wow Bruce Dell is really amazing I remember when I first saw you scan software with unlimited power I was like I can`t wait for this and now its finally here !!!
I got to try Holoverse today as a guest of Euclideon during their pre-opening testing phase. Let me tell you it's an excellent experience. Bruce is an artist and this is his creation. And like art, some will love it and others not. It's not made to impress "gamers", or compete with the top home console games. It's for families and groups of friends wanting a fun time out to experience something totally different. The current content (scenes) are made to be an accessible introduction to virtual reality. I'd surmise that their target market is the strong local tourism dollar, and their competition are other family entertainment venues. It has loads of potential for repeat customers by regularly introducing interesting new content and experiences. Well done, Euclideon!
Kids of today are born with devices, games and VR so I'm not sure what group they want to impress. Having a showcase of years old reused content and lack of a story doesn't convince that there is potential to create new content quick enough for a demanding audience.
A couple of weekends ago I took my whole family out to Holoverse again and this time tried out the first game episode, called False Eden. I thoroughly enjoyed it. There's lots of action and it was professionally voice acted. It has an original fantasy sci-fi story, which builds in emotion as the game unfolds. The boss fight at the end is a feast for the senses. Some of the coolest looking swords and fantasy weapons I've every seen in a video game and they are right there in your own hands. There's a new official video that better explains Holoverse: ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-F6MaZE9cU_c.html
You tried it out, can you see why Euclideon choose to limit their games on own-controlled hardware with custom 3D and not on any existing platform? Without that VR control, do you see it as a quality production as other modern games? You think they should continue this unique combo primarily and not doing common commercial software?
Imagine other developers optimizing this technology and further develop to see more possibilities with this new system like what voxels can do that polygons simply cannot..
I'm pretty sure he has to convert his little "atoms" (let's say the dirty word, voxels) into polygons to rasterize them to have decent, realtime speed. This tech is nice, but it's not revolutionary, it's just a voxel engine, and like all the voxel engines has a lot of limitations. Games today have really complex PBR shaders, dynamic lighting, reflections and so on. And you can use photogrammetry with polygons as well... Unless someone decides to make a GPU that's very fast to process and render voxels, polygons are still the de facto standard and are still going to be.
+TheWaveringRadiant It isn't rasterized. If it was, it would be considerably slower. It's actually probably ray traced or raycasted sparse voxel octrees.
+Void Chicken Ray tracing is slower than rasterization. It would explain why their video already showcases a lot of skipped frames, even though the RU-vid feed is at 30fps. Furthermore, their poor latency is even worse of an issue for VR, as it quickly makes you feel sick. I have a feeling that was why they chose wall projections instead of headsets. That way, you can turn your head without the system having to do more processing, but the problem remains when you are moving relative to the world. Additionally, anything that is animated is awfully hard to render in point clouds, and yet again it shows in the video. All the animations are super-short and repetitive, and I'm not sure they can rely on interpolation. The scenes are really static. The fact that they don't even mention the visible cons of their approach, as usual, gets old. Too bad, it probably has really valid applications.
+Oculus Rift The clips shown during the Innovation Summit speech above are the only official videos to date showing Holoverse content. On their updated Facebook page (link above) there are some new detailed promo pics and a brief video walking past some of the rooms. I would expect Euclideon's own RU-vid channel, EuclideonOfficial, to publish any further official videos, in preparation for the Holoverse opening.
I'm a little disappointed. I don't expect it to be perfect right away, but it seems that when it comes to animation, they haven't achieved "unlimited power" yet, or even normal power. I'm actually a bit confused on this. The evil bone-dragon thingy and the swimming turtles look alright, but then the charging elephants and a few other animations looked terrible. Maybe the reason the turtles and the dragon looked impressive was because they didn't need to make contact with the ground. A physics problem perhaps? The elephants looked like there were only 2-3 frames of animation too. I'm still excited to see if this goes anywhere, but I think it's going to be a few years yet, maybe more.
+Khechari ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-4pFAgf3PUK8.html He states that this was done by 3 people in 2.5 months. That's actually impressive for the resources they used. Imagine a whole team, or more time! If there are other gaming companies using this tech, it's up to them to display their work, they're not obligated to do so, neither is Euclideon for that matter, but they're putting themselves out there, despite the criticisms - do you get what I'm saying? From the looks of it, they're only just beginning!
+blackpanther6389 Why does a software need to be marketed with unlimited power, so much better than the rest and create controversy, shouldn't software speak for itself and create respect? Euclideon had $2 million, 30+ people and still struggling. Have you heard of the classical thing of saying one thing but doing something else?
I would imagine that there are a lot of ideas out there that get shot down, or ignored despite being wonderful ideas. It's not software, but I'm part of an organization with pretty radical ideas about social redesign of civilization. I think the ideas are good and makes sense, but it's still taking quite some time for people to see how valuable those ideas are, and the development time of these ideas took 50+ years to make. Lol, you're probably thinking that sounds crazy, which you have every right to think so. But I'm still not convinced that they're playing con games. I'm sorry, but you'll have to define "struggling". When did Bruce say that he was "struggling" I've only heard you insinuate that Euclideon is "struggling", do you have a source? To be honest, I'm not entirely sure what it would take to convince me otherwise. My experience with graphics and computer programming (I'm going to school for programming now) leads me to believe that this is legit, so at this point we're really just arguing for the sake of arguing because I think our viewpoints are pretty much sold on what we've been believing since Euclideon first surfaced, but if you would be willing to provide a source for Euclideon "struggling" I would appreciate that.
As somebody who has been modifying games since I was 13 years old (younger than that if you count screwing with build engine), I can see the potential in this technology but there remains to be answered a couple of very specific questions. In general I have a pretty good understanding of a very broad range of topics involved in this tech. 1) Framerate. In this video the frame rate appears to be very low, is that an artifact perhaps of the method of demonstration... i.e. projector? Or is this an artifact of tick time of the graphics engine? 2) Latency of input during interaction. It actually seems to be pretty good, what are the limits for latency of input and the response from display output. Less than 30 milliseconds? Pretty good demo really, but what you really need to do, is play around with a bit of netcode and test what this would be like if you networked "rooms". Sure its already being done for military applications. Probably all hush hush. The reason why not many people take this tech seriously is because it's missing a lot of the wow factor that could be achieved by say 4 people playing a simple laser tag game against each other at 60+ fps, now that, would not just get people talking if it was done sucessfully, but it would gain you SIGNIFICANT enthusiast investment from the gaming community. My 2 cents.
+David McCrae the thing is, creating all these points isnt whats causing the problem, its calculating the shadows of them.(atleast that what i would think, unless they solved this problem somehow.
With a point-cloud you do ray-tracing, and with ray-tracing shadows and realistic lighting are basically free. You're already tracking the proper angles to the viewer, which is exactly what you need to do the light and shadows realistically. The problems with point clouds and polygons are basically inverted. Shadows are expensive and actually essentially impossible to do true-to-life for polygons and rasters, but surfaces are cheap. Conversely, shadows are cheap for ray-tracing, but surfaces are very expensive. The reason rasters and polygons have basically won up until now is simply the fact that you can get away with bad shadows if you have great surfaces, but you can't get away with bad surfaces even though you have amazing shadows. Ray tracing has always been the holy grail of 3d game graphics, primarily because you don't need tricks to imitate real lighting and shadows. Ray-tracing is basically how light actually works in the real world. The problem has been, in order to make a good looking game in, say, 1080p, you need to trace the path from the viewer's eye level to 2 million points 60 times per second, and calculate the proper light and color for each point in each frame. Right now the only way to make it work is to blow those points up so that the computer only needs to manage a few thousand points. What Euclideon is basically saying is that they've figured out how to track as many points as your screen can display in a reasonably quick fashion. It doesn't look like they're up in the millions at 60fps, but they do look like they're much closer than anybody else. The other major issue with ray tracing, which was implied in the demo, is that current GPU hardware does absolutely nothing for it. So the software is going to run on CPU only, which is also why they can run it on laptops. Laptops get CPU's in the same neighborhood as desktops, so it's not too surprising they can run nearly as well there as a desktop. But realistically you're going to want some dedicated hardware for processing this stuff to get any kind of real, reasonable response out of it.
I'm pretty excited to see that their actually making computer games but what I don't understand is why would they make a fictional game like that if they have the ability to make a game look extremely realistic with unlimited detail?
+TheFlash Stickman Yeah I don't understand either. They promised unlimited game graphics many years ago. And then they sold geospatial products. And then disappeared. And then promised games. And now it's VR rooms in Australia. Who knows, maybe the tech for a detailed static scan is not as same as a realistic game?
+EuclideonFollow This is the story: Bruce marketed unlimited detail 2003 at a game conference with no success. 2010 he started Euclideon that introduced geospatial products. 2016 they are into VR but still has zero track record in gaming industry. For me is that a hoax, Bruce claimed to revolutionary game industry 2010 but never finished anything or got any credit in gaming industry. What's hard to understand about that?
+EuclideonFollow Went geospatial because they were asked to? So all publicity about revolutionizing polygon games and game fans weren't important anymore because another industry had the money? How come that years have passed, and there is no blogs, news, company reputation in the world about this tech?
+EuclideonFollow Interesting that you mention Meixner which is a geospatial company, because that has nothing to do with games. You have still little clue what the games are about and satisfied with seeing some choppy retro game logics thrown together. You don't want to call it a hoax, its just Euclideon that can change their mind anytime and you forgot what they said in first place.
Jeffery Wells To create believable 3D graphics you need to shade the pixels on the screen properly. Objects only become 3 dimensional if they are exposed to shades. As you would draw a face on paper and search for the gradients of color in order to let the face appear 3D. If graphics were only about finding the silhouette of an object, things would appear 2D, literally. Your object could consist of trillions of primitives (voxels /polys), you wouldn#t notice a thing of them if there are no shadow gradients. You would only see a silhouette containing a texture. That is why all this infinite detail becomes unperceivable in those Holoverse demos. They probably have trillions of voxels in there. So what? Why does it not look like that? Because perception of 3D objects arise from shades. Physically based shades. Each ray of light that make up the color of a pixel on the screen has a history in time and space, hitting multiple primitives along its path transporting its reflected color from one primitive to another. Every great artist in the past first studied the very nature of light as they started to paint.
"40 million polygons.. that's the count you would use in your whole level"... well CryTek threw that idea in the trash with Crysis 2, not to mention the engine will render full detailed water under every level which is wasted compute power... this is just bullshit.
It`s 2016 now. Still noone ever saw a playable demo download from these scammers. All they did is prerender some footage and called it a revolution. Hope they catch them soon.
This is only good for static scenes, the major issues are with manipulating or moving the point cloud data for motion. Each laser scanned object has far to much data to work with currently. Once the cloud data is compressed small enough and computer hardware can render 20 million polygons per square meter, then you can use this implementation.
Question. All the objects seen in the created "Real Time" world are scanned in. However, in games they don't scan in objects artists draw them and create animations. Will this tech be compatible with such animations?
there is only rumors about this because noone ever provided a proof. But as they said they can scan in all standart 3d formats you could basicaly import every frame of your 3d model animation into a "atom-dataset" and then switch the set you read your atom data from every frame while rendering. But i dont know if this will work, i just came up with this from all the informations i got about this engine
Banjo So it needs massive amounts of memory? Games tend to work with bones which is an order of magnitude more efficient. Also I'm not sure about the 'infinite' claims, there must be a finite point cloud data set, if I zoom the camera very closely to an object it will eventually run out of data points and have to fall back to some type of interpolation.
SerBallister bones only work with vector datasets not with point clouds, so i dont know why you came up with this? its pretty obvoius that the infity claims are fals, you cant finish a single raycast in infinite datasets
Banjo I think you misunderstood my post. I was referring to traditional polygon engines and how memory efficient their animation systems are. Yeah it's clearly over-exaggerating its capabilities, the point data needs memory so unless it's procedurally generated it's not going to be infinite
Scott Metzger The lack of a demonstrated animation system hints that the point cloud has been optimised off line for faster searching (probably oct-tre, kd-tree or some other subdivision method), I know just casting a ray into an unsorted point cloud is going to be expensive with lots of points. Not sure LODing a point cloud is that trivial for real time either, wouldn't you need to be removing points that have the least neighbours ? Meaning you need to check every point against all potential neighbours. Say I have an original cloud of 500,000 points, but I want to draw it in 50 points, reducing 500,000 -> 50 isn't straight forward. You could preprocess the point cloud into a tree to help you decimate it quickly I suppose. Infant thinking about it, you could make a decent system using a hierarchal point cloud. Start with a very low resolution representation, instead of points imagine large spheres that are touching so there are no gaps, you can test a ray against these to see which sphere you hit. Then when you have your hit sphere, you repeat the process for the next LOD level points inside that sphere.. ray vs sphere tests are super fast and you only need about 8 or so per ray to be able to traverse an insane number of points quickly...
one big question is how this will work with the game physics. imagine physics applied with these millions of atoms that will blow up most powerful pc in the world. :)
I can think of an easy work around for that, to be honest not much would have to change because you would have a separate collision mesh that determined how each mesh would interact with the world and other objects. That way it the computer wouldn't be trying to calculate all of these little atoms colliding, just these simple invisible collision meshes encase each object.
It depends on what you need. If you would need atom-precise collision that would probably, indeed, be pretty taxing. However, what would be a much better way is to do it as it is often done today (and in the past) by using collision meshes. That is, separate, simpler models that are not visible to the player that are used for doing physics related things. A 2D example of this is how in danmaku (bullet hell) games your collision box is actually a lot smaller than your character sprite is.
***** "NO, it will not work for game engines" Just this statement alone makes it clear that you don't really know what you're talking about. No, this tech is not going to be implemented in an existing engine since they feature their own rendering tech. But the statement that this is not going to work for game engines is a really uninformed one since it suggests graphical systems are not part of an engine and that you can just plug any rendering tech into any engine, which of course is a load of baloney. Right now it is unlikely for this rendering tech to be implemented in a game engine, but you can't say that this "isn't going to work with game engines". A game engine would need to be developed around this rendering tech. It doesn't work the other way around. You can't make a graphical technique "work" with an engine. The engine encompasses the rendering tech. "That is what critics were saying from the beginning." Nice source. Talk about not contributing anything at all. I'm not saying you're wrong. I'm just saying that you haven't said a single useful word.
***** Aha, so you do know. Well then why do you try so hard to make an ass out of yourself? You can't just say "NO, it will not work for game engines. That is what critics were saying from the beginning" and leave it at that. That adds nothing to the conversation and nobody gains any knowledge. In fact, I don't have to do anything. You made that statement. Now it is your turn to explain to me what makes it so difficult. It is up to you to come with the arguments and the explanations. It is not up to me to explain Your statement. So I expect either a lengthy post explaining everything, which would be suitable, or I expect silence or another overly negative response, which would be a cop-out. Don't get me wrong, by the way: I don't expect this technique to make it into the gaming industry any time soon. I think the current techniques are more flexible, better supported and are moving forward at a much faster pace than this technology is.
what is stupid is that Euclideon should have put their island online to prove thier point, when u come out with unbelievable technology of course people are going to jump at, and against it, it's basically like magic
YAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA!!!!!!!!!!!!!!!!!!!!! THE HOLY GRAIL!!!!!!!!!!!!!! ULTRA HD VIRTUAL REALITY ON A MID RANGE PC OR INTEL SMARTPHONE!!!!!!!!!
they have released a new video. A geo spatial entreprise scanned in a 1000 mile long highway (Australia) with photogrammetry. Its a 2 TB big model. Euclideon made a programm thats called Geoverse. With that geospatial entreprises can see their humongous nodels in real time. So yeah it must be real, they are already selling their technology. :) In another video a student (Aiden) interviews them. The CEO promised they will release the best animation the world has ever seen soon.
People think that this will render Nvidia obsolete, but this isnt the case. Graphics cards dont JUST render geometric shapes. It also renders particles, physics, dynamic lighting, etc. I think that this type of technology with just give games more room for more realistic water, light, physics, destruction, and so forth.
