Wednesday, February 28th, 2024 Posted by Jim Thacker

Sneak peek: Chaos’s new AI technologies

Originally posted on 20 February 2024, and updated with more details of Chaos Next.

Chaos has posted sneak peeks of new AI-based technologies in development at the company.

The webpage for Chaos Next, the firm’s new “AI lab”, provides brief summaries of the V-Ray and Enscape developer’s “AI principles” and nine AI-based technologies that it is working on.

They span the full length of visualization and visual effects pipelines, from generative AI for creating PBR materials from text prompts, through tools for automatically populating or extending 3D sets, to technologies for applying new visual styles to renders.

What is Chaos Next?
On its blog, Chaos describes Chaos Next as an “AI lab”, exploring ways of new integrating AI technology into Chaos products “in a way that is useful and respectful”.

In this month’s Chaos Unboxed livestream, the firm described AI as complementing traditional artist workflows, not replacing them entirely, with Chief Product Officer Kam Star describing its aim with AI as being “to give you a better more powerful brush, not to replace your creativity”.

Which AI technologies are in development at Chaos?
The Chaos Next webpage shows nine “sample technologies” in development at the company.

Some are technologies for creating individual assets: according to Chaos, one of the first is “chat-driven material creation”, due in Chaos Cosmos, its online 3D asset library.

The feature, shown above, will let users generate editable PBR materials from text prompts.

Others are geared towards entire scenes, including a ‘smart scene populator’ for placing 3D objects, and systems for optimizing scene lighting and predicting future asset use.

A couple are for manipulating final-quality images, including a system to convert product images to “instant lifestyle visuals”, due in product visualization platform Cylindo Studio, and a new style transfer system for applying new visual styles to existing renders, due in Enscape.

How are the AI models used by Chaos being trained?
The Chaos Next webpage doesn’t provide any details of the data sets on which its AI technologies are being trained.

However, a general statement of principles says that Chaos “will be transparent in our AI processes, algorithms, and collection of user data, with utmost respect for user privacy”.

Which of Chaos’s applications will AI technologies be integrated into?
The Chaos Next webpage doesn’t say which of Chaos’s products the technologies will be used in, but this blog post confirms that the list includes Chaos Cosmos, Cylindo Studio and Enscape.

Most of the example images show architectural scenes of the type generated by Chaos’s Corona renderer, or Project Eclipse, its upcoming arch viz storytelling product.

However, some technologies – like Set Extender, which “intelligently extends” part of a 3D set – are targeted at VFX work, which suggests use cases in products like V-Ray or Chaos Vantage.

When will the Chaos’s AI tools become publicly available?
Chaos hasn’t announced a release date for any of the AI technologies in development.

Read more about AI technologies in development at Chaos on the Chaos Next webpage

Have your say on this story by following CG Channel on Facebook, Instagram and X (formerly Twitter). As well as being able to comment on stories, followers of our social media accounts can see videos we don’t post on the site itself, including making-ofs for the latest VFX movies, animations, games cinematics and motion graphics projects.