Seems a bit far fetched to do all this work for a bit of curiousity.
Applying EQ to higher frequencies can have negative results because of unpredictable room interactions.
Not an issue with headphones as room interactions don't apply.
With JRiver applying EQ via Loudness to higher frequencies there is no consideration of the fact that speakers on and off axis response can vary wildly.
Not so much an issue with bass frequencies as they are far more omni-directional.
There's a reason why room correction software does not work or can be made to not work on higher frequencies.
Take Dirac Live (which JRiver supports via VST) for example the basic version only works to 500Hz (the most important) and the full frequency version can be bandwidth limited with the use of curtains.
Even if one chooses to use Dirac Live full range (like I do) it comes with the benefit that it relies on measurements from multiple positions around the room to make smart decisions about what should and should not be equalized.
Unless I'm mistaken the Fletcher-Munsen curves were based on listeners listening to headphones and not speakers.