Chaos: Doing the work AI won't

RODERICK BATES: Hello, and welcome to Doing the Work AI Won't. We're very happy to have you join us. And my name is Roderick Bates, and I'm the director of Corporate Development here at Chaos. And I'm joined by my colleague Ana.

ANA LYUBENOVA: Hey, everyone. My name is Ana. I'm a product manager at Chaos, and we're excited to host this AI class today.

RODERICK BATES: Absolutely, we are excited. And we're really talking about something that we think is incredibly exciting because AI technology does get all the attention, but at the end of the day, it's not necessarily delivering everything that people need when they want to communicate their design.

So here's the abstract. Probably some of you have read it, found it exciting. And we're going to dive into the details that make all of this possible by addressing these key learning objectives. We want to show you how you can create incredible visualizations from your architectural design project that combine both the hard data of building performance with the beauty of the actual design intent.

We also want to show how something like energy modeling, that's not normally associated with things that maybe are on the aesthetic side of design communication, can actually be folded directly into your design visualization workflow. And then we also want to show how someone that maybe doesn't have a high level of expertise can rapidly generate crowd and vehicle animations based off of Revit model geometry.

These are all things that we see as being incredibly important, this combination of data and visualization, to create really a narrative or a story with your project that right now AI isn't quite ready to be able to do.

But that isn't to say that AI can't do a lot. There's a ton of things that it can do and a ton of things it can do incredibly well. Like let's say perhaps I wanted to ask, what was the best render engine out there. Maybe it would tell me it's Enscape, and that's a really nice way of just getting right down to the data. So for sort of a streamlined search, it really does incredibly well.

And chatbots are something where maybe at one point it was just an interface for, say, customer service you might find on a website. But more and more people are starting to find ways of integrating it directly into software, maybe even generate prompts, really start to use it in a way where it's like asking a very smart person in the room complex questions, and they're going to give you an answer far more quickly, really, than any other methodology.

And of course, complex analysis, the ability for AI to take a massive, very nasty looking data set, and be able to through it incredibly quickly to derive insights. I'd say in many ways, it's almost intimidating how quickly and capable AI is, and whether those are visual data sets or numerical data sets. I think we've all seen examples of a lot of the complex analysis that AI can do.

And one of the things that we find particularly interesting in the world of architecture is that people are using it as a very rapid mechanism for assessing code and building code. Questions like, well, I have all these parking spaces, but how many of them need to be, say, van accessible for, say, ADA compliance, something along those lines?

That's something where AI maybe shouldn't be your final point of interrogation, but it's an incredibly effective way of having that question answered quickly as opposed to, say, flipping through a large book of codes or trying to search for PDFs. And then oftentimes, you'll see here on the left, it'll link you right to the source document.

And of course, image generation. I don't think there's too many design firms out there that aren't at least using AI for the creation of mood boards and other sort of preliminary information. So maybe you wanted to create something that's more button-down. It's perfectly good for that.

Or maybe something that's absolutely outlandish, the most wacky thing you can think of, like a cat on a unicorn juggling cupcakes, and maybe that cat's an astronaut. No problem. AI has your back, and it'll create exactly that image, but there's a lot that AI can't communicate.

And one of those things that it really isn't there for is about emotional resonance. It doesn't create that emotional connection. It can't give you a hug, if you will. It doesn't have empathy. It doesn't understand you as a human or the people you're trying to communicate with as humans, and part of that is about context. It doesn't always understand the nuance of a particular situation.

And I think one of the other things that is incredibly challenging with AI is that it doesn't know when it's wrong. And you can see here, this example. It's like, you ask how many letters are in the word "nineteen," and you get back "nine letters."

It's one of those things where I think it makes it difficult to really trust it at a certain level, and it also shows that there's still a place for humans to have an involvement in the process when it comes to using or leveraging AI.

And the other thing that's interesting about artificial intelligence is that when it comes to visualization and creating visualizations, it's really not doing that with accurate data. It's doing it based off of a pure pixel-based approach.

It's not thinking of it from the perception of, well, will this perform well from an energy code perspective? Or can people easily move in and out or around this particular building? It's just saying, here you go. It's neither the right or wrong. Here's a strong visual that we think meets your prompt. And of course, along with that too, is it doesn't have that human emotion, that creator's artistic touch that is so integral to the design process.

And why is that a problem, at least from our perspective, and we think from the perspective of the industry as a whole? And maybe we can go back, go back a long ways, to Aristotle's triumvirate of really what makes for powerful rhetoric. And that's the idea that you need to balance multiple elements here.

You need to have emotion, and you need to appeal to your audience with emotion, but you also need to appeal with logic and reason, these hard numbers and data and things like that, and really combine the two of those from a stance of credibility and ethics, one in which people really trust you and they know that you're unlikely to be wrong, or if you are wrong, you're going to want to correct it.

And so that's the idea of this balance here between having an argument with emotion that appeals to the audience's emotion or to the receiver's emotions, and feelings substantiated by logic and reason, and then coming together from a very credible source. And that's really, from our perspective, one of the challenges that AI has. It's not necessarily combining all of these elements that are really required for a powerful and compelling argument.

And why is some of these components, particularly the emotional component, so important from the perspective of an argument? It's because you have the power to change, really, when you start to combine emotion and data. And from our perspective, that's really what makes for storytelling, when you put those two pieces together. And you can start to really influence how people think, because oftentimes people are not making decisions based purely off of data, but oftentimes it's really that combination of emotion and data.

And so when we look at these examples, these are our books that are well known or movies that are well known, and they really did implement change. And while they were substantiated in data-- if you think of say, The Jungle, it was actually going out into the field doing investigative journalism, understanding what's actually happening on the ground in those meatpacking facilities around the turn of the century-- it's a combination also of the really compelling story that galvanized a nation and implemented change.

Or if you think of An Inconvenient Truth, it raised awareness around global warming. And sure, there was a lot of data and graphs in the context of that movie, but it was really about a narrative and a compelling story that was put together that caused it to be so successful.

Silent Spring was one in which people really understood the implications, the environmental implications, of DDT in a way that was emotional. And combining that with the scientific data that was emerging is what really frightened people and caused them to see that the implications, particularly with birds and the implications there, to see declining species counts.

And then with Bambi, I don't think there's anyone that didn't see Bambi and felt their heartstrings pulled. And that's something that generated quite a bit of anti-hunting sentiment that honestly, it's probably still felt today.

So architects, in many ways, are storytellers. They're ones that do need to appeal to an audience and communicate to them in a way that's compelling and actually really is driving their stakeholders, or project stakeholders, towards a particular decision-making process and outcome. And they are visual storytellers that use data, but the problem is, unfortunately, that sometimes the data they use isn't particularly compelling.

As you can see, these examples, we look at, say, a drawing set. Given what type of stakeholders are out there on a given project, say like a client group, this type of information maybe isn't always the best to provide the emotional aspect, if you will, and the communication capabilities that are required for those that maybe aren't trained in the field of architectural design.

And so oftentimes they communicate using illustrations and visualizations. And that maybe fits in more on an emotional level, but unfortunately, you also have to have a good visualization in order to do that. And as some of these examples we pull show, the visualizations aren't always the best. And when you do that, you're not necessarily communicating on the type of emotional level that you want, and you're losing out on the power of storytelling here by having data that people don't understand, as you saw in the previous slide, and here perhaps with an emotional message that just doesn't resonate with the audience.

So as visual storytellers, I think designers really have an obligation to think about how they combine data and the compelling design visual of their project to tell a compelling story. And fortunately, we're not the first industry that's been confronted with this challenge.

What's fascinating, at least from my perspective, is that in the world of science, they've recognized the need to visualize data in a way that is exciting for audiences that go beyond just the scientific community. For the scientific community, maybe looking at a scatter plot or a raw image right out of a simulation engine is exciting, but they know that that's not the only consumer of scientific data.

They actually have to communicate with those maybe that are responsible for funding or maybe need to make large-scale decisions based off of the data they're generating. And so they're realizing that they need to look at other ways of communicating the same data that are accurate, substantiated by that data, but at the same time also begin to bring in the visual and emotional component of storytelling.

So here's an example that I think is pretty effective, but perhaps one of the entities that's been most effective at this is actually NASA here in the United States. And they are very comfortable using video images for a variety of different topics, where they are showing all sorts of complex data-- weather systems, satellite movements, things like that-- and bringing all of this into interactive visualizations that can be understood really by anyone, whether you have a background in science and regularly report to work at NASA, or you're someone like me that perhaps has an interest in these topics, but certainly not the type of scientific background that would typically be required to understand the raw data that is underpinning these types of visualizations.

So when we think about this type of visualization process, those types of projects that are happening on a scientific level are happening at the level of the architectural design process. There are architects and designers and visual artists out there that do understand the importance of that type of visual communication, but the challenge is how do you combine something like you're seeing here with data?

And that's something that we're very keen to do, is to show you ways in which you can do exactly that. You can perform an analysis or maybe you're trying to communicate complex information, say, like energy analysis or occupant movement throughout a building or around a building or traffic flows.

How do you combine that type of information you may be wanting to communicate, and maybe move beyond just pure diagrams or graphs and charts and things like that to actually make them part of the visualization process? And for us, that is exactly what we want to show you. Essentially, how you can have storytelling workflows that combine data and visualization using Chaos products.

So maybe we'll start with one that isn't normally associated with a story, or at least a compelling story, and that's energy analysis. Energy analysis, obviously, is incredibly important. This is something that is both a hyper local problem-- it matters when it comes down to how much you pay in an energy bill on a monthly basis-- but of course, it's also a global problem, when we think about climate change and global warming.

So our response to that has been Impact, an integration with our Enscape modeling platform that works across any type of design platform you might be using, including Autodesk Revit. It's something where we wanted to combine the ease of use of Enscape that our customers know-- and of course, it's one of the key reasons why they use the software-- but to retain that ease of use in the context of an entirely new workflow for energy analysis, so they can really understand how their building is performing from an energy use perspective as they design.

And we wanted to make sure that the analysis that we generated was something that our customers could trust, something that they could look at and say, is this something that is getting me on the pathway towards compliance? And we want them to feel confident that the answer is yes.

And so to do that, we partnered with IES out of Scotland. And IES is one of the world leaders when it comes to sophisticated building performance simulation software that's used by the top engineering firms globally.

So what do we mean by ease of use? How easy is it? Well, let's just say you happen to be in San Diego. And if you can answer that question, you've answered one of the first of the three questions that are required to perform an impact analysis.

Second question is, well, what type of project are you working on? Well, in this case, let's say we're working on a school or university here in San Diego. In what year was it built? Well, let's go with 2012. I would hope you know the year that your building was built, and of course, along with that, when it was last renovated. Or maybe it's new construction, but you should know the date for that.

And if you have all those components, amazingly enough, you can actually run an impact analysis. So you can see here a user going through the process there of inputting the information relative to those three primary questions. With those three questions and along with the data that's already found with the Revit geometry that you've created that Enscape is able to pull, we're able to generate an initial analysis that understands energy use intensity, peak loads, and energy use allocation.

And the great thing that all of our users love about Enscape is that you can make changes to your design in real time, and you can see the updates in the visualization interface of Enscape to see exactly how your project's going to look as you change the design. And now in the same time, you can see the updates in the building performance of your-- or the performance, rather-- of your building.

And to improve the communication of the data and move beyond the charts and the diagrams, we've integrated the building performance results directly into the render view, allowing you to navigate that analysis and see those results and find out which rooms are performing well and which rooms might not be.

And with that information, you can create a compelling visualizations-- if it'll allow me to move forward-- compelling visualizations that are able to combine both building performance and traditional architectural visualization, creating a story of your building that has these two elements brought together, allowing for all stakeholders-- whether they be architects or engineers or client groups-- to understand these two facets of your building simultaneously, within the same interface, crafting a narrative that combines both the data of building performance with the analysis brought to you by Impact.

And a platform like this is really only as effective as the workflow that it supports. And in a context like this, this is something where we wanted to allow for our users of Enscape, regardless of where they may be starting their design process, to be able to seamlessly transition into Impact.

And from our perspective, we've created a workflow for storytelling, storytelling around energy analysis. But as my colleague will show, that's not the only story that designers need to tell.

ANA LYUBENOVA: So our next topic will be about storytelling through creating crowds, occupant movement, traffic flow, et cetera. And for this, apparently, if you're working in a CAD system, you are very limited, and basically you cannot create any of these.

So we know that many of our users have been requesting this type of functionalities for quite a while now. And looking into how to address those functionalities, we actually end up creating a new product.

So I'm very excited to share with you the newest product on our portfolio, Chaos Envision. This is basically the first time that we are presenting it in public. And it's been in a private alpha for the past few months and will soon enter in a public beta.

So Chaos Envision is an independent, real-time 3D assembly and animation tool for architectural visualizers wanting to either explore design options or produce compelling animations that tell the story of their design in the best possible way.

And I know that you guys are probably curious about how does it look like. It basically looks like this, and we're going to explore the main workflows and the key capabilities that this new Chaos tool provides to you.

And we'll start with key capability number one. Chaos Envision allows you to assemble scenes coming from different 3D tools. And in doing that, it is also able to handle ultra large scenes, which sometimes is not the case with other real-time tools.

So for example, if you have your Revit model, like your design being built in Revit, but you have the terrain in 3ds Max, or you have the building envelope being built in Rhino, you can combine everything using Chaos Envision. And let's see how it works.

So we have this Revit model that was rendered initially with Enscape. It's actually the same model that we already saw with the building performance functionality. So what we need to do is to export a V-Ray scene file, which we then import into Chaos Envision.

And then we have the terrain coming from 3ds Max. We assembled both scenes together. And then, as this is a visualization app, you can create cameras, you can modify the camera parameters, and configure various settings.

As you can see here on top, we have the cameras and variations that we're going to review later on. And here, I have access to all the scenes that I imported. And I can manipulate each and every object that is coming from those scenes.

I have a list of all materials. So basically, my Enscape materials are read and rendered by Chaos Envision. And in addition, I also have access to the Chaos Cosmos library of 3D assets. And here is the timeline, which we're going to use in order to create our first animations.

So let's explore a very basic workflow of assigning materials. But as I said, your Enscape materials are actually supported and rendered, so Chaos Envision would aim to simply enhance their look by adding some additions to them.

So you can simply pick a material, and then you can either add enhancements to it or modify its parameters. Like in this example, we modified the reflections of that material.

All right. And then, in case the Enscape design contains Entourage, it will also get rendered. In that particular example, we decided to demonstrate how you can also add Entourage from within Chaos Envision, because as I said, it is directly plugged to the Chaos Cosmos library of 3D assets.

And with the advanced scattering tool, you can scatter as many assets as you want to. In that example, we scattered this grass asset. And if you can see, the performance will not drop by any means.

And then you can also play with the lighting. There are a couple of options for environment lighting. Number one, of course, is the sun, but there is also a possibility for image-based lighting.

Yeah so, the main purpose of Chaos Envision is to let you create advanced animation, from objects being animated on paths, through crowds, phasing, or design transitions. And Chaos Envision combines 100% ray traced environment as well as the most realistic Archviz entourage delivered by Chaos Anima These are 3D and 4D animated assets.

And let's see an example of how this works. So in this first example, we are going to generate a crowd. You simply need to place the crowd and then draw the path where the crowd characters will be walking, and then select the characters from within the Anima asset browser.

And the cool thing is that you can generate a crowd even out of two or three characters because you can increase their count, but at the same time you can vary the color of their clothing, so that it looks as if these are different characters. But in our example we, of course, use more characters, and you can adjust the parameters of each individual character. And then by pressing generate, the crowd just gets generated. And here on the timeline, we can check the results.

The next example focuses on traffic. So as you can see here, we have already drawn the path, the blue line. And we simply import a car, and we link it to the path. And then we can adjust the car speed, and now the car is linked to the path. And basically, it controls its direction of movement. And we can add more cars, or even a train in that example.

Yeah, and Chaos Envision allows you to generate an unlimited number of design variations. And they can be used for various purposes, from design explorations to creating animations with the help of the variations.

So there are two types of variations that you can create with Chaos Envision, one that are only related to the lighting, and another one that are related to the objects in the scene. And both of them can be combined together, as we're going to see in this example.

So in this example, we first have a camera that is being animated, as if it's moving. And we have the timeline. This is where we actually create the animations. And here, we created another variation of the environment lighting. So we changed the sun position.

And we're going to animate through both the environment variations. And the final result will look as if the sun is moving. We simply need to place them on the timeline, and adjust their duration, and then check the result, and adjust the timeline blocks, if necessary.

And it works similarly with scene geometries, as I said. So in this example, we decided to simply change the position of some trees. But of course, those objects can be-- well, it could be any object. It could be either moved or rotated or hidden, and this will only be saved in the currently selected scene variation.

And with that, again, you can create as many variations as you want to. Like in this example, we created quite a few of them.

And Chaos Envision is compatible with the rest of the Chaos products. So it basically allows for seamless data transfer from all Chaos products that support the V-Ray scene format, meaning that as long as you're able to export a V-Ray scene, you can import it in Envision, and you can start creating your animations.

And the cool thing is that Chaos Envision actually combines all the Chaos products that you see on the current slide. It utilizes the accurate, real-time render engine of Chaos Vantage. So basically, chaos Envision and Chaos Vantage are sharing the same render technology, the same render engine.

Chaos Envision allows for visualization or scene data created with either Enscape or V-Ray to be imported and continued for the purposes of creating animations. It also provides direct access to the Chaos Cosmos library of 3D Assets, as well as direct access to the Anima library of animated people for crowd generation.

And in addition to that, it has integration with the Chaos Scatter, which exists as an addition to some of our V-Ray plugin integrations, for scattering vegetation and any type of organic creations.

And now, again, all the key capabilities at the glance. And let's see how our final video looks like.

And as I mentioned at the beginning, Chaos Envision has been in a private alpha for the last few months, and we are now getting ready to launch its public beta. So whoever of you is interested to give it a try and explore its functionalities, you can simply scan this QR code and apply for the beta.

And with Chaos Envision and with our building performance feature slash extension available in Enscape, we are getting one step closer to our company mission to democratize visualization and to provide easy-to-use visualization workflows tailored to the needs of our customers.

And with that, our presentation has come to an end. And we would like to thank everyone for joining our class, and see you again.