Yet... the Stealth and Expanse are not U-shaped at all and are supposed to have better stereo imaging.
I had contact with SamV when he was developing that. He had the HD800S and Edition X that he found to have good spatial qualities compared to some other headphones he had at that moment to test.
AFAIK he averaged the PRTF (the difference between a pinna and no pinna on his fixture) of those headphones and used that as a target assuming a headphone that comes close to it must have good spatial properties and then good score that aspect.
Personally I have not seen any good relation between other headphones with good spatial qualities and the PRTF measurement so disregard it.
The idea was that a dip at 10kHz at his fixture would be the biggest reason for this. A similar dip is observed in Amirs measurements of all headphones and is phase nulling caused by the pinna.
If it were indeed that aspect then peaking or lowering the 10kHz at that specific bandwidth would induce better stereo imaging with any headphone and would ruin the headstage of headphones that have good headstage.
The interesting part is that it does not happen to me, but I seem not to be able to get my brain to do this. BUT when I lower the peak at 10kHz in the HD800S, which it actually has but is 'obscured' by the fake pinna as it nulls there, the stereo imaging does seem to get less and 'sharpness' of the stereo image is lowered a bit.
So it seems 'something' is there. Unfortunately the 10kHz range (a peak there) also gives 'sharpness' to instruments that have such content when there is too much of it or when it is too wide in BW.
Also the HD820 (not universally loved because of its weird FR when one has perfect seal) has a narrow peak at 10kHz and this headphone, to me, also had good imaging and a wide stereo image.
Of course on HATS measurements this is nulled out so you can't see it.
So there seems to be something there BUT your brain has to do something with that. It seems that for the majority of people it does not. Only very few people report 3D effects and if they do it is recording dependent. Perhaps some trickery (phase, reverb, hall or other plugin) is used in those recordings.
As I don't do IEMS it might be interesting for people to experiment with a sharp notch filter (move it around between 8kHz and 12kHz ?) or peak filter in an IEM as it bypasses the pinna. This then should affect the imaging/spatial qualities IF it really is a peak or dip that is responsible for it. Personally I would assume it may only be a small part of the story.
To me it remains a mystery and as I cannot seem to get the same depth I can hear with good speakers (my brain can easily do that) with headphones I am not the right person to look into that closer either.
Interesting subject.... perception.