If I run a frequency sweep through a pair of headphones and all of the frequencies appear the same volume to ME, then YOU put them on and run a frequency sweep, the likelihood is that YOU will hear some dips and/or peaks because OUR hearing is DIFFERENT. Is that STILL too confusing?
There are different things going on here.
1: HpTF is individual for each person but isn't so much about tonal balance being different but treble peaks differing.
2: Individual HpTF is not the same as specific HATS minus correction curve.
3: Personal preference exists but should not be in a standard.
4: Standard conditions are not ones personal conditions which creates personal 'standards'.
5: Flat can mean different things. It can mean 'smooth' (as in no wiggles), it can mean 'perceived loudness at a specific average level' it can mean equal SPL at specific measurement conditions.
The correction curve for a bunch of headphones will differ because of HpTF of the HATS when measured.
This means the Harman correction curve for their specific HATS will have to be an average.
It's why the correction curve is 'smooth' (because averaged) which creates measurement errors for all headphones.
The higher the frequency and the closer the frequencies are in the area where the HATS pinna and earcanal have an influence the bigger the error will be.
When measuring it is best to have an accepted 'standard' which to compare to.
There weren't any that were really applicable to headphones.
Flat, as in SPL over the audible range and smooth (least amount of wiggles) measuring speakers in an anechoic room used in the 'Harman listening room' which is supposed to react similar-ish to (again averaged) listening rooms (rooms where the stereo is usually setup).
The 'flat' sound is modified in the room. This is complex because SPL measurements (not gated) at the 'listening position' differs from what we hear per frequency band. Ears are not microphones.
Harman research showed that most people preferred the sound of speakers that had a smooth response (the least amount of dips and peaks) in their 'standard room' with no modifications of the sound signature (personal preference). Those speakers measured SPL-flat at 1m in anechoic conditions and had good dispersion properties. When such speakers are used in a room (with reflections at listening position) then the recordings sounded realistic. I assume at 'studio listening levels'.
When you measure a headphone's SPL and that would be flat, as in no wiggles and SPL the same for all frequencies the plot will be flat like the perfect speaker would measure at 1m in anechoic conditions on axis.
The thing is no one listens that way to speakers.
Yes, some folks do listen to speakers at 1m (nearfield desktop setups) but not anechoic and bass boost (from the room) will also be 'added' and the brain won't be able to discern low frequencies. Most will listen to speakers in a room at several meters away in furnished rooms.
This means that
flat+smooth SPL measured in headphones is not tonally correct like
'flat + smooth' speakers in a room
These 2 have a different sound.
When the goal is to get the correct tonal balance recordings are designed for (flat+smooth speaker in a room) you will have to 'mimic' room effects. Here is where things go sour.
Everybodies room is different. Just like HpTF differs from person to person.
To measure anything you need to have a reference. For most things in life measurements can be fairly exact and when a standard is used we will get similar results.
Not so in everyones room and not so with HpTF, leaving preference aside.
So both HpTF and 'speaker in room' will have to be an average. There is no way around that.
Assuming at Harman they knew what they were doing and established standards (using human hearing + objective measurements) they arrived at their standard which (mostly) differs from SPL measurements in the lows and treble. Bass because of the room, treble because of speaker distance/room and averaging. It is turned into this standard when measured on a specific HATS with a specific configuration and averaging.
So now we have a standard that works in the Harman room which is supposed to mimic 'average' listening rooms with better than average speakers.
That doesn't mean that each individual speaker in each individual room at each specific listening position/distance will yield the correct 'tonal balance'.
Alas, for most people, that sound of their selected speaker in their room with their recordings and music preference will be 'their' reference.
That may wel NOT be the intended sound (optimal speaker in 'optimal' room) but your brain get's used to it.
Just like when listening to 1 headphone only we get 'used' to sound signature (and generally think it is burn-in of the headphone).
The closer the speakers in that room get to the sound sig of the headphone the easier we accept it as 'real' and thus how we perceive bass.
Now here's the thing. The serious listeners all EQ their room/speakers and thus, in general, have to EQ the lows. Mostly downwards in level.
These are usually EQ'ed to a 'gentle sloping' curve and thus deviate from 'flat+smooth speaker in room with no treatment'.
You get used to less 'room made bass boost'.
Chances thus are that audiophiles with treated rooms/room EQ will get used to and prefer less 'bass boost' than the vast majority of 'standard listeners (with room bass boost).
So... in short your (and mine) amount of bass boost differs and amount of treble may also.
Let's all live with our own 'truths' but lets measure according to a standard. Harman is such a standard. And while not perfect for everyone it is a scientifically determined standard and 'better' than FF or DF so why not use it ?
This shouldn't be in an LCD-X thread though.
Maybe these posts should go in another thread where the discussion can easily be found again for those looking for Harman things.