The interesting thing I found many years ago is that the better the intrinsic SQ is with regard to lack of distortion, audible artifacts, then the less important is the FR. I just accepted that as a by the way then, but just recently gained the understanding of what's happening: the field of Audio Scene Analysis explains the process of hearing as being a balance what what actually impinges on the ears, and what the brain expects next, as a result of a lifetime of acquired knowledge about how sounds behave. The clever thing is, that our minds automatically adjust what we perceive so that it always makes sense, so long as the actual sound waves coming in confirm that inner prediction; this is why real things always sound real, and we can move very close and very far away to a musical instrument, say, and its tonality and presentation are always consistent - we 'expect' a certain quality depending upon everything, and the sound waves then confirm it.
All bets are off with most audio systems, because often the sound waves from it contradict what the mind thinks should be there - the system is insufficiently competent to get the messages across well enough. Result, the rig doesn't sound good, it's obviously "fake".
So getting back to FR ... if the sound reaching the ears "fits the mind's template" then all EQ'ng is done in the head, automatically - it continually balances the the ear's input with what it anticipates should be the reality, the correlation is maintained at all times. And this gives one, in the head, convincing sound.
This may not work for all people, but it certainly does for me.