4d modeling sounds extra fancy, but it seems like the best, most obvious, way to create a simulation (treating the universe as solid, rather than hollow the way triangular modeling does). 4d modeling simply represents everything with 4-faced, 4-pointed tetrahedrons rather than triangles. It seems like this technology could be developed to create much more interactive and realistic environments much more easily. I don’t know though. What do you think?
If you want to know the future of any game engine, look at what NVidia + Microsoft does. Anything else for engine developers, when the matter is core tech, is irrelevant.
UE5 will only exist after DirectX12 is obsolete which may be over a decade from now.
And no, DX or OpenGL developers would never ditch technology based on triangles, that doesn’t make sense. Decades of R&D were spent to create an ocean of code and hardware optimized to render triangles.
There’s applications that support 4D modeling and a few 4D games. UE4 is not a modeling application, and 4D is too much of a niche for UE4 to support out of the box. We live in exist in 3 spacial dimensions, modeling in 4D wouldn’t make things more interactive or realistic in any way.
great feedback, but for the sake of argument I’m not sure 4d would be all that hard to implement, as it simply add a vertices, creating a solid, rather than hollow object. So I don’t think all that work would be trashed, it would simply be modified.
4d does not add a any spacial dimensions to the simulation, the simulation is still 3d, it simply adds a vertices to the classic triangle causing objects to be solid rather than hollow. This added dimension might greatly streamline workflows in some places. Thanks for the feedback.
There is nothing magical about it, Everything is made out of ‘tetrahedrons’ rather than triangles. Of course in the example video the guy’s thoughts of 4D is a lot more than tetrahedrons. He imagines an unseen direction the character can simply walk into that allows the char to bypass the laws the 3 dimensional space, just a game mechanic rather than any technical evolution.
In a lot of their materials they just call it Unreal Engine, there may be a version down the line that’s Unreal Engine 5.0 but it’s not going to be a completely different engine, it’s just the number they will use when they run out of numbers for 4. It’s not going to be like the switch between UE3 and UE4
Going with this. We’re just incrementing from 4.X.X, to the eventual 5.X.X. But it will still be the same engine or rather the same launcher.
The same way engines like idTech 3 have just been built upon for years, adding new features, but still keeping the same core.
No not yet, not for many, many, many years. The way game development will go in the next 10 - 15 years is the way it has been going, procedural.
Right now we’re half way between the ‘old’ way of modelling and the new, that’s making meshes then making the different textures and masks, importing them and making a shader/material. Applying it and then you have your model with a set size texture (eg. 2048x2048). This is soon to go away and the start of this can be seen today with tools like Substance Painter & Designer.
Texture authoring will improve and I believe the next ‘big’ improvement will be a standard model shared between all engines/rendering techniques, think of it like a “physically based rendering 2”. Along with this we’ll at some point get a special file format designed specifically for this use case with compression that is designed specifically for real time rendering meaning files will be a lot smaller. The next stage will most likely be the integration of these tools into the ‘top level’ software, so here we’re talking about 3dsmax, Maya, Unreal Engine and the like so the Substance software will no longer be a separate application but instead part of most higher level software. This will “unify” texturing, the ‘old’ methods however will still be supported for many, many, many years.
Modelling will be the next thing on the list to change, as of right now not much has changed in this industry. We still use 3dsmax/Blender/Maya/whatever and still have to follow the same set of rules as we did 10 or even 20 years ago, we can just do ‘more’ of it now than back then. Today we are JUST beginning to see some of the ways this will change with Houdini and SpeedTree. This will eventually “return” to the way it was back in the days with “level editors”. You’ll be able to create a wall brush (for those who don’t know that’s old terminology for a cube) like you could back then only now it’ll procedurally “form” itself into a detailed wall with appropriate edge detail and corners that aren’t “knife edge”. Basically a semi-procedural level editor that works not to disimilarly to the way it used to but that creates highly detailed and realistic levels from simple shapes. Texturing and modelling will unify somewhat.
Basically the ‘barriers’ will come down that currently exist between you and what you want to make, a lot of things we have to do today like UV Mapping, manual optimization and LOD generation as well as pixel based texture sizes will all go away and instead be replaced by “how detailed do you want this, it will use this much memory”. Advanced options will not go away of course there will be a tun to work with but it’ll be A LOT easier to make things.
As for coding, that, despite peoples hopes and predictions will NOT go away. Things like Blueprint will also not go away and will exist much the same as they do today but expand in availability and usability making it both easier for newbies to work with and stable/faster for the pros as well, meaning an artist can design an awesome system just as much as a coder and the choice whether to make it in a blueprint like system or in code will not matter beyond preference. (ie. blueprint will “generate” code like it does today but will convert to native code like c#/c++ as you design)
