1) As a coincidence, I have just been doing such measurements, because I can see, from many posts of many posters here, that the transformer input impedance concept is not well understood, in general. I do not criticize it, just noting it. I will try to stay as much technical only as possible. Transformers are a kind of "steam loco" engineering, it is basic electrical engineering and nowadays generation is not much in contact with such things.
So, this is the NA2M input impedance when loaded with 10kohm
View attachment 294126
Please note that 47kohm terminating resistor inside NA2M and 10kohm load makes a parallel resulting resistance seen at the upper end.
and this is when loaded with 1.6kohm, as in case of my Cosmos ADC
View attachment 294127
1.6kohm above 1kHz, below the L1 = 1.94H influence.
2) Why @amirm has measured the LF roll-off below 80Hz or so? Because the adapter was overloaded, saturated core and the LF shape then is not sinusoidal and rms value falls down.
At -3dBu (550mV approx.), I measure LF roll-off starting at 40Hz (still flat) and reaching -4dB at 20Hz. At lower level like 200mV there is no roll off and 20Hz is still flat.
Please everyone, try to thing scientifically, with your brains, less emotions and heart involved

. Listen to the facts and make your own studies on the subjects discussed.