The key is to have the same paths on both machines. This can be done with path substitution.
Pick one machine to do all the scanning. Then, add a path substitution on the other machine to match the path that is used to do the scan. The path substitution only has to contain the part of the path that isn’t duplicated.
Clear out the database again, and scan from the chosen machine, and the other machine should be able to access the videos.