I am incredibly excited to reveal our @UniOfYork spin-out @pxld_ai (https://pxld.ai/), now out of stealth! Game-changing relighting, inverse rendering and harmonisation technology that lets you put yourself in any world with photoreal lighting and effects on the scene. You really can now get LED volume virtual production quality with nothing more than a phone and our tools.

I have worked on 3D computer vision, inverse rendering and relighting since my PhD back in 2003-2007. Before that, I was obsessed with the idea of VR and built a primitive VR system for my GCSE DT project! My interest was less about the VR hardware and more about the freedom that would come from being able to enter any real or imagined worlds. I've remained fascinated by this challenge and, when I think back to my PhD and our primitive attempts to edit the lighting on a face using shape-from-shading, I cannot believe how much progress we have made!

Have a look:
🔹 Rock solid surface normals and albedo maps, hair-perfect mattes and photoreal relighting and harmonisation into a scene https://pxld.ai/#passes
🔹 Click through the scenes for each video and see if you can tell which is the real one https://pxld.ai/#campaign
🔹 Check out how the video you record of yourself has a physical effect on the virtual world, almost like an interactive world model https://pxld.ai/#physical

Unlike a generative video model, our output isn't just conditioned on your recorded video (where your prompt and random seed luck will determine how well you and your performance are preserved). Your performance is kept exactly as recorded, pixel for pixel.

There are so many cool things you can do with these tools at your disposal. We'll be releasing demos over the coming days and weeks to illustrate some of the potential but I'm most excited about getting the tools into the hands of filmmakers, content creators and consumers to see what ideas we unlock!

Join the waitlist if you want access or send me a video of yourself and I'll put you in a new world.

Huge thanks to my cofounders @WillRowan3 and @JADGardner for bringing this vision to life.