SuperDepth3D
can read and reinterpret the depth buffer to generate 2 views, one for each eye, for stereoscopic 3D.
I think saving the RGB image alongside depth would be a lot better than saving the 2 generated views, in case there's a better depth buffer-based stereoization in the future, plus some
fancy effects in image editing software
.
I just read that
an option to disable the alpha channel was added
a while ago, but it would be nice to implement embedding the depth buffer, which I think newer iPhones do since they have a depth camera, tho perhaps using an image format that explicitly supports depth buffer like
JXL, AVIF or HEIC
, which might allow saving the full depth data/metadata without being constrained to an 8-bit channel.