So what are your predictions for the future of 3d-graphics? Both realtime and offline?
I think we might approach 16k, as it is the resolution of the human eye, and will be sufficient for VR.
Raytracing in games is already here, so it might just become used everywhere, so offline renderers might start losing its cutting-edge position.
The industry desperatrly in need of affordable mocap for enthusiasts. The advent might come fro AI.
The whole idea of AI-aided solutions is both breathtaking and terrifying at the same time. It might outlplay us in any field. But who kknows, maybe humans will be able to create their own worlds just from imagination, or by words construction input.
Though graphivs evolves, therre must be a point of diminishing reeturn, when the progress will slow down due to being good enough for the human eye. We are still way behind this point I think. It might be 20 or more years ahred, or I’m delustional and it’s 100 years ahead.
The future of 3d-graphics
IMO, tools are more important and need more improving, not graphics.
Pretty visuals are nice, but they can only go so far. Especially when they actually begin to hold back creativity and rely on increasing manpower to get better results.
Look back at the PS4/XBO generation. The shift to PBR tech already made games look realistic yet a lot of titles started to ship broken at launch and required patches just to fix them. Or how about the fact development cost exploded, leading Publishers to use more nickel and dime tactics while also moving away from funding risky projects like they use to do all the time in the PS2 days.
I actually welcome AI if it can actually help remedy these problems. For example, I’ve said many times on CGsociety that I find UV Mapping to be the most boring and useless part of the 3D pipeline. If I can tell a robot to unwrap an entire scene in less than a second, that would be god send.
I also can’t say that slapping base textures is all that creative. I think neural network would be perfect for creating hi-res textures. Basically, a lot is not creative at all in this field. It’s both good and bad.
It’s good, because craftsmen still can make a buck by knowing the tools. And bad, as it’s a manual labor, which has little creativity.
What about retopology?
A lot can be automated and proceduralized. AI can create textured 3d models, based on photos. And we have plethora of them on the internet.
Generating humans? AI can use social media for generating it.
It could also xtract movements from videos.
Has anyone from the medical profession ever stated that humans "see"in 16K?
Not likey, because Human eyesight is no pixel based Mate.
Definitely you are right, and I stand corrected. I meant our eye won’t perceive eny noticeable pixelation after that digital verge. And it might be ample enough for the human eye to plummet into the 3d-realm.
I predict…
-VR, as in VR tied to goggles or your phone will die off like a fad or middle-ware, like pagers did with cell phones, when working holographic displays become more mainstream. Once we do not have to wear VR goggles or pull up our phone for AR, things will take off more.
-16k would be horrible, working on 8k projects is enough thanks. The amount of data streaming is already enough at 8k.
-machine learning…or AI, will just continue to aid on mundane tasks, and that is a good thing unless you are say a roto artist, or super specialized in say creating or retopo only.
I prefer better tools, output specs for high level projects are already high enough for todays hardware and networks. more detail with current resolution standards, not higher data loads please. More pixels do not equal better, if we haven’t learned anything from the megapixel battles.