upvote
The simplest way to show this is that the adjoint (complex transpose) DFT matrix is the inverse, so U^H U = I.

(For DFT there are different scaling conventions which may be taken into account.)

reply
Maybe not the simplest way but constructive proofs are always nice.
reply
Yes! It's discussed in the "Inverse DFT: reconstructing the signal" section, which relies on the earlier post https://eli.thegreenplace.net/2026/notes-on-the-complex-root...
reply