> Mind you, for movies where the majority of many shots are CGI-composited (or for works that are just plain-old digital animation in their entirety), the 3D is "free": you have a 3D master whether you want one or not, and any 2D release is a post-processed edit.
It's not free because although the assets are in 3D, the final render still needs to be done, and for stereo you need to render every frame twice, once for each eye. When I was working on animated features the extra production cost for stereo was ~10% of the total production budget.
Mind you, for movies where the majority of many shots are CGI-composited (or for works that are just plain-old digital animation in their entirety), the 3D is "free": you have a 3D master whether you want one or not, and any 2D release is a post-processed edit.
> In a 3D movie, this results in a really disorienting effect where something in the foreground can be out of focus, and despite trying to focus your eyes on it, it will continue to be out of focus, while the part of the image that is further away is still in focus.
I've always wondered whether this problem could be "solved" with eye-tracking in VR. The gear would project a ray from your pupil to the image, hit a pixel, map it back through the projection matrix to the surface of an object in the scene, and then dynamically adjust the depth-of-field so that that part of that object (and everything else at the same depth) was in focus.