UE5 VS Maxon


#1

I feel I am faithless…
David, please give me the hope this year to show me we can still sleep together…
EX6samXWsAMZBv-


#2

I hope you don’t just judge on appearance … you may have surprises!


#3

Apples and oranges.

At the moment, you can use UE4 to render your assets. With Substance Painter the results would be great. But…

You wont find a game engine as flexible as Redshift / Arnold, etc… I dont think you can get cryptomatte, multipass or even alpha rendering from UE for postprocessing in Nuke / AE.

I use Ross Voyager and It has some issues when rendering opacity into a Live Studio.

https://www.rossvideo.com/products-services/acquisition-production/virtual-production/voyager/


#4

Not sure what the point of this thread is.
However the Analogy depicted is flawed.:relieved:

As impressive as the UE 5 Demo is, (and it IS impressive!!)
Game engines are still fundamentally content consumption and playback Software.

You still needed Zbrush to sculpt those Million triangle statues
we saw…or you needed to import the quixel megascan content.

Or create the content yourself and for that you need a true 3DCC yes??
or at the very least a good modeling and texturing package.

That said ,I am NOT going defend C4D here, as I have Moved completely to Blender for content creation and rendering even VFX
shots.:hugs:

However this UE5 demo does raise the bar for everyone in the area of viewport performance at least where there seems to be universal agreement that C4D’s viewport performance is far behind most of the major 3DCC’s.:roll_eyes:


#5

I wonder how hard it would be to have the UE5 engine drive the viewport of C4D (or any DCC for that matter). I’m sure it’s massively complicated – but could it be done? I guess I’ll have to download UE5 next year and see what it’s like as a realtime renderer.


#6

Well there is, of course the live link for Maya& Iclone pro.

But that is streaming data to Unreal.

Seems it would be quite the challenge to bring the UE viewport into your 3DCC. :thinking:


#7

There’s a lot more under the hood than just drawing triangles on screen happening there. What you can see when they switch to the triangle view is that what you see isn’t what you gave. It’s getting to the triangle per pixel level but no model is made that way. Their drawing looks nice but then it’s not what you are working on is what you see. So for modeling the efficiencies mean a display that isn’t accurate to what you are actually making.

That sad as impressive as it is, and true realtime, it’s only just catching up the Clarisse viewport. Better shaders for sure but those are clever shaders not true raytrace. Billion of polygons hasn’t been an impressive term for me in 4 years. For a game engine and true clean realtime yes, but for a DCC trillions realtime in view for lighting has been a reality for quite some time. I think our record scene at dneg, which honestly was the most foolish matte painting scene, was 40 trillion. But that was a single unique object instance many times. Yorktown was 5 billion unique objects and a few trillion and you could move and light in at interactive rates.

That said there isn’t any reason not to start using unreal as you final scene assembly and rendering tool at this point. As the others said you’d still need standard dccs for authoring though.


#8

You mentioned Clarisse already. Any other 3D program can handle that? I guess Houdini…? Not sure, only tried once.


#9

No, even Houdini needs to load as bounding boxes and the likes.

The thing is the memory management and triangle display of a typical OpenGL viewport just can’t handle this level of geo. Better instancing methods can help a lot as really any demo including unreals, is a lot of instancing and we see S22 getting better at that too. What unreal has done with their tessellation in view space is levels of screenspace optimizations that are not easy or necessarily suited to a DCC. Clarisse gets around this by not using opengl and having a raytrace viewport. If you use a realtime raytracer with cinema or Houdini, you get something similar, but with the disadvantage that the software and renderer have extra conversion and memory managements that hinder it. C4D needs to store the scene in its memory, draw it in its view, and pass that scene off to redshift or octane etc. The demo of redshift in Clarisse is a prime example of how their approach to scene management for a raytrace view benefits other render engines compared to redshift in maya that has to support a viewport.


#10

They are developing a hydra delegate for Unreal, which would (in theory) allow the engine to run in any DCC that supported hydra.

But then the question is: How long until Cinema supports USD (which would result in a fundamental shift in C4D’s internal workings) and then how long before they integrate Hydra.

The question then becomes: does C4D’s new framework support a relatively easy integration of USD down the road, or was the initial concept for it drawn up in such a way that USD doesn’t fit well at all.?


#11

Hail Hydra. !!!.:raising_hand_woman:‍♀️


#12

Originally I just want to make something fun for those who are suffered with C4D’s viewport and default rendering. But I do believe that with the David’s new leadership, Maxon is running in the right direction and just hope they can make something happen faster than before
I do hope in the summer, we can see something new , like:

more options for subscription: annual pay, month pay, 3 month pay, 6 month pay, 9 month pay, days pay, even hours pay
a new Intel Ospray based Physical Render could be done (we could have interactive render viewport this way)
new Boole for hard surface modeling
more spline types
more efficient artist friendly modeling tools
more deformers
new cloud system
new pyro cluster
more useful default particles
merge PBR, standard, Node, Uber shader to a universal shader, one shader for everything like before
redesigned expresso
automatic rigging
and refresh their blog faster

Unreal is eating the boundaries of VFX, animation, motion graphic and more, one day, they could redefine the entertainment industry. look at how capital, investment and people worship it. C4D should make some new flags with something unique, so that they can still standing in the future.

Let’s see what’s gonna happen after summer.


#13

Which begs the question, why do you continue to troll… I mean post here. You say you’ve moved on, then please kindly do so completely


#14

Well I personally welcome ThreeDDudes input here. Like it or not, Cinema4D is still the ‘catching up’ 3d app - like it always has been. My old collab buddies jumped to Houdini years ago and have never looked back. Blender has turned my head too, and I welcome input from folk who have made the transition or use both. There is definitely an on-going discussion about how Cinema4D is continuously lagging behind and potentially losing users. If you some how feel betrayed by his decision to change his pipeline, thats your fragility speaking.

ThreeDDude is a hugely helpful fellow who shares his knowledge and insight on many forums. He’s one of the good guys.