TLDR: I've solved camera motion for a video clip. How can I export each camera's position and rotation to "real", unconstrained camera objects?
I'm using Blender's motiong tracking tools to reconstruct 3D spaces from still images (from an old game I love). I added in two image "frames", set up a bunch of tracking points, solved it (with 0.20px error, good enough for my purposes) and clicked "Setup tracking scene" to export the whole thing to my 3D workspace. Now I can press the arrow keys to jump between frames so I can see my 3D space from each solved camera position.
But I want to export each frame's camera position to a "real" camera object so that it stays in my scene even if I delete the empties etc. How can I do this? When I look at the properties for the 3D camera object, the position and rotation values don't change as I switch between camera positions, even though the camera clearly did move several meters.
I tried selecting the camera then clicking Object -> Constraints -> Clear object constraints, but this resets to the first frame in the video clip, so I don't see how to get camera coords for subsequent frames.