Nvidia partners with luxury EV maker Denza on Omniverse Cloud

Denza is a luxury electric vehicle (EV) brand born from a joint venture between BYD and Mercedes-Benz. And today it is partnering with marketing and communications giant WPP and Nvidia Omniverse Cloud to develop its next generation of car configurators.

The announcement was made by Nvidia CEO Jensen Huang during the Siggraph conference today in Los Angeles. The collaboration between Denza, WPP, and NVIDIA Omniverse Cloud — a cloud platform available in more than 100 countries — aims to revolutionize the car configurator experience by leveraging real-time, physically accurate digital twins and advanced generative AI tools.

WPP is utilizing Omniverse Cloud, a platform designed for developing, deploying, and managing industrial digitalization applications, to streamline the automaker’s intricate design and marketing pipeline.

By integrating comprehensive design data from Denza’s preferred computer-aided design tools via Universal Scene Description (OpenUSD), Omniverse Cloud enables the creation of a single, physically accurate, real-time digital twin of the Denza N7 model.

OpenUSD, a 3D framework that facilitates interoperability between software tools and data types for virtual world creation, plays a crucial role in building the unified asset pipeline. This implementation breaks down proprietary data silos, enhancing data accessibility and enabling collaborative and iterative reviews for Denza’s large design teams and stakeholders.

The use of Omniverse Cloud allows WPP to engage in launch campaigns at earlier stages of the design process, resulting in faster and more cost-effective iterations.

A better car configurator

Car configurators are the visual platforms that consumers use to pick the options they want on their dream cars.

The unified asset pipeline in Omniverse Cloud empowers WPP’s teams to connect their OpenUSD-enabled design and content creation tools, including Autodesk Maya and Adobe Substance 3D Painter, to develop a new configurator for the Denza N7.

By utilizing the unified asset pipeline, artists in WPP’s teams can iteratively and instantly edit a path-traced view of the complete engineering dataset of the Denza N7, ensuring that the virtual representation accurately reflects the physical car.

Unlike traditional car configurators that require hundreds of thousands of prerendered images to showcase various options and variants, OpenUSD enables WPP to create a digital twin of the car that encompasses all possible variations within a single asset. This eliminates the need for prerendered images.

Rev Lebaredian, vice president of Omniverse and Simulation Technology at Nvidia, said in an interview with VentureBeat that most car configurators today are 2D car visualizers, where everything has been pre-rendered.

“The way they do that is by essentially taking pre-rendered layers. And in the browser, they combined the layers, much like you would do in Photoshop,” Lebaredian said. “And the problem with that, in order to create a system like that, you have to render thousands or even millions of images. And if something changes in the design or you’re wrong about something, you have to go back and re-render all of them.”

So it isn’t easy to make design changes. So car companies only do that a few times a year, and that doesn’t really keep up with the needs of consumers.

“It is replaced now with something that’s basically instantaneous,” Lebaredian said. “To create a 3D configurator, it has to be real time and interactive. You don’t need to pre-render anything. It is rendered on the fly in real time. But also with Omniverse and USD, you can now build a pipeline that goes directly from all of your engineering data in the car company into a full data set that what we’re calling a super digital twin that has all the possible variants in the system already. That feeds into the configurator. It’s fluid, you don’t have to go through a batch process of re-rendering everything.”

WPP’s environmental artists are creating fully interactive, live 3D virtual sets. These sets can be based on scanned real-world environments, including those captured with WPP’s robot dog, or generated through generative AI tools like those offered by Shutterstock. These tools allow for the rapid creation of 360-degree high dynamic range imaging (HDRi) backgrounds, maximizing personalization opportunities.

“Everything was rendered in real time,” said Jensen Huang, CEO of Nvidia, in his Siggraph keynote. “All of a sudden, the car appears where you like it to be. This is one example of how generative AI and human designs come together.”

Shutterstock utilizes Nvidia Picasso, a platform for constructing generative AI visual models. At SIGGRAPH, Shutterstock announced the introduction of their first generative AI service, 360 HDRi, which enables the creation of photorealistic HDR environment maps for scene relighting. This feature empowers artists to swiftly generate custom environments tailored to their needs.

Once the 3D experience is complete, WPP can publish it to the Graphics Delivery Network (GDN) with just one click. GDN is part of the Nvidia Omniverse Cloud and consists of a network of data centers designed to deliver high-fidelity, real-time 3D content to almost any web device.

This allows for interactive experiences in dealer showrooms as well as on consumers’ mobile devices. The streamlined process eliminates the manual packaging, deployment, hosting, and management of the experience. If updates are necessary, WPP can easily publish them with a single click.

Source