A team that has made a name for itself in the field of autonomous vehicle technology is now stepping into Hollywood to bring a revolutionary cinematic-level visual AI effect called Odyssey. In simple terms, this team aims to use AI to revolutionize the production methods of movies, TV shows, and video games.
See the effects for yourself directly in the video:
Odyssey draws inspiration from computer graphics research in the 80s and 90s and the story of Pixar's founding. Its goal is to create movies, TV shows, and video games using artificial intelligence. In the process of achieving this goal, the biggest obstacle encountered is—AI controllability. In response, Odyssey proposes a new solution: a more powerful generation model.
They are training four generation models to achieve complete control over the core level of visual storytelling. These four models generate high-quality geometric shapes, realistic textures, realistic lighting (controlling time, seasonal changes), and controllable motion. A single model can achieve precise configuration of scene details, and when combined, they can generate complete videos or scenes.
Odyssey allows users to iterate continuously, and its generation workflow will be combined with professional-level version control. The founders mention: "Odyssey has been designed to easily integrate into existing workflows or systems being used in Hollywood and other regions." All content can be edited and exported in standardized 3D file formats, such as USD format.
Of course, while Odyssey looks quite impressive, sharp-eyed netizens have still found some minor bugs. For example, the hand details at 00:30 are slightly rough:
The Odyssey team has a close relationship with autonomous vehicles. Most of the founders have spent years building AV (autonomous vehicle) technology. The two core founders, CEO Oliver Cameron and CTO Jeff Hawke, are experts in the field of autonomous driving.
CEO Oliver Cameron: Angel investor for more than 60 DeepTech startups, has been engaged in the field of autonomous vehicles during his early career.
CTO Jeff Hawke: Responsible for Odyssey's directional visual AI, his research areas include machine learning, computer vision, robotics, etc. He developed an expert hybrid vision system for autonomous vehicles and mobile robots during his doctoral studies at the Oxford Robotics Institute.
As Jeff Hawke put it, "Autonomous vehicles are taught how to navigate the world, Odyssey is taught how to generate the world."
Odyssey not only turns to 3D and generates multi-layer 3D worlds but also achieves fine-grained control through multiple generation models. In addition, to create a Hollywood-level visual AI, Odyssey has recruited a team consisting of AI researchers and Hollywood artists. Researchers come from Cruise, Wayve, Waymo, Tesla, Meta, and artists are the creators behind works such as "Dune," "Godzilla," and "The Avengers."
So, what do you think of the effects of Odyssey? Can it really become the new era's Pixar? Everyone is welcome to leave comments and discuss!
Official website: https://top.aibase.com/tool/odyssey
