The only spec that matters is sensitivity. For normal speakers, sensitivity is usually listed as dB Per Watt at 1 meter. For headphones, it is usually listed as db per milliwatt at "ear distance".
There are lots of things that can affect sensitivity, and impedance is just one of those.
What I can say is that at the same "volume setting", one of these headphones will consume about 7.8 times more wattage than the other. But how that impacts volume is completely up in the air because we don't know the sensitivity.
Since I am exciting with voltage and reading the current I assume I have a ratio of current to voltage which is admittance.
More specifically, you can call this a trans-admittance, since the voltage is applied at one port of the filter and the response current is measured at a different port.
If you just talk about an "admittance" people will at least at first assume you are talking about the voltage/current relationship at a single port.
Do I convert to current correctly? Do I find conductance and susceptance correctly? Did I understand correctly that the values I measure is admittance?
I didn't follow through your calculations entirely, but you did not come up with the correct result for this case. Your result may be correct for the normal admittance to impedance conversion for a one-port network.
For converting transadmittance to transimpedance you have to consider both ports of the network.
The admittance representation is like this:
\$\begin{bmatrix}I_1 \\ I_2\end{bmatrix}=\begin{bmatrix}Y_{11} & Y_{12} \\ Y_{21} & Y_{22}\end{bmatrix}\begin{bmatrix}V_1 \\ V_2\end{bmatrix}\$
To get the Z-parameters you need to invert the Y matrix. Rather than calculate the inverse by hand I looked it up on Wikipedia, where I found the result for the Z21 term is
\$Z_{21} = \frac{-Y_{21}}{Y_{11}Y_{22}-Y_{12}Y_{21}}\$
Separating this out into real and imaginary parts is of course another algebraic effort, which I'd much rather do using software (Mathematica or whatever) than sort through by hand.
Best Answer
Impedance does vary by frequency. However, for many things it doesn't vary very much by frequency within the expected operating frequency of the equipment, and a single number is much easier to handle than a frequency response graph. So people quote a single average number.
In fact, your linked article admits this:
That's quite a lot of variation! It also explains why this causes variation in the audio quality when driven from different sources.