There are already networks to generate depth maps out of 2D video, so hooking up the output to that and a simpler VR video encoder should probably work.
Will it be jank as hell? Oh yeah. Nothing is conveniently packaged in Local AI Land, unfortunately.
My friend wants to know if it can generate VR videos too?
Take anything created, ffmpeg it into frames, then run it through this script’s stereoscopic mode:
https://github.com/thygate/stable-diffusion-webui-depthmap-script
Theoretically? Maybe.
There are already networks to generate depth maps out of 2D video, so hooking up the output to that and a simpler VR video encoder should probably work.
Will it be jank as hell? Oh yeah. Nothing is conveniently packaged in Local AI Land, unfortunately.
Tell your friend that my friend is also asking too.
Asking the real questions