Adventures in Convergence: Sustainable Design, Data & Procedural Modeling

TYSON FOGEL: Welcome, welcome. We hope everyone is doing well. Thank you for joining us for our industry talk at AU 2021 entitled Adventures in Convergence, Sustainable Design, Data, and Procedural Modeling. We're coming to you from the Autodesk Technology Centers. I'm Tyson Fogel. And I'm joined with my colleague and friend Matthew Spremulli. How are you doing, Matthew?

MATTHEW SPREMULLI: Pretty good. Thanks, Tyson.

TYSON FOGEL: Amazing. We'll get into our agenda, our backgrounds, and a little bit on what the technology centers actually are shortly. But first, let's go on a conceptual journey. And let's unpack this very much loaded title. Adventures in Convergence-- what does this mean and how does this apply to you?

To answer that question, it's best to start from a place of design intent. What is good design? And what is poor design? And what impact does it have?

Well, what if I told you that pollution and overconsumption are a byproduct of poor design. This is what Matthew and I set out to explore. It is our guiding question and very much part of our hypothesis. It's something you should keep in the back of your mind as we go through this talk.

To help ground this statement, let's explore what our poor design is costing us. Americans could save approximately $40 billion each year on electronics alone. And that's if they could repair them. Instead, electronics are tossed and new ones purchased. Or currently, 8.6% of the world is operating as a circular economy, meaning the remainder percentage is operating underneath the assumption that we can continue to take, make, and waste.

Or this other stat-- $0.20 of every dollar spent on manufacturing is wasted due to inefficiencies. This is a pretty startling margin. But this waste isn't always physical. It's economic waste-- things like redundancies and prototyping or design and manufacturing changes that cause downstream effects. Or this last stat-- adults make approximately 35,000 decisions daily. If only 1% of this is spent on our consumer products we use every day, then that's still 3,500 touch points in opportunities to make a difference as individuals.

But this is a small sample of design's impact. However, it starts to paint the big picture. It tells us there's a sense of urgency here. And there's an inherent responsibility. There's a demand to do better.

We need to shift in our design mindset. We need to be able to make things better, smarter, and more efficiently. And that's what sustainability is all about.

And convergence could be one path forward because convergence is a mindset, one that encourages others to gain a new perspective on how to solve problems. Convergence means leveraging best practices from other industries and domains and encourage these individuals to gain a new perspective to solve their own design problems. To learn more about Autodesk's position and how it's enabling convergence, we recommend checking out the website shown.

So now that we have a working understanding of the problem at hand, what does this have to do with procedural modeling?

MATTHEW SPREMULLI: Yes. What does this have to do with procedural modeling? I'm glad you asked, Tyson. As already mentioned, the converging or mixing of tools, workflows, or even personas and industries is becoming one way to explore complex problems. And one way convergence can manifest is to leverage technology from one industry and apply it to another.

What we're going to be focusing on for the duration of our talk today is how specifically the procedural modeling technology from media and entertainment, which is the same technology used to mutate this donut here in the background, could be leveraged to make industrial designs more functional and/or sustainable. To interrogate this question, we wanted to work with others who are also interested in exploring these kinds of questions. So we brought together architectural and academics and professionals using media and entertainment technology and tools to make product designed and manufactured objects. And we supported them in an intensive design make workshop.

But we also had our own ideas and approach to answering this question. And we wanted to explore this in a complementary process to an external team that we could work with. There were three parts to this approach. First, we wanted to brainstorm a set of scenarios of how we thought procedural modeling could positively impact industrial design.

And to do this, we based our brainstorm on discussions with individuals and reviewed numerous industrial design case studies. Second, we wanted to stress test these scenarios with our invited external team and other subject matter experts. And we did this through surveys and interviews. Lastly, we cross-referenced everything and aggregated our observations and lessons learned.

This two-pronged approach is summarized in the following diagram. To answer the question, can procedural modeling be used to make industrial design more sustainable? we invited collaborators on the left to conduct a design make workshop, which was paired with our own research on the right. And the intersection between these two streams was our surveys and interviews.

In the end, we aggregated everything together to get ideas on how procedural modeling can be used to make industrial design more sustainable. In this manner, we were able to get both on the ground experiments and some higher level thinking on the subject. Both groups, our external collaborators and ourselves, chose to look at a simple industrial design object as their focus for activity, including thinking about materiality.

We chose a chair for our own ideas. So look out for that as a recurring object in this presentation. For the rest of the talk, we are going to walk you through the following four subsections. First, we'll all settle on what procedural modeling even is. Then we will cover how this hypothesis was explored within Autodesk, namely through the technology centers, and the collaborator, which we call residents.

We will spend some additional time unpacking our own research approach that complemented the work of these residents. And finally, we will share lessons learned from these activities.

TYSON FOGEL: And with that intro, I am Tyson Fogel, workshop supervisor for the Autodesk technology centers based out of Toronto. I'm an avid maker, Fusion 360 user. And I love designing and building my own furniture. I work directly with innovation communities in my role to provide a level of technical expertise and fabrication consultation. And I try to do all of this through the lens of sustainability. Over to you, Matthew.

MATTHEW SPREMULLI: Thanks, Tyson. And I'm Matthew Spremulli, development manager also with the Autodesk Technology Centers. And also based out of Toronto, Canada. I help shape projects and research activities within the Technology Center community, scouting for new members to join, and helping them pursue deeper collaborations. I'm officially trained as an architect, come from a background in manufacturing, and have explored telling speculative design future stories in both my research and teaching.

Great. So now that we're all introduced, let's dive in. Some of you might be asking yourselves, what even is procedural modeling, especially if you come from a product design or AEC background. Well, in general, procedural modeling is considered an umbrella term for several techniques in computer graphics to create things like 3D models, textures, effects, such as particles or explosions, graphics, and animations. All of these are based on sets of rules or, as the name implies, procedures.

These sets of rules can be embedded into an algorithm or even made configurable by exposed parameters. A major advantage of using procedural modeling is to help create complex scenes through small simple recipes or scripts that in turn the user uses to create more of an art director experience, tweaking and refining the ingredients to get different outputs. A series of simple examples of using procedural modeling to create assets as shown on the right-hand animation, created in Autodesk Bifrost, a visual programming language integrated into Autodesk Maya.

It is interesting to note that procedural modeling is not a new technology. In fact, the first published book on the topic was authored in 1994. The same year Ace of Base's "I Saw the Sign" came out. Since then, procedural modeling has mostly been adopted, employed, and developed in and for the media and entertainment industry, mostly to help automatically create assets and behaviors seen in movies or video games. And since 1994, many technology companies saw the sign that this would be a great technology to automate workflows and created numerous tools based on the technology.

The specific tool that we explored for this talk, as already mentioned, is Autodesk Bifrost. So wait, what exactly does procedural modeling do again? As a quick recap, some examples include effects such as simulating cloth or aero behavior, or creating models that can accrete, grow, or even die-- kind of like what you're seeing in the image in the background here.

As a quick anecdote, we found that after discussing numerous tools of transdisciplinary designers in our experience, we found that they had a great way of describing procedural modeling technology. Quote, "I see tools like solid modeling as extensions of my hands, whereas procedural modeling tools are extensions of my mind," end quote. What they're saying here is that procedural modeling allows a designer to imagine new forms of complexity and behavior that would have been previously impossible to dream up.

Before we go any further, I wanted to mention that experimenting with technology sometimes requires the right environment to explore such things-- a Sandbox, if you will, that supports and encourages experimentation. Well, that's exactly what we're trying to enable with the Autodesk Technology Centers. The Tech Centers are a division within Autodesk research. However, we conduct our research not by making our own projects, but rather by supporting and observing others in their work.

This is what we call the outside network. It's a global community comprised of resident teams from industry, academia, and the startup sectors, all exploring a vision on the future of making. The program provides these resident teams with several resources at no cost in exchange for our team to observe their work as case studies to get a glimpse of where the future is headed. If you'd like to learn more, please visit autodesk.com/technologycenters.

Now, the resident team that we worked with as our external collaborator on this topic was the Architectural Association's Visiting School program, or AAVS for short. For those who are not already familiar, the AA is a world-renowned architecture and design school located in London, England. The school is notorious for exploring new forms of media and technology, especially as it applies to the future of design make.

Outside of their London HQ, the AA also hosts a global program comprised of intensive two to three week post professional programs on niche or fringe topics. The visiting school program was the ideal vehicle to explore our specific question and topic because it's not part of their typical design curriculum. And we work with some of the visiting school directors to craft a specific workshop this summer.

The result was a program entitled morphological experiments between force and form aimed at mixing technology together as it applied to industrial design objects. We worked with the AA directors to recruit five tutors affiliated both with the university and the AEC industry across Toronto and Boston to lead their own interpretation of this title in designated groups. And in turn, they worked with 24 students across 16 different countries.

TYSON FOGEL: Amazing. Now that we've level set on what, who, where, and when, let's get back to the how and how we approach the question of, can procedural modeling be used to make industrial design more sustainable? If you recall, we had a three-step process. Step A, the brainstorm-- we tried a bunch of different things here and did a lot of whiteboarding. But essentially, this step for us was trying to work forwards or backwards from sustainable design as a concept to arrive at some type of connective tissue or interplay that links back to procedural modeling.

We eventually landed on five design strategies that bridged procedural modeling and sustainable design. If you recall, much like our external counterpart, AABS, we grounded these design strategies by applying them and imagining them in the context of an everyday object. Now, before we unveil what those design strategies are, it's worth mentioning the strategies we imagined are impacted by two competing cycles. And these may look pretty familiar to you. But these cycles are intrinsically connected through the process of manufacturing.

So we created this illustration for each scenario to help you better understand where the value of procedural modeling is being realized. If you are unfamiliar with what these cycles are, the development cycle is a Sandbox all makers and designers find themselves in, whereas the product life cycle is a process by which materials go from their point of creation or extraction through to disposal. Now, if this green loop was closed, this would be a good representation of the circular economy.

Finally, our five design strategies where we imagine procedural modeling can play a role with sustainable design. And these are biomimicry, design for afterlife, design for repair, manufacturing modifications, and design for awareness. Now let's take a closer look at each of these. First on deck is biomimicry.

So this is using procedural modeling to emulate a form or process as observed in nature. Now, what does biomimicry have to do with sustainability? Well, mother nature has proven its resiliency and ability to be circular by existing for thousands of years. So a lot can be learned from the world around us. And for our chair, this could be akin to mimicking a creature's skin to increase durability.

Another great example of this in industry or in product design is by a company called Whale Power, who creates these wind turbine blades based on the tubercles of whale flippers to minimize drag and increase arrow or hydrodynamics. Now, if you shift your focus back to our development and product life cycle, the value of procedural modeling is best realized during the ideation and usage stage in this scenario.

Next up, design strategy number two-- design for afterlife. This is the use of procedural modeling to enhance functions like decomposition and effectively ensuring the transformation of a product to provide nutrients to an environment during its afterlife. So for our chair, this would mean producing and stimulating the growth of biomaterials like coral. And when discarding these, these objects encourage the growth of an ecosystem.

An industry example of afterlife would be the P-Wall designed by Matsys in 2009. And the images here show a bubbled wall panel or bubbled wall panels. And these were designed to provide habitat to a range of creatures over a lengthy period of time. And again, shifting our focus to the development and product life cycles, the value of procedural modeling in this scenario is best realized and manifests during the actual product waste stage. And consequently, material extraction can be acquired if the scenario plays out the way we had match it.

Number three, design for repair. This strategy speaks to the ability to keep a product in service for as long as possible because if we can repair or replace components in our designs, we are better equipped to prolong their negative environmental impact. But part of the challenge here is knowing when or how to replace certain components. And this is where we think procedural modeling can play a role. It could indicate wear and usage.

If you look at the chair, a rippling pattern on it could reveal new colors as it deteriorates. This could create a visual aid for users indicating a need for replacement. This is very similar to another example you may have seen floating around by a company called Discolor Tire.

They created this concept of indicating tire wear on tire treads. And in the context of our product and development cycles, the value of procedural modeling is being realized during the validation stage. Now, that may feel counterintuitive. But let me explain. Knowing how and when your products are being used is invaluable information to any designer.

Number four-- manufacturing modifications. So the way we first imagined this was the intentional use of procedural modeling to limit a high intensity factor during fabrication. A high intensity factor could be something like energy, time, or material used. So effectively optimizing the manufacturing process means we can limit the amount of areas where waste is being generated.

And for our chair, this could mean using procedural modeling to develop complex lattice structures to provide efficiency in the manufacturing process without sacrificing any level of integrity. A good example of this in industry is by a small company called Metafold. And they also happen to be a Technology Center resident.

Metafold is able to create scale defined geometries in their work-- things like lattice structures, microstructures, and metamaterials. And on the right, our development and product life cycle for this scenario, we see the value of procedural modeling best being realized during the manufacturing stage. Our final design strategy is that of awareness and cultural storytelling.

Consumers and society are oftentimes disillusioned or disconnected from how something got to its final form as well as the amount of effort or resources it took to get that thing to that point. And this strategy tries to combat that. It tries to remove this disconnection and alter consumer perception.

It eludes to the ability to encode certain data like energy consumption into a design using procedural modeling. A good example of this an industry is this bench, entitled Idea of a Tree. And this bench was wrapped, or manufactured by being wrapped using prepreg fabric infused with resin.

And this whole process of wrapping is powered by solar energy. So in the bottom image, the bench you see, the green striations indicate points of high or low energy to tell a story of embodied energy during the manufacturing process. And in our development and product life cycles, we see the value of procedural modeling manifesting during the design and usage stage.

And that wraps our design strategies, which brings us to the second phase in our approach-- stress testing our ideas through surveys and interviews with the subject matter experts because on paper, these ideas sounded amazing to us. But how would they be received by others? And to find out, we decided to expand our thinking to learn more not just about sustainable design and procedural modeling, but how procedural modeling was unfolding in the field. A quick plug here and a special thanks to those that were involved in this process-- you truly made it interesting. And we learned a lot from you.

MATTHEW SPREMULLI: Awesome, Tyson. Well, I'm excited to dive into the final part of our presentation, which is to share what we observed both from the design make workshops and the survey results. First, for on the ground results, we wanted to share some of the output from the students working under the guidance of the tutors. Both tutor led groups presented their findings and results this past summer in an Instagram Live takeover style review.

The session was attended by distinguished guests from both AEC and product design and manufacturing and was viewed by approximately 30 additional guests who tuned in. The first group, they worked on a project involving agent-based simulations, reacting to simulated forces on their chair designs. The agents, modeled after Craig Reynolds' bird flocking simulation, were attracted to areas of forced concentrations. And in turn, became the pathways for 3D printed filament or resin layers. Procedural modeling here was key to creating the interaction between the agents and the forces and seeing how they could create form together over time.

The second group worked with a digital clay slumping simulation tool that mimic the material properties of clay being deposited from a robotic printer. The attributes of the clay such as weight and sand were taken into account. It would impact successive layers printed atop. Procedural modeling allowed this group to create such a simulation that they stated, quote, "had never been done before," and allowed their group to explore new designs inspired by the manufacturing constraints exposed in the simulation.

Outside of the student work, we observed some of the following metatopics in the surveys. First, nomenclature was a challenge. Interviewees had different definitions of what procedural modeling was. And we found that the term sometimes was confused with other terms like computational design or visual programming.

We thus felt that designers, either from AEC or product design, may not be as willing to use procedural modeling tools because there appears to be some barrier to entry due to either terminology or confusion on what the tools actually do. Most interviewees stated that even though procedural modeling tools gave them access to new ways of thinking, there were still some cumbersome things to get over because of either syntax or UI/UX.

Some of the highlights here is that the current tools are a little bit difficult to capture design intent or teach in standard AEC or industrial design education because they require more time and immersion. Finally, it's not as easy to share things like code script or samples due to versioning or language preference. However, despite the challenge and/or confusion of nomenclature and syntax, there was clearly still a very strong allure and draw to using such technology. And we wanted to know why.

Interviewees mentioned that they specifically use procedural modeling tools because they can conveniently pull together several sources of data and information into one place, which is something they can't easily do in other tools. Four major such features highlighted in the interview include one, behavior, such as rigid body dynamics, particle collisions, or agent-based behavior. Two-- material properties, such as the ability to create anisotropic materials things that are directionally dependent. Three-- time, seeing things like both behavior and material properties play out, modifying or changing either form or behavior over time.

And finally, scale-- working at the extra small such as the nano or micro or even the extra large, working on large urban patterns. What was common with scale is the ability to work with large amounts of data and polygons.

TYSON FOGEL: Now, so this next observation, applications equals investment, is very much a nod to the why of procedural modeling and these features that Matthew outlined. We notice that all applications and behaviors boil down to four things-- the ability to extend beyond perceived limitations or the ability to design, simulate, or animate something that is impossible to conceive, the ability to have a sense of wondery and curiosity embedded into a design, or the ability to have flexibility and insight into how behavior influences other factors within the design.

Now, this procedural complexity though, it comes at a cost, with many interviewees noting that the learning curve here is a massive one. It's a huge time and money investment. But integral to their creative practice and identity, to them this investment was not optional. It was an obligation, as they cannot find other tools to experiment with that meet this demand for control.

And if we take this a step further, these distilled observations speak to the complex relationship between functionality and sustainable design, with many interviewees looking to leverage procedural modeling to effectively hack materials. From an engineering standpoint, this allows them to add mechanical strength for the purpose of making inherently weak and abundant and isotropic materials, and manipulating their layout to gain strength, as Matthew outlined, or to add surface articulation for the purpose of elevating a different material attribute, or to explore some level of advanced material or process simulation for the purpose of seeing how a material or react under certain environmental conditions. It's fair to say that procedural modeling here is really giving them a sandbox to digitally iterate.

I also wanted to highlight that we notice two overarching observations that emerge from these discussions. The first being when we asked interviewees to think about how procedural modeling could fit with other technology, things like generative design and Fusion 360, the overwhelming response was that they can imagine generative design being used to help narrow down high performing candidates in a design space first and then following up with procedural modeling tools to explore a subset of that solution space. In essence, this is a one-way pipeline going from generative design to procedural modeling.

But many of the interviewees could not imagine the reverse scenario, going from procedure modeling to GD. The second observation-- when we asked interviewees to think about ranking the sustainability scenarios based on difficulty versus impact, we found that scenario one, biomimicry, was less difficult to imagine creating but also least impactful. Scenario number two, design for afterlife, has a high impact and medium difficulty. Scenario number three, design for repair, was very difficult with some benefit.

And number four, manufacturing modifications, was very difficult with some marginal benefit. And number five, design for awareness, was considered a low difficulty but has a high impact. And to cap that, the real surprise here was that afterlife and cultural awareness had a high, immediate attraction that we had otherwise not originally anticipated or hypothesized. And when we think about this matrix and these tiered approach to the design strategies, it is very clear that procedural modeling has a place within sustainable design.

And to recap all of that in our key findings, interviewees were most receptive to cultural awareness. And they say that this one is an easy low lip way to alter consumer perception. Awareness could create transparency and the design and make process and with the objects we engage with every day.

They also acknowledged designing for afterlife as perhaps the most impressive and biggest opportunity in front of us. And lastly, they interpreted manufacturing modifications all very differently. This really pushed beyond our own initial intention of the scenario, with some individuals framing this as purely a manufacturing simulation problem or opportunity and others extending this concept to include procedural manufacturing-- kind of a new word there. But to define that, we're talking about the collection of historical machine data and generating geometry well-suited for the machine based on those insights.

So this brings us full circle back to the original hypothesis of can procedural modeling be used to design more sustainably? The short answer here is yes. The long answer, though, is a bit more complicated. And we need your help.

This conversation feels much bigger than us because it is quite nebulous. And we're really only scratching the surface here. So we want to engage with people like you to help bring more of this thinking to life and help further frame possible solutions.

Feel free to reach out to us on LinkedIn or apply directly to the Technology Centers if you would like to explore convergence with us. Contact information can eventually be found in the handout and posted in the chat. We look forward to all the questions you guys may have. Thank you.