AI, FILM & VIDEO

AI-generated RAW clips – pipe dream or soon reality? 🚀🎬

For professional editors and colorists the advantage of RAW footage is obvious: maximum control over colors, dynamic range and image detail.
But anyone working with generative AI tools like Runway, Pika Labs or Sora today quickly notices:
The clips that come out are usually already graded, compressed and in a fixed color space. 🎨
🤔 Why are there no AI-generated RAW clips yet?

Current generative models work with RGB images and not with log or RAW data.
There are several reasons for that:
📊 Data volume:
RAW formats contain huge amounts of uncompressed image information, which makes AI training considerably more complex.
🎛 Color management: Generative AI models currently have no native LUT or log processing.
🔗 Workflow limitations: AI tools have so far been optimized for end users, not for high-end post-production.

🎥 Workarounds for a “RAW feeling” with AI clips Until real AI RAW formats are available, there are a few tricks to make generated clips more usable professionally:
🛠 Generate with AI (Sora, Runway, Pika) at maximum quality
🎨 Convert to log color spaces via DaVinci Resolve (Color Space Transform, Dehancer, FilmConvert)
🔍 AI upscaling & detail enhancement with Topaz Video AI or DaVinci Super Scale

🔮 Outlook: will real AI RAW formats arrive soon?
It would be a game changer for the film industry!
An AI that exports directly to ProRes RAW or Blackmagic RAW would revolutionize the workflow. So far there is no tool that can do that – but who knows?
Maybe that is the next big step for AI in media production. 🚀