SIGGRAPH: Nvidia Studio 3D showcase brings updates on the Omniverse

SIGGRAPH: Nvidia Studio 3D showcase brings updates on the Omniverse

What do you do when the rest of the world is obsessed with the metaverse? Why, you go up a level and create an Omniverse, of course!

Nvidia took the opportunity at this week's SIGGRAPH 2021 to share the latest information on its Omniverse. The immersive and connected shared virtual world is now in open Beta, with artists creating one-of-a-kind digital scenes, 3D models and more. The Omniverse platform continues to expand, gaining Blender USD support, a new Adobe Substance 3D plugin, and a new extension, GANverse3D - designed to make 3D modeling easier with AI.

Free to Nvidia RTX and GeForce RTX GPU users, the latest creative app updates are now available for download with the August Nvidia Studio Driver.

Blender, Substance and GANverse

Nvidia announced that Blender, the world’s leading open-source 3D animation application, will include support for Pixar’s Universal Scene Description (USD) in the 3.0 release, enabling artists to use the application with Omniverse production pipelines. The open-source 3D file framework gives software partners and artists multiple ways to extend and connect to Omniverse through USD adoption, building a plugin, or an Omniverse Connector, extension or app. An experimental Blender alpha 3.0 USD branch that includes more advanced USD and material support will be available soon for users everywhere.

In addition, NVIDIA and Adobe are collaborating on a new Substance 3D plugin that will enable Substance Material support in Omniverse. Materials created or imported in Substance can be adjusted directly in Omniverse. 3D artists will save valuable time when making changes as they don’t need to export and reupload assets.

Lastly, Nvidia is to release a new Omniverse extension, GANverse3D – Image2Car, which makes 3D modeling easier with AI. It’s the first of a collection of extensions that will comprise the Omniverse AI Toy Box.

GANverse3D was built on a generative adversarial network trained on 2D photos, synthesising multiple views of thousands of objects to predict 3D geometry, texture and part segmentation labels. This process could turn a single photo of a car into a 3D model that can drive around a virtual scene, complete with realistic headlights, indicators and wheels.

The AI Toy Box extension allows inexperienced 3D artists to easily create scenes, and experienced artists to bring new enhancements to their multi-app workflows.

You can download the Nvidia Omniverse open Beta here.

Managing Editor

Steve is an award-winning editor and copywriter with 20 years’ experience specialising in consumer technology and video games. He was part of a BAFTA nominated developer studio as project manager for the UK’s first fully interactive digital TV channel. In addition to editing, Steve contributes to, and, as well as creating marketing content for a range of SMEs and agencies.


No comments
View options
  • Order by latest to oldest
  • Order by oldest to latest
  • Show all replies