🧵 Could Pseudo-3D rendering work

Anonymous at Sat, 11 May 2024 18:00:22 UTC No. 16171056

Just a stoner thought I have had for a while. Would it be possible to create a photorealistic” 3D” animated film without actually rendering. All footage of things, live action, or 3D, is displayed digitally as being composed of pixels, which are squares of a single color. If you knew where to put each pixel, hypothetically you could and also made sure they properly conveyed depth you could create a photorealistic 3D image without having to render anything in 3D software. Further more you could stitch together a series of these flat 3D images and make the next frames if you kept track of where pixels would go (an informational map of the parts of an image if it was brought into 3 dimensions, perhaps generated by AI) when they disappear and know which ones should appear when another part of an object is shown, allowing an animation to be created. This is similar to how the original Doom game was rendered, a strip of 2D images that displayed a part of the strip when a coordinate on the level was put in through movement. While generating these images would be hard, it would allow entire films “live action” films to be made with nothing being filmed.