well the real beauty of this technology is that you can transfer an object made of polygons into an object made up of the unlimited points this engine uses. so, theoretically bethesda could make tesv with the regular polygon system, then update it to use the point cloud data system rather easily.

However I dont think bethesda should do that because this technology is still very young and almost seems too good to be true. but if its capable of what they say it is, then this is the future.

Actually, Oblivion's Gamebryo engine renders all models in the area - whether you're looking at them or not. This is precisely why there are separate interior/exterior cells because otherwise the engine would try to render the entire game at once. The word 'render' means to calculate and draw the polygons as dictated by the game models, not necessarily just what is in view in the game world. Some engines are optimized to not draw (or render) what is not in view, but not all of them do this - such as with Gamebryo.


For some reason I thought games like Half-Life 2 already used pre-rendering to some extent to optimize the game. I could be grossly wrong here, but I am under that impression.

Oh, my mistake then, that would be a nice performance boost for a TES:V though, didnt realize they didnt do this.

I still dont think that this point system is the next best thing, it would make every 3d program out there incompatible with this format of model for one... And whats this about their algorithm? That means its something the company owns, that if any other company wanted to use this they would have to go through this one company for their one algorithm. Sounds like a monopoly, sounds expensive. Sounds like Nvidias PhysX thing. Not everyone is going to jump on board, and its generally just a advertisemant/w.e to push people to buy nvidia cards.

What kind of confuses me though is that what do four points make up? A plane. When you lay a texture over an object, no matter if its made of points or shapes, there will always be an empty space between the two points somewhere no matter how small. Also, wouldnt it be waste of time and performance to render an infinite amout of points on the side of a wall when you could just have a few points to make up that plane?

I dont know why it is myself, but this thing sounds so much like BS, i just really dislike this whole idea for some reason.

Instead of unlimited geometry I'd like something "revolutionary" akin to the vast forests of Oblivion. Imagine Skyrim covered in deep snow, where prey and man alike leave their tracks in pure white snow....

It would have to be something like "Unlimited multi-variable calculus to re-create every concept of atomic physics in TES"





Umcrecapt for short

Actually, given the explanation of how the system works it would need very little resources. Just about any dirt cheap modern computer could run this system, the amount of data sent through the system at any given time interval would be the equivalent of a photograph whose resolution is equal to your monitor's resolution. The system seems very efficient on that front.





Actually in later, more recent videos, they address how artist will eventually, although the software isn't yet in place to, be able to take polygon rendered object and convert it into point art. The idea being that artist start to render the objects on the maximum number of polygons, and then port them into their software package. You might ask yourself "If they can render the objects with more polygons in the first place, why not just do that and avoid using this system?" They don't do that because when you are processing an entire game cell full of objects, all of which are rendered so finely, you would need a very very powerful computer to handle all that data. This system doesn't necessarily improve the ability of the computer to render graphics, because people would still render them as polygons first. What it allows the computer to do is run higher level graphics in real time by using fewer system resources.

A good anology might be to imagine you were having a dinner party and you want to have an uncle over, for the sake of the argument you have two uncles and can only invite one. Your first uncle is hilarious, but he eats EVERYTHING, your second uncle is also hilarious, but he will not eat ALL the food, so your other guest can eat too. Well the first uncle represents polygon graphics, the second represents this system of graphical rendering, how funny they are represents how good they render graphics, and how much they eat represents system resources in the computer. Both uncles are just as funny, just you wouldn't have to pay for a catering service if yo go with the second, less resource consumptive uncle.



Yes and no. More data would be pushed through your processor, but an amount which is not even comparable to what your GPU would have to process in current graphics engines. As for being similar to what is already done, it is not very alike at all. Your computer currently does only allow you see the same end number of pixels, this is true. However that is the end product of the process it is running. With this system your computer searches for information relevant to each pixel on screen. Normally the computer runs through a process which retrieves information (both relevant and not relevant), reads it, edits it, and readies it to be placed on screen. You cut out the entire process of editing the data by including that with the process of retrieving it, it doesn't need to ready the data to be output by cutting it into pixels and outputting it because it only asked for data relevant to those pixels already, and then you also throw out the need to retrieve unnecessary by first determining what data is necessary.

In simple terms current graphics would be similar to taking an anolog picture and cutting it down into pixels, there is no true number of pixels in the picture as it is anolog. So you must first process far more data than you need because you have to process the ENTIRE picture and create the information for the pixels from scratch.

With this system its the opposite. Your computer knows precisely how many pixels it is looking to place in the proverbial digital copy, so rather than process the entire picture, it would just process information relevant to the pixels it is requesting.

Also the demo only svcked because like the narrator explained, it was programmer artwork not a professional artist art. As a programmer I can tell you most programmers are not artistically skilled. Quite frankly I was surprised it looked as good as it did.



All software consist of algorithms, games not being an exception. An algorithm is simply a step by step approach to resolve a task, so all they mean by their algorithms is their software, in this case: unlimited detail.

Secondly, how textures would be rendered in this system would work quite differently than polygon based graphics. Instead of having flat surfaces which have textures laid over them, individual points are each assigned a color, when viewed from a distance they make a texture and/or image. Its similar to pointillism. Plus there would be "no empty space", unless the artist intended it. Put simply, each point represents an individual pixel on your screen, and while there is distance between the pixels on your screen you cannot detect this distance because the pixels are so small and glow with such intensity. As you stare at the screen right now surely you don't notice the gaps between the pixels of your monitor. Even if you do, this problem is not resolved with the polygon rendering, and it i a hardware problem rather than a software problem.

I'm sure it would look pretty enough for me if they released it tomorrow. I'm more concerned with how TES 5 plays than how it looks.

@ blade_jenko, thanks for replying to all my questions and speculations It cleared up some thought about it.

About the models...more detail in each model would mean either longer release dates or less models for a game. As it stands a model creator isnt going to spend a week on a single rock to give it the most detail it can handle (assuming hes creating the model as a 3d object and having their product convert it?), hes got other things to work on to meet deadlines. Current technologies can make models pretty detailed, but devs arnt going to put max detail in every little thing, they simply do not have the time for it, not is it really needed. Its not just about because computer systems cant handle it.

What this would boil down to imo is the difference in how much time to detail an artist can make compared to the current system. Maybe the current technology really does limit artists, if so then maybe this new technology really is needed.

We'll see where they take this. If it catches on or flops

Just one more thing about its performance benefit, companies that need performance boosts should firstly do what some game engines do. Not render things not in view! If they really want to take a step in this direction, wouldnt they take a step at a time until they really honestly need to make the expensive jump in technologies?

Im probably out of my bounds of knowledge now so this will prob be the last i comment on this, can only speculate and form opinions so much before you eventually just need to wait and see. (tes:v anyone? lol)

Actually you just reminded me I forgot to mention what was the most important part of my original post. Do I think bethesda will use this? Nope. Why would bethesda waste massive amounts of money to run its graphics engine on software which is neither completed, or being marketed to devs yet. That being said, I think the rendering software is genius and will probably be the next big thing in gaming and graphics development.

As for developing that much extra detail, its not much harder in some ways, but all in all it would definitely take more time to do a single item when modeling. But at the same time you would only need to create one model for all distances. As it is now artist have to create several models for the same items, because as player's move towards an object they replace the object with increasingly more complex models, but when a player is far from an object they replace it with a simpler less "pretty" object to save computing power. Thats how they get draw rates (how far you can move from something and still have the object visible, like when you move too far from an npc in oblivion that it doesn't show them until you move closer again) in games so high now.

it's going to be grat when ppl start to use this.

but seems like most are set in there ways so may take some time before we see this a lot.

and i love the idea of work more on one model then work on like 5 of the same models .

This technology is interesting, too bad it's experimental and derives so heavily from the industry standards that it would demolish the assumed at least 1-2 years that Bethesda has already invested into TES V. Most developers and most modders are used to polygons and the like, anyway - while this is pretty, I'd really not want to switch over to a new standard anytime soon. It'd just be a major pain for everyone, and slow down Bethesda even moreso. I think Bethesda is pretty much on the marker, although they do need to learn about this little thing called http://www.umbrasoftware.com/.

Am I the only one who thinks that this level of detail is unecessary, anyway? Both Morrowind and Oblivion were still pretty good-looking for their times, and I certainly prefer Morrowind's surreal grungey detail and Oblivion's stylized lighted plasticky bloom-spammed detail over the Crysis method. That is to say, I may be interested in NPC detail or lighting, but I could care less about whether or not a game can render the individual veins of twenty-thousand 3D modelled leaves on my computer http://www.umbrasoftware.com/.

As interesting as this is, polygons still reign supreme in this day and age - something may become of this line http://www.umbrasoftware.com/ in a few more years, but as it stands, at least shelve it for TES VI or even VII. Someone should also send a letter to the people who put together this showcase, and tell them that they could have created a tech demo that is a bit less strainful on the eyes - I don't know what I was seeing, it was just some dissonant rainbow of Mesoamerican structures and goofy foliage dangling from every other inch of it.

Pretty much.

The fact is, Beth has already started developing TES:V, that's pretty much a known fact. It's not like they would trash 2 years work to do this. It's really not "that" great.
And anyway, gameplay > graphics.

The "technology" is not viable. At best it would work for static pre-rendered environments. The pre-rendering is required to generate so called search trees. Generating those is slow, but using them for display would be fast. However, as soon as something moves real-time the search tree of the moving object needs to be re-inserted in the system. That is a slow process. For simple objects this may work, but imagine a forest on a windy day, or the sea with clouds at sunset or a collapsing building on fire. In theory you could even pre-render all of that, if you would have near unlimited disc space and extremely fast disc drives. So in practice re-insertion would be the only option left. And there would be too many insertions for the samples given and the system would come to a shrieking halt.

you may want to post there other like 3 parts they have for this may help ppl get the idea be hind how this works.

As cool as this is, I would much rather that time be invested in adding depth to the story and gameplay instead of spent working on graphical perfection. Oblivion certainly wasn't photo realistic, but I still had no trouble finding myself lost in its world. Adding more realism to the movement of the world and creatures/characters in it is a much more effective path to immersion. At least to my mind, strange, robotic animations (made painfully obvious in Oblivion's third person view) which make a character appear detached from the world and anything but alive are a much stronger immersion breaker than the occasional low resolution texture. So as impressive as this technology is, I don't feel it is really the right direction for TES V.

Ugh, I hate the way that looks. It's so lumpy! And it has this strange sort of uncanny valley feel to it.
It's pretty much just another interpretation of the voxel concept anyway, so I guarantee that it'll die down within a year or two.

The thing is, this isn't worth delaying TES:V for. Period. I really could care less about the graphics, and any true TES fan will agree with me. Delaying a game for "LOLRLLYGUD GRAPXIX" = not worth it.
We want story. Depth. Epic questlines stretching across deep valleys and canyons, filled with monsters.
Give me TES:V, and I don't care if it has MORROWINDS graphics, so long as it has an insanely amazing MQ and guild quest and great houses to join. I'll be happy with that.

That's not trough for Bethesda though, instead of Model scaling they just add a lot of fog.

While I wouldn't go so far as to call BS, I doubt 16 months is going to be enough time to make this practical. And even if it is a good estimate, I wouldn't want them to wait that long to resume TESV development. It's a very interesting concept, but not delay-worthy, to me anyways.

In response to the argument that it would take the developers too long to make everything so damn high-poly:

As far as I know, developers nowadays usually make high-poly models first, and then look what they can take away and have it still look good. At least, I've often heard modelers from the modding community say so; additionally, I've heard that many normal/parallax maps are actually based on those high-poly models, so that they need to do them anyway. (They make a high-poly model, then the low-poly model, and then they let a program make the normal/parallax maps depending on what is missing in the low-poly model.)

Other than that, nobody would be forcing developers to make high-poly models. It is just that there would be no difference in performance between high-poly and lower-poly, so that they should do it unless they want their game to look worse than it could with the same performance.

I wonder if this can be used in correlation with polygons, that way this might suffice for static environments, while polygon tech takes care of animation.

As I said before, I'm 99% sure if you ask any hardcoe TES fan if this is worth delaying TES:V, they will say no. Graphics aren't what makes a great RPG. This isn't worth delaying anything for. Maybe when it's more practical, and maybe on TES:VI? Sure, I don't care. But not worth delaying a current in progress game that's already being worked on.
And anyway, I honestly think it looks atrocious. Everything in that video looked like it was a cutscene from a N64 game.

Who are you talking to? Nobody says that Bethesda should delay their work on TES V (except for the title of the thread, which I didn't take too seriously).

By the way, yes the video doesn't look very good, but they're probably just a bunch of programmers and simply don't have the artistic skills of game developers. What they're advertising is a new, speedy way of rendering a scene that allows you to use as many polygons as you want without compromising your game's performance. Which is pretty cool.

Well, they are claiming it but haven't offered credible proof. There is no such thing as unlimited voxels (their technology is based on volumes, not polygons), because there isn't unlimited storage, and there aren't unlimited instruction cycles. In particular, they haven't made any kind of proof that they can compute any but the most elementary animation of small parts of a scene in real time. The storage and power to render those things has to come from somewhere, and they can't just pull a rabbit out of an empty hat and say "here it is".

In the video they say they can convert polygons to the new system, so it wouldn't truly slow progress and it won't make the models they have already made useless. And what's the point with the 'the video looked atrocious', it's not how everything looks using that technology, it's just there example of mdoels which aren't made by professionals. If they would make Oblivion or Morrowind with that technology and not editing anything it would look the same but it would just lower the system requirements.

Off course no I don't want TES V to be delayed for this but I think it is (if it's true) a great piece of technology.
At the moment many new games are focusing on graphics rather than gameplay, off course this isn't going to fix it. It might even make it worse in some cases, but if the graphics of TES V (or any game) require less system performance they can add bigger scripts and far more heavy/advanced AI right? Things like scripts and AI coding slows down the game, but if the performance hit of graphics is removed a lot more power can be used for things that DO make the game better. I don't need TES V to have 100% best graphics, but if the video is true then this technology would made large numbers of NPC's, open worlds, great AI possible because it enables the game to use all your PC's power for stuff like that.

I don't know much about this, so I could be totally wrong.

Anyway, I hope the thing they say in the video is true because it has many advantages and in a few years most of it's disadvantages might be gone