I seem to recall a prior Epic technology had something to do with dynamically scaling models to fewer polygons automatically. I'm not sure if that is completely automatic, or works as I'm assuming, but I imagine there's probably some amazing synergy with this technology.
It's generally called "LOD", dynamically changing the level of detail of an object based on the distance from the camera.
It has been a pretty widely used optimization technique for a long time, and not specific to Unreal Engine (though I can't find which 3D renderer introduced it first).
The impression I got was that instead of dynamically changing between a few different pre-made models for an asset at a specific point, it could automatically scale a very high quality and poly count model down to less quality (that is, generate the lower quality model automatically from the high quality one), and on the fly or maybe JIT, but I'm not sure if that's accurate or if that's also old tech.
Edit: See your sibling comment, which came in just after I originally replied here.
I was in the back room of Epic Megagames in 97 when Tim Sweeney showed me his automatic LOD on the Skaarj model. It was pretty basic compared to the current tech, but I think remember him saying it didn't require lower poly meshes/maps to be generated by the animators/artists. I don't know if the LOD was pre-baked or not at the time, but given that we were running the first MMX chips it probably was.
> Nanite virtualized micropolygon geometry frees artists to create as much geometric detail as the eye can see. Nanite virtualized geometry means that film-quality source art comprising hundreds of millions or billions of polygons can be imported directly into Unreal Engine—anything from ZBrush sculpts to photogrammetry scans to CAD data—and it just works. Nanite geometry is streamed and scaled in real time so there are no more polygon count budgets, polygon memory budgets, or draw count budgets; there is no need to bake details to normal maps or manually author LODs; and there is no loss in quality.
Yes, I believe that's what I'm recalling, I remember there was a cave walk through demonstrated along with it.
I would think a technology that claims the more detail the better and that it will deal with worrying about optimizing what can be handled would work very well with a high quality scanning technology to make very detailed real-world based models.
Very well deserved. Having worked on 3D capture and reconstruction for the last 10 years I can say that CR is really way ahead of everyone else technologically and it was only a matter of time before someone would buy them out. I'm glad it was Epic.
Not OP, but possibly because they are working to upend steams monopoly on game distribution and offer an arguably better engine than the venture backed/pressured unity.
They're a monopoly the way apple is a monopoly. Maybe "anti-competitive" is a better word. There are other choices, but they have such a large share of the market "locked in" to their specific platforms that developers are forced to work with them or get out of the business. Good luck not having an iOS version of your app or not launching your game on steam, etc. With proper competitive pressure, the price of distributing a game or app should go down over time, not stay pegged at an arbitrary 30%.
> What makes you think that the Epic Store won't be exactly the same given half the chance?
They will, but it's pathologically impossible for either side to see that literally every option is elbow deep in corporate evil. Something has to be celebrated.
I'm not a fan of Steam simply because it is a form of DRM (and I hate the way library sharing works) but this is such a bs comparison it's hard to take seriously.
Steam is not remotely equivalent to the Apple Store in terms of anti-competitiveness. For one, I do not have to jailbreak my computer to install games outside of Steam and neither does doing so violate the TOS of Steam. In fact, steam literally lets you add non-steam games to it's library as well as let you access features such as the overlay and Steam input.
If anything, Epic Games is the one who is practising anti-competitive behaviour here by trying to force their way into the market by buying out developers over providing a better service.
If they ever took the dominant position in the market over Steam, I would be shocked if we don't end up in a situation that is worse off for both the customers and developers.
Also, "the laws of competitive pressure" only really work out well for the customers unless we have many truly viable competitors. What is far more likely to happen here is an oligopoly and we know how well those turn out.
I'm not familiar enough with the specifics of anything Valve/Steam might be doing, but I don't think they're necessarily anti-competitive (at least not in the same way Apple is). One way to think of this is that it doesn't take Steam being bad or anti-competitive to still have people benefit from alternatives and competition, so we can all be happy about more competition to Steam.
Apple, on the other hand, does many anti-competitive things to do with locking people into their whole ecosystem, such as restricting their software to run on their hardware (and their hardware to run with their software in the case of phones), and their store to run on their software, etc.
It's not so much that they have a walled garden, but that the garden is within a walled city and the city is within a locked down nation with mostly closed borders.
If Steam is anti-competitive (I don't know, perhaps they are) it may be best to note how, especially if you going to equate it to Apple's situation. Otherwise, people will likely see and focus on the ways they are different and/or think Apple's situation is similar, when it may not be.
An anti-competitive company doesn't make contributions to open source projects that are only tangentially related to their own product, nor does it let you integrate arbitrary software with their product.
Epic, on the other hand, doesn't allow any form of reviews on their store and routinely forces developers into a one-year exclusivity period before they're allowed to sell anywhere else. Tim Sweeney himself even said he doesn't think Epic is capable of offering anything new and can only compete through exclusivity deals [0]. If that's not a damning review of the company's competitive practices, I don't know what is.
Photogrammetry is what they say. It's a technique that involves stitching together loads of still photos to reconstruct a 3d surface. It's not real-time capture, it's for creating 3d models.
Can someone familiar with actual game development comment on how this works in conjunction with typical polycount concerns for games? My experience with environment scanning (notably limited to Matterport) has been that the models it spits out are... really bad in many ways, most notably the mesh geometry is an absolute mess as you would expect from estimating things based on scans. It seems like if you wanted to use scanned or photogrammetric-based models in a real game where performance matters, you'd end up having to basically recreate them. Hoping I'm wrong and someone can explain how this fits into a real-world asset pipeline.
The key is to have sophisticated software to "filter" down the content into reasonable geometry (rather than the grossly inefficient geometry you'd get out of typical photogrammetry) and then into LOD levels and very optimized streaming representations for that geometry. Anyway this is why unreal engine 5 comes together with current-generation console runtime SSD streaming, it's to facilitate this sort of streaming large quantities of geometry for virtual environments.
Obviously your result is only gonna be as good as your source imagery, but a large part of the magic is in how precisely the photogrammetry is able to solve and compute correct camera poses so as not to inject artifacts, combined with the automated generation of efficient meshes.
It seems like traditional asset pipelines never really had a need for crazy optimized LOD generation. There have been sophisticated mesh optimization tools, but usually it's only a small fraction of the manual effort to make some LODs compared to modeling the actual full detail model. It's usually done manually, but I don't know for sure, never worked in the industry. But you can imagine how this is a non-starter for when your full detail model came out of photogrammetry. I haven't been following it that closely but my take on it is that Unreal Engine 5 is a paradigm shift in that it will liberate you from even needing to consider levels of detail, much the same as you don't consider releasing memory in a garbage collected programming language. This is clearly possible once those concerns you state, "input is an absolute mess", "having to recreate models", become automated and reaches a level of quality that equals what a human can do.
I would expect in unreal engine 5 you will be able to import assets in traditional 3D formats (in addition, of course, to a bundle of photos) and it will recommend you provide it with the highest possible quality input. And you'll just be able to straight away instantiate hundreds of thousands of copies of it into your scene stretching into infinity and your framerate will be fine.
Disclaimer: My experience with Unreal engine is very limited and close the zero. However, I've worked in the Game dev industry for more than a decade. And have experience in mesh optimisation for various platforms.
I don't think the mesh can be directly used in games for any platform. However, the texture capturing with reference geometry is a great asset. With the scanned mesh, there are surface reconstruction algorithm that can use the original mesh as reference and create a mesh close to what a 3d modeller does for gaming. There are still a lot of manual work to optimise and make it ready for tessellation. I am not really sure, if there are any recent innovations in this field to automate this at the level of details needed for a game.
If there is, I can only imagine GTA inside a city with this level of details.
Usually a 3D artist would retopologize the model in a tool like ZBrush. For some models they are automatically "decimated" where the model is simplified using an algorithm. Decimating is not that common though in high end studios though.
With Unreal Engine 5 a feature called Nanite is supposed to allow high polycount models to be handled without preprocessing using some kind of runtime algorithm that simplified geometry.
These models also need all the lighting information extracted and turned into textures for use in a game engine(albedo, metalness, roughness etc). Most tools for automatically extracting the lighting information get it wrong most of the time leaving the artist to recreate it with a tool like Photoshop.
This could open up customizing games by adding your own scanned models of things (and head/body) in game.
Really hoping to also see some leaps/breakthroughs in photogrammetry tech i.e. significantly less pics with sharper results and faster rendering (not asking for much, eh).
This is the first moment that has made me pull my head out from under the many deep layers of Unity sand/code I’m in - if this is seriously a thing I will at the very least use this tool for this and then low-poly it and import into Unity.
35 comments
[ 3.7 ms ] story [ 86.6 ms ] threadIt has been a pretty widely used optimization technique for a long time, and not specific to Unreal Engine (though I can't find which 3D renderer introduced it first).
Edit: See your sibling comment, which came in just after I originally replied here.
> Nanite virtualized micropolygon geometry frees artists to create as much geometric detail as the eye can see. Nanite virtualized geometry means that film-quality source art comprising hundreds of millions or billions of polygons can be imported directly into Unreal Engine—anything from ZBrush sculpts to photogrammetry scans to CAD data—and it just works. Nanite geometry is streamed and scaled in real time so there are no more polygon count budgets, polygon memory budgets, or draw count budgets; there is no need to bake details to normal maps or manually author LODs; and there is no loss in quality.
https://www.unrealengine.com/en-US/blog/a-first-look-at-unre...
I would think a technology that claims the more detail the better and that it will deal with worrying about optimizing what can be handled would work very well with a high quality scanning technology to make very detailed real-world based models.
> What makes you think that the Epic Store won't be exactly the same given half the chance?
The laws of competitive pressure.
What makes you think that the Epic Store won't be exactly the same given half the chance?
Steam is not remotely equivalent to the Apple Store in terms of anti-competitiveness. For one, I do not have to jailbreak my computer to install games outside of Steam and neither does doing so violate the TOS of Steam. In fact, steam literally lets you add non-steam games to it's library as well as let you access features such as the overlay and Steam input.
If anything, Epic Games is the one who is practising anti-competitive behaviour here by trying to force their way into the market by buying out developers over providing a better service.
If they ever took the dominant position in the market over Steam, I would be shocked if we don't end up in a situation that is worse off for both the customers and developers.
Also, "the laws of competitive pressure" only really work out well for the customers unless we have many truly viable competitors. What is far more likely to happen here is an oligopoly and we know how well those turn out.
Apple, on the other hand, does many anti-competitive things to do with locking people into their whole ecosystem, such as restricting their software to run on their hardware (and their hardware to run with their software in the case of phones), and their store to run on their software, etc.
It's not so much that they have a walled garden, but that the garden is within a walled city and the city is within a locked down nation with mostly closed borders.
If Steam is anti-competitive (I don't know, perhaps they are) it may be best to note how, especially if you going to equate it to Apple's situation. Otherwise, people will likely see and focus on the ways they are different and/or think Apple's situation is similar, when it may not be.
Epic, on the other hand, doesn't allow any form of reviews on their store and routinely forces developers into a one-year exclusivity period before they're allowed to sell anywhere else. Tim Sweeney himself even said he doesn't think Epic is capable of offering anything new and can only compete through exclusivity deals [0]. If that's not a damning review of the company's competitive practices, I don't know what is.
[0] https://twitter.com/TimSweeneyEpic/status/114368034400623411...
Which is hugely exciting because this is phenomenal tech and in the hands of potentially everyone, could produce amazing stuff.
Perpetual enterprise local license dropped from 15k EUR to 3.5k USD
Lowest tier pay-per-input cloud credits dropped from 20 EUR for 8k megapixels to 10 USD for 14k megapixels
Obviously your result is only gonna be as good as your source imagery, but a large part of the magic is in how precisely the photogrammetry is able to solve and compute correct camera poses so as not to inject artifacts, combined with the automated generation of efficient meshes.
It seems like traditional asset pipelines never really had a need for crazy optimized LOD generation. There have been sophisticated mesh optimization tools, but usually it's only a small fraction of the manual effort to make some LODs compared to modeling the actual full detail model. It's usually done manually, but I don't know for sure, never worked in the industry. But you can imagine how this is a non-starter for when your full detail model came out of photogrammetry. I haven't been following it that closely but my take on it is that Unreal Engine 5 is a paradigm shift in that it will liberate you from even needing to consider levels of detail, much the same as you don't consider releasing memory in a garbage collected programming language. This is clearly possible once those concerns you state, "input is an absolute mess", "having to recreate models", become automated and reaches a level of quality that equals what a human can do.
I would expect in unreal engine 5 you will be able to import assets in traditional 3D formats (in addition, of course, to a bundle of photos) and it will recommend you provide it with the highest possible quality input. And you'll just be able to straight away instantiate hundreds of thousands of copies of it into your scene stretching into infinity and your framerate will be fine.
I don't think the mesh can be directly used in games for any platform. However, the texture capturing with reference geometry is a great asset. With the scanned mesh, there are surface reconstruction algorithm that can use the original mesh as reference and create a mesh close to what a 3d modeller does for gaming. There are still a lot of manual work to optimise and make it ready for tessellation. I am not really sure, if there are any recent innovations in this field to automate this at the level of details needed for a game. If there is, I can only imagine GTA inside a city with this level of details.
With Unreal Engine 5 a feature called Nanite is supposed to allow high polycount models to be handled without preprocessing using some kind of runtime algorithm that simplified geometry.
These models also need all the lighting information extracted and turned into textures for use in a game engine(albedo, metalness, roughness etc). Most tools for automatically extracting the lighting information get it wrong most of the time leaving the artist to recreate it with a tool like Photoshop.
In our eyes, we do not think of triangle or mesh : We think of geometry patterns. flat rectangular patterns, curved patterns, ect...
--> And this limits enormously the weight of the environment we visualise and memorise.
This is what I do at https://free-visit.netAnd this explains why my kitchen works on even a low end Smartphone : https://free-visit.net/fr/demo01/demo02
Really hoping to also see some leaps/breakthroughs in photogrammetry tech i.e. significantly less pics with sharper results and faster rendering (not asking for much, eh).
This is the first moment that has made me pull my head out from under the many deep layers of Unity sand/code I’m in - if this is seriously a thing I will at the very least use this tool for this and then low-poly it and import into Unity.
Wow, y’all. The future is here.