Possible future...

[lasse1309]

we're on it

of course you would have to build a lot more for a scene where you can actually move freely: but there's again the question whether you want that or not. the "zen garden" demo by epic shows in a very nice way how you could navigate through such scene.

 

fromm discussion with clients i know for exampe the ego-shooter style walk through is nothing you want to sell an expensive apartment with.

this has to be done different - how is still to be elaborated

 

 

I actually cannot see why you should hate it. It can be very laborious, of course -

hehe, maybe i sound a bit too positivistic in the text but tbh, i am very excited. i made my bad experiences with different other real time engines. cryengine may have a better dynamic lighting, but the whole shading and import thing is hell on earth! honestly, i prefer baking and unwrapping a bit to not get it done at all

[lasse1309]

I'd like to see that.

try it

[philippelamoureux]

try it

 

I actually tried it with a Evermotion scene. But I used LPV so it doesn't count...Not 100% sure but even if I'd flatten mapped all these buildings, there would've been seams problems. Shadows been broken. That's another problem for me.

Here's my recording with shadow play on my gtx 670. Had a version with trees after that. Never dared trying the lighmass way tho. :-(

[lasse1309]

oh, yeah that's definetely a big scene but well - maybe one would spilt the building up into different parts you can basically instance.. and when getting closer you would change the sun become dynamic. the lightmass then would only do the "GI" and there you would not se broken shadows.

 

of course it's a bit labourious, but i know it's possible :')

 

but you are right: the time saved in "production" in terms of rendertimes will be consumed at a different point in the process, for sure!

cheers

Edited December 17, 2014 by lasse1309

[philippelamoureux]

oh, yeah that's definetely a big scene but well - maybe one would spilt the building up into different parts you can basically instance.. and when getting closer you would change the sund become dynamic. the lightmass then would only do the "GI" and there you would not se broken shadows.

 

of course it's a bit labourious, but i know it's possible :')

 

but you are right: the time saved in "production" in terms of rendertimes will be consumed at a different point in the process, for sure!

cheers

 

The fact render times are pretty much inexistant is interesting for freelancer or small studios tho. Now you can make cool stuff without having to spend much on hardware!!! Looking forward to see your future work! I'm also always keeping an eye open on ue4 dev and updates! Might try it again someday!

[Scott Schroeder]

Here's the thing, and I'm trying not to be Mr. Grumpy Gills, but all of the "Wow! Look at what UE4 can do for arch viz!" scenes are incredibly simple and devoid of any detail. They are all the black and white, spartan furniture, 1-2 room studios. Show me a 5,000 sq ft house, interior and exterior, fully detailed out. Show a me a condo development where you can explore the property and the various condos in depth. Show me a 50,000 spectator football stadium fully built, with people, inside and out. Then make changes at the 11th hour and see how well the process holds up.

 

With larger scenes and more complex stuff in there, the whole auto-unwrap starts to fall apart. If you want your real time at anything above 1fps, you are going to need to start optimizing your lightmap unwraps to give the most UV space to the faces that need it so you can reduce the size of the lightmap required to show the object still in nice lighting detail. The whole, just render everything at 400k doesn't work that well when you have to deal with many many different types of objects.

 

Again, I'm not saying scenes like in the article are not superbly high quality. I'm just saying they represent about less than 10% of all arch viz type work out there, so you need to take this real time stuff with a little bit of salt. You know the common thread behind all of the great UE4 work? It's not UE4. It's the fact that the person using the software is an incredibly talented artist to begin with.

 

Grumpyness over.

[lasse1309]

Here is few items from search results, after feeling deja-vu reading through the fundamental question of the thread (disregard real-time vs render, just think "industry killer"), remember, no matter the ease, quality and price of cameras, photographers have managed to scratch out a living.

 

Some of these posts are classics, peruse them, notice a thread emerge, don't forget to note the date of the posts...

 

http://forums.cgarchitect.com/6184-future-architectural-visualization-3.html#post49233

 

http://forums.cgarchitect.com/23988-changes-our-industry.html

 

http://forums.cgarchitect.com/29542-impression-autodesk-2009-danger-cg-arch-industry-disappearing.html

 

http://forums.cgarchitect.com/61031-chasing-dying-industry.html

 

these are a great read!

honestly, i dont think the "real-time thingy" will replace our regular work in near future or will ever at all. Of course off-line rendering will become faster and possible quality of real-time engines will decrease - but i see it more as an add-on. actually it's a different product for different purposes. animations for arch/real-estate is only for prestigious projects that want to "show" something or justify a luxurious price as well today. expensive things are hard to sell and every "gimmick" wellcome to set you apart from the competitor.

 

the regular set for the regular-joe-developement will stay and I don't expect this being replaced by "real-time", at least not in the way we think of it today.

 

 

Here's the thing, and I'm trying not to be Mr. Grumpy Gills, but all of the "Wow! Look at what UE4 can do for arch viz!" scenes are incredibly simple and devoid of any detail. They are all the black and white, spartan furniture, 1-2 room studios. Show me a 5,000 sq ft house, interior and exterior, fully detailed out. Show a me a condo development where you can explore the property and the various condos in depth. Show me a 50,000 spectator football stadium fully built, with people, inside and out. Then make changes at the 11th hour and see how well the process holds up.

 

With larger scenes and more complex stuff in there, the whole auto-unwrap starts to fall apart. If you want your real time at anything above 1fps, you are going to need to start optimizing your lightmap unwraps to give the most UV space to the faces that need it so you can reduce the size of the lightmap required to show the object still in nice lighting detail. The whole, just render everything at 400k doesn't work that well when you have to deal with many many different types of objects.

 

Again, I'm not saying scenes like in the article are not superbly high quality. I'm just saying they represent about less than 10% of all arch viz type work out there, so you need to take this real time stuff with a little bit of salt. You know the common thread behind all of the great UE4 work? It's not UE4. It's the fact that the person using the software is an incredibly talented artist to begin with.

 

Grumpyness over.

 

you are right in everything! this has to be tried out - and of course that requires a lot of work. it is not done within no time - but that also is not the case doing it the "traditional" way. all I can say is, the UE offers a lot of possibilities for optimizing things.

 

honestly, what i myself really liked about the unreal engine is the possibility to get rid of the AOish look you always get in other real-time engines and the really high quality of GI!

and thats also the reason why i took this very white scene. in darker environments it is always fairly easy to hide the faults

 

best

Edited December 17, 2014 by lasse1309

[CliveG]

Where's my picture of a grumpy cat when I need it?

 

Step back and look at the bigger picture, ignore the little points "UE4 that, CryEngine this" and look at the original statements:

 

1. Vray and other brute force render engines will be useless. (Edit: this is because it's too expensive to do photorealism to the smallest detail)

 

This is just plain wrong, photorealism is all around you, look at the adverts in your in-flight magazine or on TV..... chock full of CG photo-realism. People are prepared to pay for the small details of photo-realism now, do you think they're going to go away or say, "actually let's go back to the garbage looking stuff of 10 years ago, because it's cheaper!"

 

Do we think companies such as Chaos Group are going to just roll over and let go of their strong position in the marketplace when technology evolves or do you think they will evolve with it, bringing products like Vray along for the ride, improving and adapting all the time so they can stay in business.

 

A little look at history would tell you that if these technologies were in any way a threat, AutoDesk would have bought them by now.

 

2. Anyone doing CAD today, will be able to do visualization.

 

Refer to my original post on this thread, but obviously they can do so already with many products, but of course the results are s%%t, unless they are creatively talented.

 

3. There won't be a need to outsource the task, more than it is today to outsource CAD.

 

This will always depend on the quality of output deemed acceptable and the in-house resources available.

 

4. We'll see in-house or outsourced render farms and vfx and post production firms closing doors.

 

Opposite, I believe we will see more of all of these, even the render-farms. There will always be people wanting to push the limits. Push the limits of the software and push the limits of the hardware.

 

As an analogy look at another "Photorealism" industry - Photography. By definition we all achieve photorealism every time we pick up a camera of any sort, so why would we need to pay anybody for taking photographs? Yet there is still a demand for photographers of all stripes, Studio, Action, Sport, News, Fashion... you name it. So how could that be if any jerk can achieve photorealism when they click the shutter? Because the results are not as good (marketable etc.) as when a professional does it. Go to London, I think there are more professional photographers than taxi's.

 

That tells you that Photorealism isn't some finite destination, once we (or the game engines) reach that zenith, "well we all may as well pack up and go home boy's we've done it" Instead it's human nature, we just keep exploring and creating.

 

If you think otherwise you've got a very naive or incomplete sense of the industry.

 

Maybe I'm missing something in the original idea of the post, but no matter what happens with game engine development. Offline or realtime work is always going to require a creative to interface with the suits and the techies, full stop. The inference that once game engine development achieves "easy" photorealism that Vray is dead, creativity is un-necessary and outsourced talent is redundant - is (I think) wrong.

[lasse1309]

there will always be painting, drawing, photography and visualization.

 

the only thing which might happen many years down the road is there won't be a distinction between "real-time" and "off-line".

both are heading into directions that one day they will meet and merge! still there has to be a human that rides the machine and gets the aesthetics out of it.

 

btw: merry christmas, everyone

 

Edited December 18, 2014 by lasse1309

[philippelamoureux]

there will always be painting, drawing, photography and visualization.

 

the only thing whihc might happen many years down the road is ther won't be a distinciont between "real-time" and "off-line".

both are heading into directions that one day they will meet and merge! still there has to be a human that rides the machine and gets the aesthetics out of it.

 

btw: merry christmas, everyone

 

 

That is a nice animation and great use of Ue4. GJ!!!

[Amen]

I concur with Lasse.

 

I think we will have more physiks stuff like fluids, cloth... that will be better integrated.

But someday everything should be realtime.

 

Jan

 

there will always be painting, drawing, photography and visualization.

 

the only thing whihc might happen many years down the road is ther won't be a distinciont between "real-time" and "off-line".

both are heading into directions that one day they will meet and merge! still there has to be a human that rides the machine and gets the aesthetics out of it.

 

btw: merry christmas, everyone