Just to remind people of the level of variation we are talking about, here are the responses from the ears of 15 different people to five different pairs of closed-backed headphones. IIRC one of these was the Stealth. From Sean Olive on Twitter.
https://twitter.com/seanolive/status/1630389960904957953
You don't even need to have different people. Moving the Expanse in different positions on a jig already gives you differences. Here you have the Expanse on a miniDSP test jig in 5 positions (with HEX compensation);
- Magenta - MID; Headphone centred, so that the ear is (approx) located in the middle of the chamber
- Red - FRONT; Headphone moved backward, so that the ear is located at the front of the chamber
- Cyan - BACK; Headphone moved forward, so that the ear is located at the back of the chamber
- Yellow - TOP; Headphone moved downward, so that the ear is located at the top of the chamber
- Dark blue - BOTTOM; Headphone moved upward, so that the ear is located at the bottom of the chamber
Averaged:
So what did we test with the different people? Were the headphones set up consistent? How would you define consistent (position to what, pressure, angle)? What makes sense, is that the seal is consistent for variations; this is also what Sean Olive is stating:
https://twitter.com/seanolive/status/1630392145965350912
You can also hear this very easily when putting up pink noise, and move a headphone around. The timbre changes considerably, and some headphones are more sensitive to it than others.
Not all measured deviations regarding HTRF per person are relevant to the Harman Curve. People have different shapes, and hence a neutral reference speaker measures differently when the measurement includes the person's shape. A neutral reference speaker would still be assessed being "neutral" by those people (assuming they can assess), despite the measured differences. The Harman Target, an average weighted curve, only makes sense for the jig for which it was defined. It doesn't matter that a headphone measures different on other jigs or other people's ears. And as being mentioned a couple of times before, the Harman target is not a golden reference, just a good approximate starting point.
As good as I could, I've recently measured the Expanse on my own head (without any compensation applied!), and compared it to the (MID) miniDSP jig measurement
- Magenta - MID JIG; Headphone centred on the jig, so that the ear is (approx) located in the middle of the chamber
- Red - MID EARS; Headphone centred on my own head, so that my ear is (approx) located in the middle of the chamber
I understand the drop in the low frequency region, as I cannot completely keep the seal when measuring with a UMIK-1 microphone on my head. But from say 60Hz to 7kHz I would call both measurements pretty consistent. What stays consistent with all measurements is the peak around 3-4kHz, that I attenuate with about 3dB, and to my opinion improves the perceived sound considerably.