The edge of architectural visualization at Gensler

We spoke to Scott DeWoody, creative media manager at Gensler about all things architectural visualization. Over the past 10 years, Scott has worked with numerous clients, including NVIDIA Corporation, ExxonMobil, Shell Oil Company, BP, City Center Las Vegas, and many more…

Tell us a little about your role

I oversee Gensler’s visualization technologies, managing rendering workflows, training, documentation and R&D into new technologies such as VR/AR/MR. I consult on a lot of projects too, figuring out what might be the best solution to a specific problem. For example, a couple of years ago I worked out of our San Francisco office on the NVIDIA headquarters. I was tasked with bridging the gap between a design team’s workflow and NVIDIA’s in-development hardware. I worked on IRay and Visual Computing Appliance (VCA) solutions for network attached rendering; which has since morphed into the DGX station, a deep learning hardware platform.

It’s great to work for Gensler, employees are presented with big opportunities, especially with the sort of clients we have and the variety of projects. We work on everything from the 600m tall Shanghai Tower to wine labels. I’ve been here about 10 years now and have always been able to try new things and work on exciting projects with incredibly talented people. Oddly enough, I don’t really have experience elsewhere. I got hired right out of school, a week before I graduated. When I walked in on orientation day, Gensler had something like 2,500 employees and 37 offices, now we’re at 5,500 with across 44 offices. We’ve grown a lot since 2007 and I feel like I’ve been a part of it.

What design tools are employed by the team?

Our core design platforms are SketchUp, Rhino, Revit and to a degree 3ds Max. V-Ray is our core visualization engine, we also use a little IRay and a handful of other apps. We do in-depth work in the Unityand Unreal gaming engines and produce some interactive visualizations. We manage our network resources and essentially create a global render farm by using Deadline.

How does the team set out to produce powerful and engaging imagery?

This definitely varies per project and per designer, since we have so many. One constant, however, is that people want really good quality outputs. With V-Ray 3 across all of our platforms, I feel like that’s a baseline we can easily achieve. I’ve been using it since version 1.4 so I’ve seen it evolve over the past twelve years or so. It astounds me what Chaos Group keep producing and our designers just enjoy using it. The new interactive mode allows rendering whilst simultaneously working on a project, not having to stop-start over and over again is a big win.

The latest features have given our designers whole new view on rendering quality, what they can expect. It’s kind of up to the artist how far they want to take things, from very conceptual, to diagrammatic, to as photo-realistic as they can make it. The system is flexible enough to do anything.

Materials are the latest endeavor we’re just getting dialed into. We use a system by Avail, it’s designed for the AEC industry to provide shared libraries. We’ve got 1500+ V-Ray materials loaded in there, some of it is stock and some custom. We use Adobe Substance Designer, a procedural material builder to create virtually anything; carpet, wood, marble, stone, you name it we can create it and make it look crazy real. Gone are the days of tiny tiles that scale terribly.

The full library is synced out across all the offices, so no matter where you are, you have access to drag and drop into your design. Right now I'm creating sort of a default set of go-to materials, the team can just grab them and know they are going to work. Everyone will be using the same palette to get their first renderings off the ground.

Is there a notion of a ‘standard workflow’ at Gensler?

The majority of our designers follow a similar workflow. The Microsoft Store project is a great example of this; we were chosen to work on the new retail outlets to help bring Microsoft’s hardware to the high street in an engaging way. The design team started in SketchUp, as they always do. They can quickly dive in and mass out their ideas-the abundance of downloadable textures and entourage also speeds up the design process. Our retail teams use SketchUp a lot, they’re probably some of our biggest power users. With Microsoft, they were able to iterate different options extremely quickly and even present design changes in the app on-the-fly.

A lot of visualization is done in SketchUp and this is where V-Ray 3 comes into play. We can now design and render all in the same package. And, with the introduction of V-Ray 3 for Revit, the team can share materials and move freely between platforms. For some projects, we go into 3ds Max for the production of high-end renderings.

Later in the design stage, we’ll look at the real-time rendering and utilize a mix of Fuzor and Enscape. There are many on the team that loves these extensions. Some prefer 3ds Max interactive, which is based on the Stingray engine. For larger projects, it’s often the case that we’ll go down the gaming engine route too. For Microsoft, the client was in Seattle and design work took place in LA. To help work around this we created an interactive Unity app. It was a pretty powerful tool. My colleague Alan Robles was able to show the client how the huge wraparound LED video walls would translate into the design.

When we go into a game engine, effort levels increase exponentially. Right now we are figuring out how best to optimize that pipeline. There are people out there that have set out to solve the issue and we are trying to speak to all of them. If you figure out how to get a model into a gaming engine really quickly, you've won. A lot of designers jump into one button real-time render apps. The output looks extremely consistent. You’d think that’s a good thing but with limited customization, all firms start to look the same. We want to avoid that.

When the client team is happy with how things are looking, we enter the design development stage, that's typically then when we start to move into other software such as Revit.

What does collaboration look like for you?

The style and level of rendering are vast and wide because we don't look at it as an addition to what we’re doing, it’s just a part of the process now. It’s been a goal of mine to have a consistent platform of rendering across the board. So that no matter what app you're designing in, you should be able to get a good rendering out. V-Ray upgrading to version 3 across all platforms helped me solve that problem really well. It’s helpful to not only get consistent expectations, but with the same version, you can use the same materials and additional features. This lets you set up a pipeline across platforms. Not only is the look and feel essentially the same across the project, the render quality is consistent too. This couldn't easily be achieved a year ago.

What does the future hold for design visualization?

I think we’re at this interesting point in visualization where real-time rendering is becoming extremely accessible, and a lot of people are jumping into it. But at the same time, current real-time rendering isn’t ray tracing [simulating the effects of light as it encounters virtual objects], they’re vastly different. For some people real-time is ok, but for others, ray tracing is where things need to be in the end. I’ve got a feeling that those two worlds are going to collide and there will be a point where we won’t call it ‘rendering’. It’s just going to be how things look out of the box. I’d compare it to photography, it used to be that you needed an expensive camera to take good pictures, but now everyone has a decent camera in their pocket. The same thing will happen in AEC, whether that’s access to more powerful CPUs and GPUs or the cloud comes into play and you're no longer rendering on your hardware. Like taking a photograph, click a button, and boom, I’ve got my image.

Tell us about your experience with Mixed Reality at Gensler

We’ve had some really good experience on projects where we use HoloLens in tabletop mode instead of building a physical model. The client can get a real feel for the design early on and instantly cycle through options.

There are definitely use cases for HoloLens throughout the entire design process, but we think it really shines when you’re on-site and have the model 1:1. It is awesome for both us and clients to walk around the actual space.

For the creation of the new Gensler Los Angeles vertical campus, we used the HoloLens when designing a bridge to connect the existing building with the adjacent tower. Back when Trimble was developing SketchUp Viewer for HoloLens, Alan Robles got in there early to do a proof of concept, dropping the custom designed bridge into the real-world environment worked like a charm.

Advanced distributed rendering with V-Ray Swarm, image courtesy of Chaos Group.

Rendering a 360 degree panorama for VR, image courtesy of Chaos Group.