Back to the testing method....
I see a pattern that kind of alerts me. The HP is an electro mechanical reproducer. The HP usually use one single full range device. it plays (on test as on our head) in a close environment. No room response, no axis angles (on axis, off axis etc') and it also has some isolating function (not much reported in Amir's tests!).
The response of such device should be continuous. I do not see why it should show such fluctuations and variations, with slight Fr. change.
Electronics, as Amp's and other pure SS devices, don't do that. Usually, once they variate, the attitude is kept. Roll off at high Fr.
or the same at very low Fr.
The test is kind of a sealed (black) box, when the HP goes on that head like dummy and than it is tested. I still would be happy to know what
HP Amp. is used, What Mic. amplifier is used and what kind of test pattern of AP is applicable.
When speakers are tested, there are some rules kept, as on axis and distance, to measure efficiency. Fr. is more complicated and need some controlled environment (noise, reflections etc'. Some of a speaker Fr. may come from crossover and multi elements structure. All that is not valid
with HP. No crossover, no crossover distortion, no elements overlap (-3dB) and many more "no"'s.
View attachment 134993
Look at 8.5kHz or even worse at about (?) 17kHz. The 17kHz is deviate by 20dB in about 1kHz (it's more like 5% Fr. deviation. Down 20dB and back up 20dB. 20dB equals to x100 amplitude.
Now let's compare the same at 500 Hz and down. Here, the response is with very smooth variations. The deviation is also no more than 7.5dB. That's about x5 only!
This chart is kind of making no sense to me. Especially that an electro mechanic device response should look more like the 500Hz rather than the 17kHz. Electro mechanical devices, are less "jumpy as Fr. rise. As the mechanic part has a mass, it is still a coil moving in an electromagnetic field
and this mechanism is practically tend to slow movement not to increase it. On this chart, the higher the Rf. goes, larger is the deviation!
This is against physics of standard mechanics or electromagnetism. Sr. Newton or Mr. Maxwell would be very unpleasant with those results.
Going back to the test, this dummy that the HP is placed on, what is it standing on? The same table with the AP test equipment / PC?
Does it vibrate? do you have a sensor (vibration) sticked to the table, that shows the output vs. Fr. from 20Hz to 20kHz?
My guts feeling are that the sudden Fr. change with amplitudes of up to x100, within very little change in Fr. is not the HP property.
If so, the entire HP testing should be revised. Something I remember from Physics lab, is that if the test results do not fit your expectations, you need to check your test methods before publishing the results.