Since I only use over-ears, we are probably working a bit at cross purposes. I don’t really have any magic formulas for achieving any of the above though with over-ears, other than simply to aim for a neutral response, with maybe a little Fletcher-Munson added if you listen at lower volumes. Many audiophiles prefer to listen at higher volumes though where Fletcher-Munson is less useful.
For best results, start with a headphone that’s already close to your target FR, with less distortion, and less modal breakup potential, and good extension in the bass and treble. Because large changes with EQ (especially at the ends of the audible FR range) just don’t work as well.
There are also some good reasons why you might prefer more or less upper mids to get better imaging, detail, soundstage, or whatever for different recordings. But I think that’s mostly related to the way the content is mixed, mastered, and produced. And probably not much more.
Most good pro audio mixing boards have EQ’s to control the levels in the upper mids at around 2 or 2.5 kHz, because this is where the phantom center on stereo speakers can drop out. And it’s where the midrange and tweeter drivers on speakers tend to cross over.
Since headphones don’t suffer the same issues, a mix or recording that’s been crafted to compensate for these flaws in speakers might tend to sound more hot or strident on headphones that do not have a similar type of flaw or “BBC dip” in the same frequency range.
Beyond that, I dunno. 
I have my own model for neutral (DF+SP) that I prefer, that is rooted in the Harman research, but also takes it a few steps further imo. I believe the best model though is the in-ear response of good speakers in a semi-reflective room. And there are certainly some ways you can experiment with getting close to that using mics in your ears, or maybe doing other kinds of listening tests to tune your headphones.