BIG Innovates: The Ultimate Architecture Workflow with D5 Render and AI

JENS MAJDAL KAARSHOLM: All right, great. Thanks for tuning in to our session here today, where we will basically be covering how we work with D5 at BIG and how we've found it to become a super interesting tool in our workflows here at BIG.

Super quickly about myself, my name is Jens Kaarsholm. I'm the director of design technology here at BIG. I'm based in Copenhagen, hence my screen is a little bit dark right now. It's almost midnight here at the moment. I've been with BIG for 6 and 1/2 years. And I'm overseeing the Copenhagen office, but also the global tech stack and the DT side of things.

Together with me today. I also have Jeffrey Espinoza, who is the customer success manager for D5 Render. And Jeffrey was brought in very last minute and, to a degree, myself as well. Because it was actually a colleague of mine who was supposed to do this talk and run the show and do pretty much everything. So I managed to get Jeffrey in to help me do a little bit of live demoing that he will show you guys later. So that's really great. So thank you, Jeff, for jumping in last minute.

Super quickly about the office and who we are. We have six offices all over the world. Four core disciplines being landscape, engineering, architecture, and products. And we actually have-- this slide, I just realized, is a little bit outdated. Because we just had a partner promotion around recently. So I think we're in the order of 26 partners by now. And we're getting close to 800 employees.

And we're cofounded or not cofounded-- we were founded by Bjarke Ingels himself, who is still the creative mastermind behind everything we do. He's still in on pretty much all projects and has the last say in, yeah, most projects and things we do. And at BIG, we have this approach we quite often refer to as LEAPP concept, which is basically that we have Landscape Engineering, Architecture, Products, and Planning in-house.

Obviously, with architecture being the predominant or the main service we do. But we like to say that we can do anything. So we do everything from small lamps to pavilions to floating student housing in Copenhagen. We've done big office buildings in New York, headquarters for Google. Master-- sorry-- master plans the size of a city. We're also doing master plans almost at country scale.

And we've also recently been doing this plan for the planet, which is something I would definitely encourage you guys to go and check out on YouTube. If you Google or YouTube plan for the planet and then Bjarke Ingels, there should be some very interesting Ted Talks with some quite bold ideas on how we could save the planet. But some of it, I think, will really resonate once you listen to it. So definitely go and check that out.

Very briefly, just a quick look at some of our projects. And I think what this slide here shows is really the complexity we work with and also in how many places we have already completed and built projects. We have-- we've not built in every single continent yet, but we're getting quite close, I think. And we will get more back to this biosphere here in Sweden in a little bit.

Here's also another handful of projects. I'm sure you will notice some of them. There's the famous ski slope on top of a waste energy plant in Copenhagen that we completed a few years ago and a handful of other projects.

And lastly, I also just like to show this slide when I present about BIG because this is really the mantra we have or still have. It's something we came up with 15 years ago, which is this less is more attitude or mantra that we have to remain positive about the obstacles and see obstacles more as challenges and really make sure that we take on our yes hat when we go to the office in the morning.

So for the talk today, I'll start out by doing an intro to our design technology department at BIG and how we're structured. And then I'll hand it over to Jeff to show a little bit on how we use D5. And Jeff has been instrumental in helping us getting going at BIG. So he's familiar with how we work and our workflows even though he's from D5. But he's familiar with our setup.

Then I'll come back to how we use AI a bit because I think AI is especially interesting in relation to D5. Because D5 is not only just a real-time rendering engine. But it's also having a lot of AI capabilities on top of it. So I'll touch upon how we use AI at BIG.

And then I'll hand over to Jeff again to show some of the more AI-related features as well. And then at the end, I will just come back and conclude on with some final thoughts and so on. So that's the aim for today.

So yeah, let me first jump into the design technology section. Because essentially, a couple of years back, we decided to basically take a lot of these subdisciplines within an architectural practice, being BIM, like the tech-related ones, being BIM, computational design, visualization, for that matter, VR and XR as well, simulation, fabrication, and AI, we sort of took all that and put it under the same umbrella, which we now refer to as BIG Design Technology or just BIG DT.

And obviously, being in the DT team means that it's a lot of support work. It's a lot about helping our colleagues getting going. It's also a lot about training and optimizing workflows and so forth. But a lot of it, especially for someone like me, it's also about navigating the sort of ever expanding digital ecosystem of tools that we work with.

This slide here is a little bit old and outdated. So we're not using everything you see here. But there's also some tools that we're now using that is not here. And also, all the AI tools that we're using has not been added into this slide. Because then, that would almost be a mirage of bubbles all over the place. So that's intentional. But I think it's just more to paint the picture of how complex this, you could say, landscape of software is becoming.

Last year at AU, I was doing a talk on BIM. So in the case, you could say, curious about how we're set up around BIM. You can go and check out the talk I did last year on the whole BIM side of things at BIG. But today, we will focus on this and then, as mentioned before, a little bit on AI as well but predominantly this and, more or less, primarily about how we use D5.

Because I've now, as you can see, slotted D5 into the ecosystem here. And it sort of sits nicely next to Enscape, V-Ray, Twin Motion, and the likes down in the visualization corner of this map here. And it's really something that we only discovered recently. So it's a tool that we started buying into less than a year ago. So it's still somewhat new.

And that's also part of, again, you know, constantly researching and checking out new tools and things. That's when we sometimes discover these either small gems or small things here and there. And then all of a sudden, we realize, oh, there's something here, which was the case with D5 in our case. I will quickly add that Enscape is still the predominant real-time tool we use at BIG for the more day to day, daily, quick and dirty visuals, et cetera. But I'll come back to why I think both Enscape and D5 actually complements each other quite well and how we see a good fit for both of them in our toolkit at the moment.

But I think to start with, I think if we're talking visuals at BIG in general and renders, we obviously go a long way back with V-Ray. That was sort of our first primary visualization tool 15 years ago. It was the main thing we would do. We would work in Rhino and pretty much only in Rhino, AutoCAD and Rhino. And then we would render in V-Ray.

Here's just a very, very recent image from a project we're doing in Bhutan, where we're designing this beautiful master plan of an entire mindfulness city in Bhutan, which is a project I would definitely encourage you to check out on our website. It's an absolutely stunning project. So this is an image done by V-Ray, traditional V-Ray photoshopping. What's interesting with this image is that, actually, a lot of the cut-out people in this scene has been AI generated, which is quite nice, I think.

So we're still doing a lot of V-Ray classical in-house. Also, obviously, we also put a lot of images externally. But we still do a fair amount of V-Ray images. This one is also done in V-Ray, again, using a lot of AI cut-out people, but, again, completely done in-house with more traditional 3ds Max, V-Ray style workflows. This one is, by the way, from a project we had in Prague for this new Philharmonic Hall in Prague that we're doing.

Then on to how things then expanded. Because then around, I think, Enscape came out around, what was it, '15, '16-ish. And I initially-- when I used to work at Foster's for a couple of years, I initially helped bring Enscape into Foster's. And then as I moved from Foster's to BIG, I sort of took Enscape with me. Or at least, I told BIG that we definitely needed to get Enscape up and running immediately back then. So I've been playing a big part of why BIG is using Enscape.

So we've been using Enscape for quite a lot of years now. So that's also why it's a heavily embedded tool in our package. Here's just a quick example of one of the renders done out of Enscape that we recently submitted for a project. This is a project called Dymak headquarters in Odense in Denmark. Fantastic project. You should definitely also check that out on the website. It's this circular, round timber building with this highly complex roof. But it's a really, really beautiful project.

Then we've also been having a little bit of Lumion in the mix of things. But it has never really kicked off on our side, at least. I know a lot of other offices have been using Lumion a lot. But it has never really picked up at BIG.

We've also been testing a lot with Twin Motion, not only testing. We've also been using Twin Motion for a few projects. But again, somehow, it never really started to fly somehow.

Lumion, nowadays, is pretty much sidelined within BIG. We don't really use it as of right now. And I would say Twin Motion is also-- it's not that we don't use it, but it's not really being used that much. And I guess part of the reason for that is also that D5 is now coming in and has sort of squeezed nicely into the middle here.

And I think the reason why D5 is important here in the middle of things is that I think it does a lot of stuff, especially on the animation side, that it just it somehow just feels more easy and natural to do than with some of the other tools. And then with all the AI components on top of it, I think that's why we think, right now, it's really an interesting tool but also an interesting, you could say, workflow that you can create and do when using it.

So that's why this is sort of our render, not the pipeline, because it's not that we always go through D5. We're only just, as I mentioned, getting started with D5. So we're still in the early days of rolling it out. But I think it seems to be a pretty good toolkit, having these three in play at the moment.

So what is D5? For those of you who are completely brand new to the tool and have no clue, D5 is a leading real-time rendering solution or one of the leading ones. Maybe I should say it that way. And as mentioned already, it's also powered by a lot of AI-driven and cutting-edge technologies in the back, which is, again, why we find it interesting as well.

And I think, personally, as I see, I think what also makes D5 great, in many ways, is that it strikes a great balance between speed and quality. And as I also put it here, there's also just like a very nice and seamless integration between a lot of our favorite design tools and workflows. And then it's also user friendly.

It's maybe not as pressplay as Enscape is. And that may be also why Enscape still has the upper hand when it comes to the very sort of quick and dirty, get something up on the screen, and walk around. But it's still pretty user friendly and easy to go with here as well. It's like you don't really have to do much to get going. And it's very intuitive as well to use. So yeah.

But again, I don't want-- I mean, obviously, it's hard with a talk like this to not become salesy. But I really don't want to sit here and try and make or sound like a sales guy because that's not my purpose of today. And it's also not really necessarily the bigger purpose of the talk, per se. I sort of just want to show some of the reasoning behind why we have started looking at it and why we're starting to implement it.

Another reason for this is also that, as mentioned before, it supports a lot of the design platforms we use. We're obviously heavily Rhino, Revit-based. So the fact that it supports those two is a big tick mark for us. But we also use 3ds Max, which is also supported, and every now Blender. So for us, it ticks all the boxes we would need from a design tool, which is obviously a big positive as well.

I'm not going to go through all the bullet points here. You can check those out if you watch the recording or, yeah, on the website, the AU website afterwards. But I think, basically, it does a really good job on both architecture, also landscape. Actually, in a way, I would say especially on the landscape side. Because the amount of assets that's available combined with these scatter tools and procedurally generated things you can do as well, I think there's just a lot of stuff when it comes to landscaping that is really easy. And I'm sure Jeff will also touch upon a little bit of that because the example project we have lined up is also having a lot of nature into it.

And then I actually also think it does a really, really nice job when it comes to interiors. And what I also find interesting with D5 is some of these sort of, you would say, animated things you can put into your scenes as well. So you can add like fire. And you can add smoke. And you can add fog and other things like that really can help on the mood and the vibe of a scene, especially when you do the video sequences and so on.

Right, so in a little bit, we will jump into the live demo. And here, we're obviously going to, as I mentioned, we're going to look into how we use it. And before I hand over to Jeff, I just want to, in a way, introduce the project that we have picked for this live demo today. This is a project we have up in Sweden called the Biosphere.

It's a one-bedroom hotel. So there is only one room. And it's in the middle of the forest up in Lapland in Sweden. It's a project we completed back in 2019. And essentially, the idea for the project is that you have this hotel room that is basically completely, almost encapsulated by this swarm of birdhouses all over the sphere that makes this beautiful wooden sphere right in the middle of the forest with the hope that a lot of the birds in the forest will actually make use of these birdhouses.

Here's just a quick glimpse of the interiors. Oops. That was a little bit quick. And then, as you can see here, the birds have actually already started using. At least, they're coming for the food. I assume they also use the houses as well. But at least, they come for the food. And here also just a beautiful image of a more snowy scene.

So we've picked this one to show today. I would have loved to show some of the more actual work we do at the moment also with D5 or in relation to D5. Obviously, with, as you can imagine, with most architectural firms, most of the stuff we do at the moment is on ongoing projects that are not published at this point. Hence, I could not. I tried, but I could not get any of the latest and greatest stuff that we're doing to show here today. But I can assure you that we do a lot of cool stuff in the kitchen these days that we will hopefully be able to share with you all at some point soon.

But I will hand over to Jeff now and let him take it from here and show you how you can, quite quickly, build up a scene around this birdhouse project in D5.

JEFFREY ESPINOZA: Great. Thank you, Jens. I'll just go ahead and start sharing my screen. And I think that was actually the perfect segue and the perfect tee up for getting started and developing this scene just because it does feature this like lush landscape as well as this exterior and then all those intimate, close-up shots that I think D5 Render's really well-equipped to handle.

And so as you mentioned earlier, D5 Render is a standalone visualization platform. So what that means is that we provide you with a bunch of different plug and play plugins for your preferred modeling software. So here, we have Rhino. And we have the D5 Render plugin installed.

And so the process of getting this into D5 Render is not too dissimilar from what your current viz workflows might be. And so for instance, once you have your model set up, all you really need to do is just click on a geometry piece, make sure that it has a material applied. And so the material doesn't really need to be all that fleshed out. It really is just a matter of making sure that it has a material name, which is what D5 Render is picking up on.

And D5 Render can also read any scene or lighting data that you might have in Rhino. So if you've already spent some time setting up some lights in your Rhino model as well as some camera views, as we can see here, we can also go ahead and import those just by a simple click of a button.

JENS MAJDAL KAARSHOLM: And here, you can also see that it's the actual Enscape views from previous scenes we might have done in Enscape that we're now bringing further into the next tool, which I also think makes it interesting that you can sort of reuse those and bring them further.

JEFFREY ESPINOZA: Right. Exactly. So we're not really trying to replace any current workflows that you might have. We're really just trying to expand on them. And so for instance, once you're ready to transfer this over to D5 Render, all you need to click is just this D5 Render button. And it'll automatically import all your geometry, your materials, and any lighting that you might have in here.

So the idea is that you might work with two monitors. So you might have your Rhino screen on one monitor and your D5 Render screen on the other. So now any modeling changes that I do in Rhino will automatically be updated in real-time in D5 Render, as you can see here.

And so I think that the real add value here is that you can start to get real-time feedback on your designs. Because you can then get to see them in these sort of photorealistic environments. So here is a fully fleshed out scene. And now I can jump back into Rhino. And I can continue to do all my design work in Rhino and start to see what those design changes might actually look like from within D5 Render, as you can see here.

And so one of the other added advantages, as Jens mentioned, is just how well-equipped D5 Render is at handling large-scale or complex geometry. So if we take a look back at our Rhino model, you'll see that I really only have my architectural components. And then here in D5 Render is where I went and started to build out my scene by adding all these different nature assets as well as some snow and seasonality.

And so again, in order to do that, all you really need to do is just utilize our asset library. So currently, we have close to 14,000 different assets. And that will include different models, materials, particle systems, as Jens mentioned. Like if you want to have any fire or water features in here, you can do that as well.

This new scatter feature, which I think was really helpful for the development of the scene, which we can cover as well. And so again, you can utilize any materials that you might have made in Rhino. And but you could also go ahead and utilize some of our high-quality PBR textures from our asset library.

And so when it comes to actually building out the scene with all this like lush greenscape, one of the tools that we provide to you is this D5 Scatter. Because, again, like providing you with a large asset library is just one part of the solution. I think the other part is how do you quickly and efficiently build out your scene without having to manually place every little object one by one.

So if I go into my scatter feature here, I'll just go ahead and hide trees. And I'll hide the rest of the trees here as well. And so you'll notice that all of this, all these grass elements here are just part of this one scatter feature. And the way that the scatter feature works is that, if I just go ahead and click on it, you'll see that we have this global preset and this content preset.

And so you can think of the global preset as more so the patterning and the layout of all these different nature assets. So for instance, I can go ahead and click on this checkerboard lawn. And it'll apply this checkerboard patterning to my landscape here, as you can see here.

And now all I really need to do is, for now, I'll just I'll scale this up, so that we can take a closer look. But this is all being controlled by just a simple 2D image, a high-contrast 2D image that D5 Render's picking up the different values on. And so you'll notice that D5 Render went ahead and created two different subareas.

So now I can just click on this initial subarea, hit Content Preset. And then I can select one of our different content presets here. So I'll do this floral hexagon turf for now. And it'll only apply to this one scatter area. So again, pretty quickly, you can start to alter and play around with the scene and how you want the layout to happen. So I'll just go back to the initial preset that I had earlier.

And so again, just pretty quickly in like a matter of minutes, you saw me import that Rhino file into D5 Render. And then utilizing like our asset library and our scatter feature, I can very quickly go ahead and start to populate the scene with different types of assets. And so for the last part of this demo portion, I just wanted to talk about the exporting.

So again, like having these beautiful designs and renderings here in D5 Render is one thing. But then, how do you actually export them and get them in front of other eyeballs? So what you'll do is you'll click on this button here.

And then you can pretty quickly render out your image or pano image. And since the D5 Render is a real-time renderer, that means that the rendering results that you're seeing here are pretty much going to be identical to the final rendered image. It'll just be a bit more crisper, a bit sharper. And the rendering speeds will also not take you too long. It'll be done in a matter of minutes.

And we also allow you to do any video recordings. And so again, like we really want to templatize every part of the workflow. So when it comes to videos, all you really need to do is just create a new clip, select your starting point location. And then you can go into Template. And by clicking on one of these templates, it'll apply that animation to your camera path, as you can see here. So by following this workflow, you can create something that's pretty cinematic and ready to be rendered out and presented.

And so another part of this demo portion that I probably should have prefaced on was that we are recording this through Zoom. So the rendering quality that you're seeing here might not be as sharp as in my screen. But hopefully, the results that you're seeing are good enough. But yeah, thank you.

JENS MAJDAL KAARSHOLM: We have some comparison images afterwards that we can show. But yeah, we tried to fix the pixelation. But it's tricky with Zoom.

JEFFREY ESPINOZA: Yeah. But yeah.

JENS MAJDAL KAARSHOLM: But maybe I can also quickly add, what you just showed there, I think is also part of one of the things that we really find interesting with these sort of, you could almost say, like preset, template-style ways you can set the cameras to operate and work because it really means that you can very, very quickly, based on a few camera setups. And again, these cameras could be something we have done in Enscape.

Like the design team might have been sitting in Enscape and setting up some cameras they liked when they were in the very early-concept sketching phase. Then someone else can take over those cameras and then, quite quickly, convert it into a beautiful scene with a lot of scenery, trees, landscapes, et cetera. And then based on those cameras, just add some very quick movements. And in almost in a matter of minutes, you can have a pretty cinematic video shot. Yeah.

JEFFREY ESPINOZA: Exactly. Yeah. So again, we really want to make this as quick and painless as possible. But yeah, that covers the demo portion for this. Once-- for the second part of this demo, I'll be showing you some more of the AI features, what I think-- which I think are really fascinating and like, again, just more so workflows enhancements. So yeah, I'll join you guys in a moment.

JENS MAJDAL KAARSHOLM: Yeah, so thanks, Jeff. That was really good. And I think, as we just mentioned, the Zoom quality is obviously playing a little trick on us here when it comes to the live demoing. But we tried to put some comparison shots here next to each other of the build project on the right and the D5 render on the left.

And obviously, there is a difference. And I will also quickly add that we had to do this very last minute. As I mentioned, my colleague ended up having to pull out of this talk in the last minute. So Jeff only had a very, very few days, if not even like a few hours to get this scene up and running and going. So kudos for him to get it running in the first place.

I'm aware that the trees may not be the exactly correct trees and all that. So there is a whole bunch of things that we could have further adjusted and done to get this image to be even more correct. But even with this, you could say, very quick turnaround that Jeff was able to do here, I think we're still producing some pretty decent stuff, especially if you take into consideration that the image on the right is the real finished project.

Here, again, obviously, there is a little bit of a difference. And you can clearly see that the one on the left is computer generated. But again, if we would have had a little bit more time and would have tweaked it a little bit more, I'm sure we could have gotten even closer. But I think, once you get to some of the next ones, you will see that it is strikingly close.

So again, this one is I think it's pretty good for the very short time that Jeff had to actually work this out. I think it's quite incredible. This one here, again, and I know the bridge might not look exactly like it. It is. But that's basically because the design changed for the bridge when they build it. It's the Rhino model that is off and not D5.

But I think the visuals, again, I think it's pretty good. Again, the trees could be different and all that. But like, generally, I think it's pretty good. And actually, for the sake of it here, I actually swapped the images. Because it's pretty hard to see what's the right and wrong one here. So now the D5 is actually on the right, and the real one is on the left, just to confuse the viewer a little bit. Because I think this one is definitely tricky.

Right, so I want to jump into to the AI part of things here and also talk a little bit about how we use AI at BIG, how we approach it. Because I think it's interesting before we jump into the AI section of what D5 does in relation to AI to sort of frame a little bit our thinking behind it. So as I mentioned earlier, this ecosystem of tools is almost like technology explosion that we've seen over the last couple of years.

And obviously, within the last year or two, we've been seeing all these different AI tools popping up all over the place in this ecosystem. And there's almost like a new one on a weekly basis coming out, which is also why it makes it hard for someone like me and others to stay on top of every single one of them. Hence, we obviously have to pick and choose sometimes and go with the ones where we have the best stomach feeling about it.

That being said, we do test a whole bunch of them. I think, in the last six months, we've tested definitely more than-- I think this slide or this information I've put in here is actually a little bit old. It's more than that now. But we've been testing a lot of different tools.

And I think it was also part of this and as part of some of this research that we also started to stumble upon D5 more and more. Because D5 is actually one of the few, you could say, software pieces here in this map that actually have started to incorporate AI features directly into their previous offering. Where a lot of the other AI offerings we see on the market, they are more like standalone, startup AI firms that are serving a very specific purpose. Where I think what's interesting about D5 is that they have already sort of started to incorporate some AI functionality straight into their tool, which I think is, obviously, really, really interesting and also why we think this is really going to change our way of working. I'm

Sure most of you are familiar with Midjourney. And this talk here today is not to talk about Midjourney, per se. But I think this image or these images here show quite well how rapidly things have changed. You know, same prompt year and a half ago, actually, close to two years ago, or actually is more than two years ago. But yeah, sorry. But again, I think it just really shows how quickly these things have increased in quality.

And for us, we've been using Midjourney and some of the other image-generation tools quite a lot since they first came out. I think we've done in the order of 40,000 different images by now on our different image-generation accounts. I think we mostly use it for mood boards, for getting inspiration, and getting new ideas, and so on. That's our primary use of it.

We have also increasingly started using it for things like, as I mentioned earlier, with these V-Ray renders I showed where we now also use it to generate cut-out people that we then Photoshop into our scenes. We also use it for more stupid stuff like clip art that we put into presentations. We-- yeah, and a number of other things like graphics and graphics for presentations and other things.

So it's also serving a purpose outside just the sheer design part of it. Even something like generating ideas for how a diagram in a presentation could look like. And we are also increasingly starting to look at, can we actually start to make use of some of these image to video generators? I believe the one you're seeing here was done with the latest runway version. So we are really looking at a lot of that as well.

And I think there's surely some stuff in this video here that you will see change from the original geometry. But that might mainly be someone like me who knows the project very well, that would pick up on that. This is, by the way, our Kistefos project up in Norway, close to Oslo, where we made this bridge across a river that forms part of this big sculpture park. So you could say the bridge becomes almost a sculpture in itself and also an extension to the existing museum. And so a project that takes three boxes instead of one.

But yeah, I think I might pick up on the details that something is changing as the video gets closer to the building. But if you mix this into a series of video clips and animations, a client or someone like who's not so familiar with the project may not notice any difference or any change. And for them, it's just the perfect video that really shows the story that we want to tell.

And we actually have recently submitted a video for a project where we did use a lot of image to video generation for the video clip of that project. So it's definitely some of these things are definitely here and will only get a lot better in the coming years to work with. So yeah, as I mentioned, we've done, I think, like 40,000 images since January '23. So that almost equals like 20 or-- sorry-- 61 images per day. I think these numbers are actually may be a little bit outdated by now.

But we do a lot of research around AI. And we also we're also constantly looking at, how can we gain more control over our outputs and more control over, you would say, the AI engines and, yeah, the output that it provides. And then on the sort of more production side of things, we are definitely using a lot of stuff like a lot of things like Magnific AI to upscale and enhance images. We're using Photoshop, Firefly a lot for small, quick changes here and there and adjustments to images.

And this is also why, for us, something like D5 that now has a lot of these AI components drizzled into their tool is making things really, really interesting. I think we're-- and this is not just us, but like the industry in general are getting to you can already prompt to 3D, as most of you probably know by now. I don't think we're at the point where we can prompt an entire project out in 3D. But at least, we're definitely going to get there at some point.

And I'm sure the Autodesk guys is probably already hard working on a lot of AI BIM-related stuff. And for that matter, hopefully, I would cross my fingers for some AI-generated documentation as well in the coming years. But at least, this is where we seem to be at the moment in that sort of progression thing.

And then I think when talking about AI, it's also important to-- or at least, I quite often refer to all the image stuff as the more sexy AI stuff and then all the text and all the ChatGPT data stuff as the more non-sexy or boring AI. But actually, I think some of that more boring stuff is actually some of the most-- maybe the most powerful stuff and where a lot of the focus actually should be. Obviously, today, we're obviously mostly talking about all the visual aspects. We're mostly on the sexy AI today. But I think a lot of the non-sexy stuff is definitely also worth spending a lot of time on.

And I think at BIG, we've sort of come up with this approach that we really want to see AI as something that aids our design process and not something that drives it. We do not want to become known for being a Midjourney architecture firm. We want to drive what we do. We want to maintain our philosophy and way of designing buildings and projects.

But we obviously want to leverage AI to the widest extent and especially where we can see it become used as a tool and something that really is, as mentioned, aids our process. And I think this is also where-- because D5 have made this statement at some point where they were saying that they believe AI is not only a tool. Or no sorry. That AI is only a tool and designers are the masters.

And I think this, you could say, blends very well with the way we see it as well. So that's really why we think there is a good fit between what they do and what we do. And then touching upon how they came into the picture, it was really when we started realizing these sort of features like AI atmosphere match. You can put an exemplar image into it and then get some of the same look and feel into your scene, AI-generated material texture maps, and Ultra HD textures AI generated.

And also, this makes things seamless. So you don't get the seams between textures. I think that's also just something I never thought about that as being a feature. But when I saw it, I was like, that makes so much sense to have a feature like that and especially with tools like AI on the market these days. So I think it's really well done.

And then lastly, they are also now incorporating an AI enhancer. I maybe would quickly say that I still think that something like Magnific probably has a little bit of a better output as of right now. But the fact that a firm like D5 is now incorporating something similar into their pipeline means that I don't potentially have to buy both Magnific and D5 in the long run.

Because that's also one of the things, for me, as, you say, someone in charge of all the software and all our budgets around software, for me, having all these different small AI applications that I constantly have to pay subscriptions for and remember passwords for and how many seats do we have here and there and so on like the administration-- sorry-- administrative side of things when it comes to all these accounts is just completely blowing up to a point where it gets too much. So that's why, from my point of view, as an administrator of all these softwares, I think getting some of these features built into our existing tools or some of these new tools just makes a lot of sense. So I'm obviously very hopeful to see more of the more established software companies take the same approach.

So yeah, here, just quickly showing that cycle of things that you can now do with some of these D5 capabilities when it comes to AI. I'm sure Jeff will show some of it in a second. So I'm not going to take anything away here. So again, I will hand over to Jeff in a second to show some of that.

I just want to quickly frame the project we're going to show now. So this project is a project we're doing in Kansas. This is the School of Architecture in Kansas. We're making an expansion to their, yeah, the school there.

So we're envisioning this timber building, this cube, makers cube, which will be predominantly made out of timber on the inside. So here's just some of the official renders we did for the project recently, just to show you what we're envisioning. And we are already into looking at making prototypes for how can we actually put it together and get all these complex joints of timber to look as sleek and nice as possible. And as you can imagine, it's going to take some quite serious wood craftsmanship to get those corners to look completely perfect.

But yeah, I'm going to hand it over to Jeff now to show a few AI-related examples on this project here.

JEFFREY ESPINOZA: All right, yeah, so here we have that Kansas campus project. And so again, we have a whole bundle of AI features within D5 Render. And hopefully, these are just meant to speed up your workflow, get rid of any repetitive steps in your workflow, not really to take away any authorship or design decisions.

And so one of those features here is this AI atmosphere match. So all you really need to do is just click on this icon and upload a reference image. And this could be an image you found online or maybe your studio has done in the past. And so if you just want to go ahead and reflect that same mood into your current project, you'll just upload that reference image, take a screenshot of your current project, and then hit Start Generating.

And so again, hopefully this will alleviate any tedious tasks of having to play around with a bunch of different sliders in the Environments and Effects tab, just trying to figure out what each little button does. And again, using this method, it's also pretty simple to create studio standards, studio presets, onboard new users. So that again, you don't have to mess around with all these little hidden features within the Environments and Effects tab.

And so here is what that final result looks like. And so as you can see, our AI model was able to pick up that this is a winter scene. So we went ahead and added some weathering effects in here. But since it didn't go ahead and play around with all my different settings, I can always go in here and tweak it just a little bit more, so that I get that exact look and feel that we were looking at our reference image.