At the point that you are testing it at, everything under 400Ohms will sound like crap whether it is variable or not. There is a pretty good reason for the 1/8 rule of thumb. So, all graphs measured with a source that is within this 1/8 limit will look the same(ish) (it might look a bit different, but only because of the differing power structures of the source amplifier).
Having said that, you are right that the NP are indeed variable in impedance (rising as the FR rises), because of the very nature of their design (in fact all BAs are variable), made using inductive parts. Which, as you cited is the exact opposite of an ortho dynamic driver, having no inductive parts. But, at this level of output impedance, anything with lower impedance will not sound good. Having such a low impedance, the NP really needs to have an output impedance of less than one for proper sound reproduction. Any higher than that, and you start to skew the FR curve. This actually would also hold true for any sort of transducer.
Also, because you used the LO into your IEMs, it will exacerbate the poor sound reproduction because as @firesign pointed out, it isn't meant for headphones/IEMs. Assuming it is a true LO, this means that there is NO voltage swing, and we all know that you need that as you turn up the volume to properly dampen or excite the drivers. If you want to properly test using a high output impedance, I would order some (varying) filters that will add more output impedance to the mix, and use the PO where the amplifier can properly do its job.
Thanks but I don’t want to test anything.
I used the LO of the Gryphon because it was an easy high impedance source that I had to hand. I know with a planar IEM off the same LO source it makes no practical difference to the sound so it was a convenient quick and dirty check in the impact of a very high output impedance to very crudely establish that the impedance versus FR of the NP was indeed nowhere near flat.
I was therefore interested what that plot looked like just for academic interest.
I have several sources that have an output impedance of between 0.2 and 1.0 ohms so I am good in terms of what is technically appropriate.
It seems that perhaps that plot isn’t readily available, all good.