core/thread to it. Video games could look VASTLY better if this tech was used simply for the ground+short grass (and, of course, it could be added to other things).

People are doubting this guy, but it really does look like promising software. His presentation needs work, obviously, but he's an amateur, you can't expect him to have AAA skills.

 

[Slayer-33]

This is relevant to my interests, will keep watch.

 

[Let me in]

The second video is much more impressive than the first.

Also, this guy loves the word "Unlimited."

Love the way he says it... EVERY time. Unnnliiiiimited detail.

Also, the presentation factor is awful. The logo looks like it's from some Chinese shmup or something.

 

[User Name Here]

There was an update posted on UD's facebook page that I thought you guys might like to read about. Check it out: http://www.facebook.com/topic.php?uid=105528666147262&topic=952

They seem to be getting support from the Australian government.

Euclideon Receives a Commercialisation Australia Grant of $1,984,652

The Euclideon new 3D computer graphics technology with infinite geometry project has been supported with assistance from the Australian Government through Commercialisation Australia.

The grant will support programmers to convert the system from PC to game consoles, mobile phones and other platforms. Most 3D technology is created for computer games using angular polygons. Euclideon has developed what may prove to be a ground-breaking advance in 3D technology by creating a search algorithm that will result in 3D images with vastly higher geometry - while saving costs and offering greater portability between platforms. It will have benefits for architecture, games, mining, the sciences and many other industries.

The Minister for Innovation, Industry, Science and Research, Senator Kim Carr said "I am pleased to help talented researchers and entrepreneurs capitalise on their good ideas. Commercialisation Australia helps our most innovative minds take their products to the market, creating opportunities for business and high-tech, high-wage jobs for Australians. I wish all grant recipients success and look forward to seeing their inventions make a difference to the lives of Australians and people around the world.”

In past times the road from invention to commercialisation was a difficult one, and we can be certain that many good inventions disappeared before they could serve the public because of the many difficulties inventors had. Good technologies benefit us all and the Australian Government in their wisdom put schemes in place to ease the burden of those who wanted to bring new technologies to the people by offering financial assistance. I feel very fortunate to be Australian and live in a country were opportunities and government assistance is granted according to ability and is available to all Australians regardless of background. The core of this vital service is the efforts of the Commercialisation Australia team. I am aware that many hours of reading and careful investigation was required in order to find the inventions that are ready for market and we here at Euclideon would like to thank Commercialisation Australia and the Australian Government for the support they have offered to us and to all companies that are helping to push the boundaries of technology in this country.

 

[Scrow]

fuck you australian government. bring in an R18+ rating for video games already, you backwards, un-progressive, dick bags.

 

[mr_nothin]

i was walking down the street and this guy had a table set up with three coconut shells and he put a $5 bill under one and swapped them around and said that if i paid $5, i could guess which one it was under and if i was right i'd get to keep all the money...

so i did it and i won...

and then this chump takes out a $10 bill and puts it under and mixes them up again and asks me if i want to double or nothing and i'm thinking "oh man ur about to get so pwnt" because really how often do you get the opportunity to turn $5 into $20?

i'll report back later with how it goes!!!

The suspense has been killing me

 

[TheOGB]

Goddamn... and this was all in the beginning of last year?


Sssshhhhhheeeeeiiiit, why is this not all over the fucking place yet?

 

[balladofwindfishes]

New Video Showcasing Natural Settings comparing it to the first two Crysis games.
Some re-used footage from the last showing, but some newish stuff.

Wow, if they can pull this off and it runs on today's hardware, that'd be impressive and would make future hardware revisions pretty interesting.

That's assuming it's not some kind of crazy scam or something.
I wish they'd show an animation playing. Even if it's just a crude walk cycle.

 

[tci]

That's impressive. Let's hope it really is as good, and will be used in games soon.

 

[ShdwDrake]

Thats pretty sick

 

[xclaw]

Why do I now hate polygons and when can I play a game with this?

 

[Naked Snake]

New Video Showcasing Natural Settings comparing it to the first two Crysis games.
Some re-used footage from the last showing, but some newish stuff.

Wow, if they can pull this off and it runs on today's hardware, that'd be impressive and would make future hardware revisions pretty interesting.

That's assuming it's not some kind of crazy scam or something.
I wish they'd show an animation playing. Even if it's just a crude walk cycle.

I want to believe.

 

[Bowdz]

Heres to hoping this works more or less as advertised.

 

[goodfella]

Glad it's not dead.

No pop in open world game? Unlimited cars etc? Yes please.

 

[KScorp]

I believe that it's quite an amazing technology, and I hope that it will eventually be the epitome of graphics technology. The way they explained it makes sense, in that it works like a search engine, but it would also be pretty useless if its a pain to develop with.

 

[squidyj]

Heres to hoping this works more or less as advertised.

Long story short? It doesn't.

 

[MrHicks]

watched vid
still don't get it

atoms or someshit

 

[eosos]

Long story short? It doesn't.

Source?

 

[panda21]

There was an update posted on UD's facebook page that I thought you guys might like to read about. Check it out: http://www.facebook.com/topic.php?uid=105528666147262&topic=952

They seem to be getting support from the Australian government.

sounds like the Australian government bought themselves a lot of snake oil. gotta keep those snakes oiled up i guess.

 

[inky]

This can't be animated though, can it?

Seems useless if it can't be but for the most basic stuff like ground and rocks.

 

[XANDER CAGE]

Little Australian games company invents technology 100,000 times better

Fishing for sympathy, I see. And still barely any lighting (is dynamic lighting possible with voxels?).

And zero animation, too, which will be interesting to see if it ever comes to fruition.

 

[M3d10n]

Heres to hoping this works more or less as advertised.

Every graphics technique has a trade off. The difference between what's used and what is not depends on whether it's worth the trouble or not.

I seriously doubt someone can make something "100,000 times better" on the same hardware as everyone else without some significant trade off. In this case, the "unlimited" detail must be stored somewhere: the models themselves are made of "point cloud data", which can take lots of storage space. You really cannot really have unlimited detail, unless your level is made entirely of mandrelbot sets.

Fishing for sympathy, I see. And still barely any lighting (is dynamic lighting possible with voxels?).

And zero animation, too, which will be interesting to see if it ever comes to fruition.

It is possible by using a deferred renderer: render the voxels into normal/depth/color buffers and run lighting as a post process. However, this is a little trickier if things are actually being rendered on the CPU.

 

[Ravidrath]

The improvements are impressive, but he needs to stop pretending the game industry will have any interest in this without showing animation and collision.

Also, I'm guessing it takes a pretty monster machine to get that running, since there isn't any way to accelerate it yet. This means it's probably completely infeasible for consoles in the near to medium term, and since that's where most of the money in the industry is coming from at the moment, it's not particularly attractive.

 

[subversus]

I believe, haha

For real. Still, what about animation, destruction and collision?

 

[Frost_Ace]

Ok now I want a open world game developed with this technology :O

 

[Slappers Only]

Its not voxels not polygons and not raytracing its a search algorithm that picks only the points that you are looking at. You dont need a great graphics card too thats why Nvidia and ATI refused this technology because it would render them obsolete.

Nice.

 

[Shadow of the BEAST]

 

[maniac-kun]

Its not voxels.

 

[Mister_Bubbles]

"Let's take a look at this rock. If you'll permit me to say - it looks... raaather reaaal."

I love the voice-over guy.

 

[Shao Kahn Brewing a Stew]

"Let's take a look at this rock. If you'll permit me to say - it looks... raaather reaaal."

I love the voice-over guy.

My reply: No it doesn't look rather real, you moron! It's brown as shit! Fix the lighting, fix the colors, fix the atmosphere, eliminate the repetitiveness even if it's a tech demo. It's not just the polygons where millions of dollars go. It's the physics engine, the A.I., the gameplay mechanics, the voice-actors and sound designers, and more!

 

[balladofwindfishes]

if this technology is entirely static and cannot animate, maybe it could be combined with polygons to make it work.

Like characters and moving things use Polygons, while static background objects use this new tech.

That's assuming it doesn't need a 40gig text file to render a mushroom though

 

[Mister_Bubbles]

My reply: No it doesn't look rather real, you moron! It's brown as shit! Fix the lighting, fix the colors, fix the atmosphere, eliminate the repetitiveness even if it's a tech demo. It's not just the polygons where millions of dollars go. It's the physics engine, the A.I., the gameplay mechanics, the voice-actors and sound designers, and more!

Yup. Voice-over Guy does keep saying 'We're not game developers", though, so they're obviously aware of how much is missing from the demonstration.

 

[Aesius]

If this engine is so powerful and mind blowing and revolutionary, you'd think they would have hired better artists to make their tech demo instead of making excuses as to why it looks so crappy.

 

[Bollocks]

I don't know about this tech :/
In fact I think this is amateurish and just tries to generate a lot of buzz with absolutely no grounds of being the next best thing as it makes you believe it is.

First the videos are awful, some of the worst editing to show of the tech and I have yet to see collision detection and lighting with this thing. (Even after a year!!)
I think once they have implemented this the difference to a traditional polygon based 3D engine will be minimal if not more expensive.

Maybe more prominent studios will get inspired by this but I don't think that these guys will be famous for it/can pull it off

 

[MadOdorMachine]

Fishing for sympathy, I see. And still barely any lighting (is dynamic lighting possible with voxels?).

And zero animation, too, which will be interesting to see if it ever comes to fruition.

Well, I'm curious if this can be used in conjunction with traditional graphics. That way you can have your static environment use the voxels and anything that moved use polygons. I suppose that's just a stop gap for the time being until they can get it all using voxels. It would be nice to finally get rid of pop up, fade ins, and sub 60 fps framerates though.

One thing that I think is the most promising about this tech is that you can use it with a 3D scanner. That will make creating items in games a lot easier, faster and in much higher detail. Imagine game character's body being scanned in from a toy - any action figure. Now imagine a separate face being scanned in from a real life plaster of a model. All of their clothes and accessories could be scanned in. All of those little pouches, grenades and guns could be real independant objects on your character. Suddenly you can have a very photorealistic at a low budget. You still need to address physics and animation, but you could easily and quickly end up with a game modeled after something like this:

 

[Genesis Knight]

They've gotten basic animation working for Voxel graphics (there was a walking ogre on YouTube a while back), but I just don't get how collision is supposed to work with this setup.

 

[Chev]

Its not voxels.

It is voxels. Specifically, sparse voxel octrees. They've chosen to call them infinite detail atoms or whatever because that sounds fancier but from the demos and descriptions it is voxels.

They've gotten basic animation working for Voxel graphics (there was a walking ogre on YouTube a while back), but I just don't get how collision is supposed to work with this setup.

The same way as polygons, ie it doesn't, not at a reasonable speed. Instead you use bounding volumes that kinda match but are technically unrelated to the data used to draw on your screen, just positioned in the same spots, and you only go through the actual geometry for very specific things that just involve casting rays.

 

[dragonelite]

Not sure but if this is the future why doesn't CGI use it.
Still have to see animations from these guys.

 

[AllIsOneIsNone]

This is the future, mark my words. If graphics are going to mimic the real world, fractals are where were headed....

Edit: The most descriptive part is 7 minutes on. Basically the only part of an object that renders is the number of pixels you have on your screen. It picks up what is visible to you and renders it as pixels through complex algorithms. Would this be correct?

 

[sn00zer]

would it be possible to use this type of technology for graphics,but use traditional polygons for collision detection? Physics models dont need to be super high res (at least not this hi res) and are usually lower poly then the objects mesh they are attached to

 

[Chev]

Not sure but if this is the future why doesn't CGI use it.

It does, for the things that are more convenient to do with voxels (for example a lot of the edible stuff in Ratatouille, or liquids). But apart from that there's no reason to steer away from polygons when your geometry budget is high enough to have virtually infinite polygons anyway.

would it be possible to use this type of technology for graphics,but use traditional polygons for collision detection? Physics models dont need to be super high res (at least not this hi res) and are usually lower poly then the objects mesh they are attached to

As mentioned earlier yu don't use polygons for collision, because they kinda suck. You use bounding volumes, so how you draw your graphics doesn't matter anyway.

 

[adelante]

"What we mean is, if I wanna make Super Mario, dragon or unicorn, I can't go into the forest, tranquilize and laser scan them; they're not there."

I got a chuckle out of that line. Voice over guy is awesome....and so is the tech, can't wait to see it with different lighting.