On one hand you could argue that the ear is just the last flawed transducer in the reproduction chain and correction for this could be a benefit. On the other, it is followed by a pretty powerful "DSP" anyway. But who knows what parameters that uses for correction.
And how would your internal DSP interpret a sound system corrected to have a flat response for both ears when everything else you hear is not corrected externally?
My ears are flawed, I could not honestly claim otherwise, nor have I.
My internal DSP seems to be working well enough, though, with the bandwith it has available to work with.
Any live sounds in the room, above the transition frequency (which most are), I should expect to be transmitted to my ears from the source without any particular modification in frequency response, or "flat". (fact check for that below)
Everyday sounds in the room do not excite the low frequency room nodes to a noticeable degree. Full range music does, and at reference levels, causes some dissatisfaction and demands a little attenuation in that area, if you can consider 20dB hot "a little".
For the rest of the audio band, I will, for now, entertain the thought that my speakers might be a little deficient, and allow the microphone to suggest some changes, and turn the automated DSP loose to calculate a adjustment according to the psychoacoustic principles it follows (AcourateDRC) to that area.
My stated "goal" is to make the equalized sonic RTA at the listening position match the raw electrical RTA of the recording being reproduced. If frequency response is the dominant factor in how something sounds, I figure if I can match peak levels across the band, it's a useful and relatively achievable target.
The automated DSP will also make some timing adjustments, again, most obviously to the bass region.
Matching waveforms between that received at the listening position versus the waveform present in the source won't work due to the remaining room interaction (reflection, echo, phase), else I would attempt to closely match the signal's wave at the mic.
"Flat" EQ achieves the stated goal, and presents the recorded signal to the listening position, with only a minimum of obvious errors, I still have a mono bass hole at 48Hz (which usually disappears when the material has the bass recorded in stereo), and a dip at the frequency calculated for "dipole bounce off the wall behind", also present with the monopole speakers, which surprised me to some degree.
If you think the highs are too bright with a "flat" EQ, consider the natural roll-off inherent in any reasonably well recorded piece of music. There seems to be enough attenuation in the highs already present in the recording.
If not, I have a setting saved on the DEQ2496 to provide a "scientifically acceptable based on preference testing" 10dB slope from 20Hz to 20kHz,
I find that slope accentuates the bass again, and don't use it, but it is there to be recalled if somebody should demand it.
---
Dissipation of high frequencies in air:
http://www.sengpielaudio.com/calculator-air.htm
Ok, so at my 3m listening position my "flat" EQ could be argued to be 0.39dB "hot" at 10kHz vs what's on the disc or in the file or streamed.
Below 4kHz less than 0.078dB and becoming even more less with decreasing frequency.
Maybe I'll fix that sometime.