Comment by lg5689

Are you thinking of the int multiplication result instead? The DFT result uses a computation model with unlimited precision reals.