@aaron I think what you have in mind is pretty much what the color engine is already doing. Structure emulation is image dependent ie localized to images specific pixel coordinates. Think of the film structure emulations as operating on a 2D image plane as opposed to color engine transforms operating on a 3D gamut. It's the same image data represented in different geometrical shapes. If color operations would depend on the 2D distribution of pixels on the image plane, they could not be represented in a 3D LUT and would have to be evaluated for every single frame of video. This is technically possible (de-noisers for video work like that for example) but it's not required for faithful film emulation. I've actually been playing with a prototype 4D LUT format that stores temporal information in the fourth dimension of the cube that would adapt the color transform behaviour across varying frames of an input video signal but found that it just doesn't make enough of a difference. I'm pretty confident in saying that the color of almost all film stocks can be stored in a single 3D lookup table. Temporal and localized adjustments for video make more sense for color "correction" tools imo. Think automatic rotoscoping, face tracking, complete re-lighting with AI etc. Check out WonderDynamics for some truly groundbreaking tech in that field!