Regarding age-related hearing loss: it is
not like an EQ. The physiological processes behind it cause a change in the perception of loudness.
Hearing is both cognitive and nonlinear. The discussions about hearing loss mostly look at it as a linear process, i.e., a straightforward mechanical deficiency and correction through EQ.
What happens is that overall relationship between SPL and loudness starts to change quickly as you get on. I've seen studies which show that different subjective levels of loudness become compressed, with younger people requiring more of a difference in SPL to consider a sound at one level vs. another to have distinctively different loudness, and these variances are not the same across the audible range.
So
@Frank Dernie is right. You can't just boost highs or choose a bright headphone/speaker and expect a natural response with hearing loss. It could easily not map to your subjective loudness map.
The cognitive aspect is also important: by "hearing" (in the cognitive sense of recognizing, contextualizing) you are able to place and settle the auditory event into larger patterns like rhythm, tempo, melody (which is a distinctly Western thing, by the way, in the sense of a melody being an abstraction from a more general musical movement). For some people, it is more than enough to follow music by listening mostly to lyrics and accompaniment, or to follow certain beats. This requires less tonal balance and less of an empirically-oriented active memory.
So there is much to be said about personal hearing practice. This is why a built-in treble/bass boost works, as does treble droop (you are forced to listen harder, or perhaps focus less on details). As soon as you learn to judge tonal balance, these features are far less attractive in their own right. From this perspective the main feature of blind testing is its requirement for the listener to adopt a hearing practice that centers on empirical features like tonal balance, spaciousness, envelopment (edit: this is probably why mono listening with speakers tends to produce more discerning results than stereo or multichannel, since those are necessarily more "computational", more inherently cognitive, than mono). Some people are better at it than others, and many are able to learn that practice and carry it with them.
Not saying that this means the Harman curve is necessarily neutral, but it's somewhere around there. It's best feature is the demand for smoothness and no sharp discontinuities across the spectrum.
