Creating Materials for the Hyundai Genesis G380
During NVIDIA GTC 2016, Hyundai and Allegorithmic shared their experience of creating a full set of materials for the Hyundai Genesis 380 using Substance Designer.
Last week at GTC 2016, Allegorithmic unveiled a new project with Hyundai for virtual materials creation for the Hyundai Genesis G380. Here is an overview of the conference presentation by David Nikel, Digital Model Manager at Hyundai California Design Studio, along with Allegorithmic’s Jérôme Derel, Chief Product Officer, and Pierre Maheut, Industrial Design Content Lead.
What are the Hyundai design team’s challenges?
According to David Nikel, Digital Model Manager, Hyundai California Design Studio has three main challenges when it comes to materials:
- New materials are difficult to create and change.
- Materials need to be of the best quality and as close as possible to reality.
- Material information is not shared between teams.
How does building virtual materials address these challenges?
The use of Substance and procedural texturing allows creative designers, as well as color and trim designers, to explore and iterate on material creation. This includes every aspect, from color to material to finish.
For example, the central console of the Hyundai Genesis G380 is made of a unique 3D patterned plastic. Within Substance Designer, it was possible to define the the finish of the 3D pattern with a custom proportion of glossy and matte. It is also possible for designers to interact in the same project at differing levels of complexity: for instance in Substance Designer, it is possible to open parameters directly
At the same time, designers can iterate the 3D pattern itself. Changing the input of the graph creates additional variations, which can then be rendered in high definition in the Substance Designer viewport – within the context of the full interior – thanks to NVIDIA Iray technology. With this iterative workflow, designers can experiment rapidly, make choices, and visualize the final production car early in the design process.
Using the MDL standard to create virtual materials reusable company-wide
In order to attain the highest level of quality, we decided to use NVIDIA’s Material Definition Language (MDL) to describe the materials. For instance, the wood insert of the dashboard started as a wood texture, which can come from a Substance (procedural texture) or a scanned material. We then added a clear coat on the surface using MDL graph edition. Substance Designer is the first and only toolset that allows creation of MDL materials using a node-based approach, which is a quick and efficient way to create a library of MDL materials.
As Material Definition Language emerges as the new standard for material definition, the created MDL materials can enrich the company library and be shared between teams: Design, Engineering and Marketing. MDL is already compatible across several rendering software such as NVIDIA Mental Ray and iRay, Otoy OctaneRender 3, and others to come.
Why is Substance more than just a texture?
In talking with David about his former projects, we discussed where and how Substance can help design teams to save time and iterate quicker. The challenge is to generate the front grid of a car, based on a hand sketch created by one of the exterior designers, as quickly as possible. This time, we have recreated the pattern procedurally within Substance Designer,we could also have used an SVG file from Adobe Illustrator Using the Height, Normal, and Opacity channels, we were able to simulate a fully modeled 3D grid using a simple surface and a Substance.
According to David, “Iteration in the early concept phase is something our designers want or need more of.” By allowing the design manager to compare two grids side by side within the background of the Hyundai California courtyard environment, it is possible to make decisions before any parts are produced or modeled in 3D, saving time while providing physically accurate high-end images.
All images courtesy of Hyundai
See video of the full GTC session here: