4 ms·

Can someone familiar with actual game development comment on how this works in conjunction with typical polycount concerns for games? My experience with enviro

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.

unphased 6y ago
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.
stubish 6y ago
Just an onlooker with no actual experience, but https://www.reviewgeek.com/42502/epic-games-announces-unreal-engine-5-heres-why-its-seriously-impressive/ https://www.reviewgeek.com/42502/epic-games-announces-unreal... is mostly claiming that polycount is no longer an issue, handled transparently by the Unreal engine.
codetiger 6y ago
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.
t8y 6y ago
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.
- lostgame 6y ago
  I agree. ‘Decimation’ seems to be the industry standard for referring to the reduction of polycount in models. :)
tmilard 6y ago
Only thinking about mesh geometry count is, my belief, a dead end road... 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. - Simple as that - This is what I do at https://free-visit.net https://free-visit.net And this explains why my kitchen works on even a low end Smartphone : https://free-visit.net/fr/demo01/demo02 https://free-visit.net/fr/demo01/demo02

unphased 6y ago

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.

stubish 6y ago

Just an onlooker with no actual experience, but https://www.reviewgeek.com/42502/epic-games-announces-unreal-engine-5-heres-why-its-seriously-impressive/ https://www.reviewgeek.com/42502/epic-games-announces-unreal... is mostly claiming that polycount is no longer an issue, handled transparently by the Unreal engine.

codetiger 6y ago

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.

t8y 6y ago

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.

lostgame 6y ago
I agree. ‘Decimation’ seems to be the industry standard for referring to the reduction of polycount in models. :)

tmilard 6y ago

Only thinking about mesh geometry count is, my belief, a dead end road... 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. - Simple as that - This is what I do at https://free-visit.net https://free-visit.net And this explains why my kitchen works on even a low end Smartphone : https://free-visit.net/fr/demo01/demo02 https://free-visit.net/fr/demo01/demo02