Might be worth adding the optimal source impedance for noise (en / in) too. The AD797 is conspicuously absent. I presume the in values are for symmetrical input impedances only in some cases?
The NE5532A noise current values appear to come from the typical graphs that TI added a couple of years ago. The values in those graphs contradict what is in the datasheet tables and the graphs look rather suspicious to me: the 1/f corner frequency for the noise current is lower than for the noise voltage, which is the opposite of what you normally see with a bipolar input stage.
