UE4 Realistic Rendering Example

Some pretty cool stuff:

https://docs.unrealengine.com/latest/INT/Resources/Showcases/RealisticRendering/index.html

Honestly that still looks like something accurender used to spit out a decade ago but impressive nonetheless. BTW does it still exist?

Perhaps, but Accurender would have taken all night to get to that level for a single image whereas this is navigable in realtime.

Perhaps, but Accurender would have taken all night to get to that level for a single image whereas this is navigable in realtime.

Yes, that makes it impressive. I still find it a little awkward and dizzying navigating it as in a video game.

saw it at polycount thread today as well ( I haven't had time to play with demos myself ), and I am impressed as well. Of course, the AO look is always nauseating, but nothing that some baking outside can't solve. But the speculars are just bomb.

I know, I need work to slow up a bit so I can put my UE4 license to work. I dig how they used RGB masks to define areas of texture, like on the coffee table. I'm expecting to see some stunning works coming out as UE4 picks up more steam and has been on the market a bit longer.

I have U4 and this demo scene running on my PC right now......

In the beginning it was a bit of a pain in the butt, since I just could not find a way to export and import meshes properly from Cinema 4D. But after a lot of swearing and googling, I found a way to do it, and now it works....

So far I am quite stoked, it feels like Archviz in real time. You can change and apply materials and colors, without hitting any renderbutton and waiting. I think for architects, especially in the concept phase, it can be a powerfull tool, since you can walk through your rooms try different things and really "feel" it. As we all know, you can cheat a lot in an image with the camera angle and lense. If you walk trough it is another game......

I don't think it will replace common Archviz, but it will become a powerfull addition to it......

I am absolutely sure it won't spread :- ) It will be niche specialty, which I I think is good. Atleast with powerful game engines as UE4 and CE(3).

My friend from Slovakia did some of the first tutorials for "arch-glass" in UDK(UE3) some time ago and this is what that shader look like:

https://forums.epicgames.com/threads/789955-Realistic-Glass-Materials-in-UDK

I know it's simpler now, but not that much.

In all honestly, there are so many people who struggle to create simple Vray glass with 3 ticks of button, that the know-how and craftmanship required to master-up realtime, will simply stay out of their means forever. It's not Lumion type simple.

CryEngine which I work with is quite bit simpler than this, and more natural for single guys coming out of 3dsMax off-line combo, although I myself don't have the PBR build which will come public in May.

My brother already published his cleared version of CE on Github if any programmer wants to compile it. We might release that as some sort of vizualization package&support 'whoknowswhen' later, a lot later.

https://github.com/MatusT/MiniDLL

You can't compare Unreal 3 to Unreal 4. In unreal 3 you where lost woithout any C++ program skills. U4 is total different. You don't need program skills anymore. It's all visual.

This is what the glass Shader looks in U4:

Imo it is not harder or more complicated than a good realistic vray shader.

But hey, nobody is forced tu use it, I just think its awesome, and you can easily safe HI-Res Images with it too, no problem......

As I said, a simpler, but not that much ;- ) I've had the UE4 licence since day one by the way.

I wouldn't really go as far to say "no program skills", even with blueprints you need the same state of mind, which is more important that knowing certain C classes by name.

All in, all out, quick-learners will obviously pick it up, no matter what, but it's still way behind the easyness where off-line rendering evolved over the years, and look how many mastered that. I would say you're more exception to confirm.

Hello Guys,

I been following the discussion and I like Scott comments in several threads. I visited all your website guys and you did a very impressive and beautiful work. I think that I'm the oldest into this discussion because John Dollus is for only one day,.... younger than me !!!!!!

I think that the group of technologies required for real time solutions in the ArchViz industry are almost there. May be they are and I didn't know. In any case, if they are not. I believe that they are only a few month away from us.... not years.

I been in the industry since 1992. I started with 3D Studio R2 & DOS so I saw mostly every little step. My brother and I were pioneers in Argentina. We created ArX Solutions. We were only 5 guys in the whole country working something called Arch Viz. It was the time that a simple instruction such "redraw all" of a small CAD file of single family home in Autocad took me three hours..... In 1997 we started to work with VRML. In 1998 we created a demo real that Intel Corporation used it in all the presentations to introduced Pentium III in South American. We created it in VRML 97. We used a basic geometry and baked all the textures in photoshop. It was a very artistic and manual process. So interactivity is not brand new from my personal opinion. I'm sending you an image from that 16 years old file. (Makes me feel so old....)

But real time solution came to me alive again when I found Oculus Rift in Quickstarter. I reminded me how excited I was exploring a simple chair in VRML !!! Since then I been doing some research and I run some test. Up to date and it is a personal opinion the best solution right now for the ArchViz is EU4 for interiors, CryEngine for exteriors and Unity for web/movil.

I love VR and it is my passion so Juraj, ….....you are not alone !!!

I think that a good VR experience requires a combination of multiple factors and all of them sound that are near to touch base.

The real time engine,..... the state of art of the file, ....the graphic card.... the processor,..... the ram, .....the display, etc..... TOO MANY THINGS must be balanced for us.

The part that is a big problem from my perspective is the production pipeline and business itself.

- To meet the necessary level of quality with the typical time frame is very, very difficult. So we must be realistic in term of expectations.

- More time means more money. So if we would like to start doing this as part of our portfolio of services we will need to marketing it as a brand new product.

- Like any brand new product means time consuming and more money.... : (

- The Lack of flexibility. For example, in Max, any change of geometry is not necessary a big deal. In any RT engine is a small tragedy based on my experience. We found it in the past in the hard way....

- Lack of good examples.! in ArchViz, I found only a few so so. In fact the Realistic Rndering that Scott shared and it is the EU4 marketplace is so so in comparison with a mid quality rendering. In my opinion, I couldn't find ANY excellent ArchViz demo in real time yet. So unfortunately it seams to me that we don't have anyone else that we can ask. So we must be ready to go where no man has gone before......I spent several hours at night looking for good samples or artist. All of them seams to me that are focusing all the efforts to represent a new mother ship or detonated architecture. I couldn't. If you could find anyone who can produce a RT space like the white room from Juraj for example, please share his contact with me please !!!! I will hire him, I'm not kidding at all !!!

- All the engines and most of the hardware were created for the game industry. The budgets and resources of any medium size game developer company is by far larger that any of top 5 largest ArchViz firms in the US. So these are tools that weren't created for us. I always remember that we are free riders of his industry. Thanks god that they exist. Without them our computers, 3D Studio Max and the plugins could cost more than hundred thousand. This is not a joke. When I started, the Ferraris were the Silicon Graphic and with the basic software (Alias, Wavefront or SoftImage) costs more than that. The graphic card, RT engines, Oculus Rift, etc were designed for their industry so we are the one who must create this transition and learn a new process curve. I'm totally agreed with David Crabtree. (he posted in another thread)

So, this is the deal. I would like to run several RT test in the short term. If you are interested to be part of them, join me in this process. I have computers, space, oculus rift and some resources.... but more important, the determination to go straight to the bottom of this. I want to find the answer to two questions.

1-Are the available RT solutions (Soft & Hardware), ready for the level of quality demanded by an ArchViz industry?

2-Do we have a product now that is waiting for us or it something that remains as a potential product for a near future?

If these are you question too and you are interested to find these answers. You are very welcome !!! I'm in Miami, FL so anyone is welcome to come to town and visit me.

Saludos,

Nice post Patricio :- )

Regarding high-quality archviz, on high-scale, I guess ENODO is the top player but it's mostly due to scale, and since they orient heavily towards "serious gaming" nowadays, it's not focusing too much on realism.

I got first intrigued by this video from argentinian guys Tearte. Simple room in Unity, but it's the baked GI from Vray that trully makes it look good. I don't think the real-time GI or baked lighting (from Lightmass or similar tech) will be able to compete anytime soon with that.

I suppose it could be made much much better looking now, with PBR builds of engines ( I am eagerly waiting CE3's May public build), varied shading enhancement ( like Marmoset Skyshop for Unity),etc...

Juraj,

I saw this video before and I agreed, it good. But the minimalism in this case and the lack of different space make this difficult to define. I think that the level of light is good but the textures are not killing me. I'm missing some glossiness and reflections. I think that is the problem that you can find if you only bake to texture. If you work over the texture and edit the material adding some leverage, the scene can be much more realistic.

Any way, I think that this kind of videos reinforced my concept. Some artist have been doing some individual research but based of the lack of time or money, the test can't jump to the next level.

To illustrate my point let me caputre one frame of this animation and let me send to you as a rendering. How many point will you give me for the quality of my rendering from 1 to 10?

What I would like to answer is, can we get a RT solution almost 95% closer to what both consider a 9 or 10 points rendering? I couldn't find anything that as a rendering is more than 5/6. What do you think?

I agree with your sentiment, I think so likewise, I am not interested in real-time viz if it looks absolutely subpar to current level of off-line visualization. It is because of that I find Lumion and those alike to be great tool for internal development, but don't believe it can used for serious marketing presentation.

I have zero time to focus on it right now but I likewise plan on extensive research during summer on this. Or the moment I get my PBR build of CryEngine :- ) So far my belief is that the quality can absolutely be reached, by the workflow and effort to reach it will be painful and painstakingly expensive.

I already jumped and I'm in that journey right now so I will keep you posted. : )

The image of the old VRML is https://www.dropbox.com/s/j52aovt21auuajp/ArX.jpg

Another UE4 archviz rendering :

, ,

Mark Rein, The VP of Epic Games just tweeted "Museum kiosks, medical applications, movie-making, automotive/real estate/product visualizations are examples of royalty-free uses of UE4!"

"And no additional cost other than the $19/month subscription."

Pretty cool imo!

Saw it too :- ) Very cool ! It's really taking off

I found a little plugin for 3ds max to speed up the pipeline between 3ds and UE4. I have not played much with it but it looks promising. here it is :

http://www.polycount.com/forum/showthread.php?t=134683

Philippe,

Thank you for your post.

Did you create it?

I would like to know more about it.

No I'm not the author of these images. They've been done by a guy named ''mabone''. It's all I know. Read that on ''unrealdevs'' twitter account.

Nice little plugin for 3ds max to speed up your pipeline between 3ds max and UE4 (export, collisions, etc). can be found here. Again, i'm not the author!!! :-P http://www.polycount.com/forum/showthread.php?t=134683

Interesting thread. I've used Unity before now I think i'll have to look at UE4. I often look at demo's like this and they look great but how long did it take them to do it.

Dan,

Do not discharge Unity. Unity will release soon the 5th version in a few weeks with several improvements. (http://unity3d.com/5?gclid=CMaRntq7zL4CFSJqOgodBSIAiA) In Unity you can achieve a very similar level of quality if you push hard with the baked textures and material edition. EU4 has a better finishing and impressive looking but it has some very serious problems with reflections right now. Unity and UDK can handle much better but it will consume a lot of resources. I personally don't have a very clear winner.

CryEngine is now available on Steam for $10/mo with no royalties.

What to devote time to learning? Guess that is a much better problem to have as opposed to the previous problem of not being able to afford these packages in the first place.

Try both ;- )

IMHO (my perspective), CE3 is still way worse for actual game developing, with almost ignorant approach from super-silent Devs. There are also lot of trade-offs...

But for visualising, it has easier workflow, is more similar to traditioanl rendering with the simple material system with preset shaders, has better dynamic lighting out of the box,etc..

BTW, currently, you can't use the Steam 10 dollars version for any other purposed than indie dev. The EULA outright denies Architectural visualition (it names it directly).

Let's see what comes further.