Unreal Engine in Architectural Visualization and why AI makes me sad. Which way now?

And then uh something changed for me in the in 2015 16 when I discovered Unreal Engine and I thought this is it. This is

now going to proliferate through the industry. We always have to stay one

industry. We always have to stay one step ahead. Think about what can we do

step ahead. Think about what can we do that AI cannot do yet. Uh at the beginning of those AI images you kind of did you you wanted a control you you you were saying I cannot really prompt it

exactly you know like I give it a general prompt and it gives me something but I want to fine-tune it. Maybe I want to make this stone material slightly lighter. And now there are tools that

lighter. And now there are tools that that do exactly that.

Something I've been passionate about since I was I don't know probably 7 years old. And uh as a computational

years old. And uh as a computational designer, someone who actually uses computers and is concentrated on software in architecture, I uh I really

I really want to uh make a couple of points about the visualization and and discuss it and maybe um make you think about the progress of it and where where

it's all going and how do you as an architect or someone that works in visualization or even in gaming industry like fit there. I have to say this every time I make a video because sometimes

somebody's watching this channel for the first time, but I'm an architect with a PhD uh uh that worked on um on a lot of large scale projects, airports, Las

Vegas sphere and so on, generating the design and the complete 3D models, product fabrication file, production files. So mostly working on the

files. So mostly working on the production side, helping a lot uh to helping architects a lot in the design phase, especially now with with the startup I'm involved with called Cosmic

Buildings where we're building houses, rebuilding LA after the fire. But u uh I'm trying to unite this this this entire chain right from the from the design uh to to the production. So

computational design is all about you know programming and Rhino Grasshopper and so on, writing plugins for Revit and so on. But there is this like side thing

so on. But there is this like side thing which is visualizations. And when I was in high school even 9697 dating myself here a bit I bought a book

3D studio max 2.5 I think. So I learned how to render was fascinated. It was the the global illuminations appeared for the first time. I was able to do like

photorealistic renderings in 99 2000 where like the first year of of college and uh it was pretty early for that but it was fascinating. I was reading about the rendering, how it works and the the

global illumination with the bouncing of the lights and how the materials are calculated and so on and it it's always fascinated me. I never got into it

fascinated me. I never got into it really because I thought it was something that a lot of people are doing and the competition is high. So I never wanted to become like a professional

visualizer or renderer. And then uh something changed for me in the in 2015 16 when I discovered Unreal Engine and then I thought this is it. This is a

paradigm shift. All the architects or

paradigm shift. All the architects or the visualizer in architecture are going to switch to Unreal within six months.

I'm already late. I have to start offering this. This is my way to jump

offering this. This is my way to jump back into the visualization. The logic

was sound. You have renderings. You have

images. Then you have the next thing which is better which are videos where you have cameras and and and going through the building and so on. But

that's all static. You cannot interact with it. So what's what's better than

with it. So what's what's better than that? photorealistic immersion, right?

that? photorealistic immersion, right?

You enter into the building, you walk around, you look where you want to look, you change materials, you move furniture and so on. And I thought, this is it.

I'm already late. I even registered a company called Surreal. You can still find it on surreal.com.

S O Real Surreal. the the idea like we started using Unreal Engine start offering like VR tours and this even desktop like basically video games. We

could even export it for PlayStation and so so you can walk around the house and do exactly that like change materials and so on. You have seen this so far but my idea was uh we even like we got

couple of projects we got paid for it and I thought this is it. This is now going to proliferate through the industry. 10 years later I don't see it

industry. 10 years later I don't see it proliferating at all. I mean there are still companies that do this kind of unreal jes and tours and u there's nothing really happening um if if the

most what I see from the architectural practice we still deal with images and still deal with videos now I'm not going to go into why that is because I have some thoughts on my own and some

analysis it has to do with like with budgets first of all like you have to get paid for the by the architects do they have low budgets so they can only pay for a couple of renderings and so on

they if they use something like this for competitions. Again, competitions don't

competitions. Again, competitions don't accept like XF files with like video game files or VR. Like they still want images or sometimes videos and that's

enough. And then you can go to the

enough. And then you can go to the developers, the ones that for example create condo buildings and then try to sell them. And you can say, look, you

sell them. And you can say, look, you can sell your apartments better if you have your customers can walk through and change the materials. There are problems there. For example, when I was starting,

there. For example, when I was starting, I was in Stoodgart in Germany and uh the developers had no problem selling up the apartments, which is still the case, I

think, in in cities, most cities today, where they just get an put an ad that they are about to build a building and they sell most of the apartments. They

don't even need the normal rendering.

So, they didn't need they had didn't have the demand for that. And then

additionally, they really didn't want the users to choose their materials or whatever. they didn't want to give them

whatever. they didn't want to give them that sort of a control and especially like pay more for that. So there are other reasons why that didn't work. I'm

not saying it's not working across the board. There are companies that actually

board. There are companies that actually do this what I attempted to do, but I think there's still a small number. I my

my idea was that this is going to replace rendering at all. We're all

going to have video games and walking through our future buildings. And uh so I love Unreal Engine and like I I I loved working in it. Uh I made a YouTube video when Unreal Engine 5 came out and

it was at the moment by my my most viewed video by far. It reached 100,000 in a few days. And I explained the technology because behind I explained these new technologies that Unreal

Engine 5 brought which was the the nanite this this optimized way of rendering which was lumen the optimized way of calculating light. And so I explained it in a little bit more

technical detail and that was uh very interesting for me because there is something about CGI and there is something about creating materials uh since couple of since many years now we

have this physically based u materials where we we can define the normal maps and so on. So I'm not going to go into details of explaining that nor am I

expert in that but that's something that very that interested me very much and so now basically I want to make a cross-section of where we are today and just to make a point that it interests

me very much I mean I I this this is the stuff I read right the game engine architecture like this is a thousandpage book that describes really really really

minute details like everything about creating a game engine which we will not do anymore. I guess there will not be so

do anymore. I guess there will not be so many game engines created from scratch when we have Unreal and similar ones.

But that's just the the the emphasize how much interested and I and why it makes me sad that I don't know how to utilize that. Right? So this video is

utilize that. Right? So this video is more of more questions than answers and I am asking you how to utilize this, how to use Unreal Engine uh how to use great

visualization and how to create the demand or where to find the demand especially in the time of AI. So let's

let's jump forward and look at the AI today. So this is the reality today. If

today. So this is the reality today. If

you are someone that renders and someone that visualizes, you're in a very tough spot. I told you I'm a part of a startup

spot. I told you I'm a part of a startup called Cosmic Buildings. We build a lot of houses in Los Angeles, luxury single family homes. We have u several

family homes. We have u several construction sites ongoing um right now.

And because we use we create our own house configurators, building configurators that gives us the basic shape of the building. So you you see the walls, the windows, the doors and so

on in that normal kind of a u viewport that we use in Rhino because we mostly do it in Rhino.

And then you can take that as an input to nano banana or whichever model uh you're using or whichever model will be next and it will give you a rendering.

And the cool thing is now is that uh at the beginning of those AI images, you kind of did you you wanted a control.

You you you were saying I cannot really prompt it exactly, you know, like I give it a general prompt and it gives me something, but I want to fine-tune it.

Maybe I want to make this stone material slightly lighter. Maybe I want to make

slightly lighter. Maybe I want to make the doorork knob slightly more golden or whatnot. And now there are tools that

whatnot. And now there are tools that that do exactly that. So they do the rendering and they automatically recognize the outlines of the window, of the toilet, of the doorork knob or

whatever you want. So you can click on it and like in Photoshop except uh you can prompt it. You say click on the doororknob. It selects the outline of

doororknob. It selects the outline of the door knob. You say change the material of the doororknob, change the material of the toilet. Make the

reflection on the video slightly higher.

Make this stone slightly more marble or whatever. So you can do this kind of and

whatever. So you can do this kind of and now the and and then you use that image as input for video. So now the the it can assume from several images it can

assume the 3D dimension of your house.

So you can even create like videos rotating around the house and so on. And

this is getting even better and better.

So it makes me sad because uh there is there is the glass half full and glass half empty uh component of this. So the

glass half full is that we will all become these creative directors now, right? So we all have we are all this

right? So we all have we are all this uh film directors, design directors, creative directors, call it how how you want that have these agents and have these tools and we kind of just tell it

how to do what to do.

[gasps] The glass half empty side of that is there a lot of us that love the technical stuff like we have like like to make the machine with our own hands and not just tell the robot how to do

it. We love to open Substance Designer

it. We love to open Substance Designer which is a program for uh creating like this materials uh realistic procedure materials and then take three hours to

create a perfect concrete material and then you can see it in the rendering you know and know and now AI is robbing us of all that. So I even remember buying the new substance designer right before

like midjourney and um especially Nano Banana came out and just saying to myself like what's the point? Why would

I use hours to create a material when you can prompt the perfect material or even prompt it directly on the image?

So the questions are like what are new frontiers to discover like is there a frontier where we can now go back and do something technical yourself or is that

lost forever? What whatever you're

lost forever? What whatever you're creating, you're automatically creating like a basic uh distribution normal

distribution of of of uh of that content of that media whatever it is and then most of it will be average some will be crap and then a very small part of it

will be uh will be excellent. So now uh and if you think about it as a pyramid again at the top is what's excellent and now we're just going to be raising the

bar of of of what's excellent. So maybe

something that a movie director was able to to create or render 20 years ago maybe not now everybody's going to be able to render that or

create that and then we'll be jumping higher and and uh striving for much more. But not to branch off too much. I

more. But not to branch off too much. I

want to come back to the to the architecture and computation and rendering and visualization.

Uh and uh see if we can draw some conclusions from this random thoughts. I

want to mention this is not a structure video. I wrote down here I want to talk

video. I wrote down here I want to talk about Unreal Engine and architecture and why AI makes me sad. But I want to see if like I can sum this up and uh maybe draw some uh lessons and maybe some

inspiration for all of us. So the sum up is if you want to open a rendering company like think twice uh because I am

not paying you at the moment. I we are paying AI to do our renderings for us and we see the tools developing that direction. What we what we see what I

direction. What we what we see what I see underused is this area of Unreal Engine.

I mean Unreal Engine, game engine or whatever you want call it the just this dynamic exploration of of rendering. How

to dynamically walk through the space, change the space, experience the space, VR or not VR. It doesn't have to be VR.

You can do it on a TV, on the PlayStation or whatever. But this idea of of going through a space and exploring different variations like let your imagination work I think that is

not not explored. We do see expansion of AI with like this Google uh uh how is it called? Worlds. Um I forgot something

called? Worlds. Um I forgot something worlds. uh the uh the world models sorry

worlds. uh the uh the world models sorry the world models where you can prompt a world model like prompt some surrounding and how it basically works is as as you

are walking through that surrounding an AI is generating images 16 20 I don't know how many times per second

following your path of of movement so what you what looks like a 3D space is not really a 3D space it is a um sequence quence of AI generated images

frame per frame but they are made so that they they they figured out the perspective they figured out the lighting they figured out the physics so if you say start uh start spill the

water from the top the water will spill and flow in a way that's physically correct because AI figured it out as a side effect somehow so you will see these worlds and maybe

you will start manipulating those worlds and the question is how like we always have to stay stay one step ahead. Think

about what can we do that AI cannot do yet. And I see that that uh AI did not

yet. And I see that that uh AI did not enter this area of of game development uh that much at least from our

architectural visualization point of view. There is the Unreal fast fest in

view. There is the Unreal fast fest in June I think in Chicago. So maybe I will be proven otherwise. Maybe there are some tools where we can all prompt 3D worlds into existence.

But for me, the big question is what we can do. And uh I would hope and I would

can do. And uh I would hope and I would work on on on uh on developing some systems that capture the imagine of

architects that can finally make that jump from uh from images and videos into this interactive kind of visualization.

And if you're a young person interested in this fieldization as much as me, uh please

try to think about ideas how to to make this dynamic rendering a thing, right?

And how not to lose the battle to AI. I

mean, the battle to AI will be lost eventually. It has to be.

eventually. It has to be.

But for the perceivable amount of time, think about how can we stay creative directors of this and how can we raise the bar? What's above that bar? That's

the bar? What's above that bar? That's

still not raised in my opinion. We're

still doing the rendering images and videos that I can do. How can we raise the bar of architectural visualization?

taking hints from the gaming industry so that we can spark that imagination that I felt for the first time when I saw photorealistic

l renderings in the end of '90s where I saw how global illumination works when Nvidia u introduced RTX uh cards and with the

real time real time tracing and so on.

So what is the next thing that can inspire us in this CGI world? Because

that's something I passionate about and I'm guessing a lot of you and I don't want to lose the battle to AI this early. So those were my thoughts about

early. So those were my thoughts about visualization although as I said it's not really my area of narrow expertise

but it is a rare area of passion and that I hope we can uh solve together.

uh you like, you subscribe, whatever is No, subscribe is is good. Subscribing is

good because then you'll see a lot of these videos where I ramble on without anyone stopping me. And um you can go to LinkedIn. I'm still active there from

LinkedIn. I'm still active there from time to time. Aside from that, see you here in the next video and stay free.