Since Adobe’s Photoshop beta, more artists have been introduced to frame expansion. Interestingly, this workflow also benefits video shoots.
Matte paintings aren’t anything new, as Epic Light Media points out. However, the skills needed to create a matte painting have been greatly diminished. With AI-generated images, we can expand our scene with more ease than ever.
Previously, smaller productions would shoot their talent on a green screen. This meant that the fake studio set would be entirely created in post. You wouldn’t dare to expand an existing shot. It would take too long and risk looking fake. Photoshop makes this work easier than ever.
What I love about Thomas Manning’s setup here is that he’s shooting vertically. He’s getting the absolute maximum resolution that he can. The entire 16:9 shot is somewhere closer to 17K, which is an incredibly useful resolution for punching in.
The Future
Manning isn’t moving the camera here, but what if he could? Generative AI tools are being used to create 3D objects, in some cases just from original 2D source material. If we moved away from Photoshop’s 2D generated images, to a 3D model, we’d get two benefits.
The first benefit of creating a 3D scene is that you can move the camera. Perhaps not a whole lot (like turning around), but enough to sell a handheld shot. The second benefit is the ability to dial in the lighting. If you know where your lights are placed in the real world, there’s every possibility that the same lighting placement could be added to your 3D scene. As fake studio sets become more and more ubiquitous, I'm sure we'll see real 3D set extensions arriving for consumers within the next five years.
