That's because there's an important distinction between loudspeakers and headphones, even for as simple a thing as achieving a flat frequency response! Let's examine that.
For example, if person A set up a room with a particular set of loudspeakers and tuned it til it's frequency response is flat at a listening position, then if a mixing engineer mixes at that position and produces a recording based off that, and person B else sets up another room with a flat frequency response at a seating position, then you can be sure that when that recording is played back in person B's room, the spectral balance of the sound at person B's place is wholly accurate compared to person A's system. Now say you and me both go into person B's room and have a listen. Perhaps I like the sound and you don't. But whatever our complaints are, the frequency response of room B cannot be part of problem (if our goal is to reproduce the sound of the mixing studio of person A), because it is flat as much as we can make it and reproduces the frequency response of the room of person A used to produce the recording. In other words, Getting a flat frequency response from a pair of loudspeakers at a given listening position does not depend on a person's physiology.
Now say remove person B's room and you listen to the recording over headphones. Well now, how do you make sure you get a flat frequency response for a given person over the headphones? Without knowing your own head and ear response...you simply cannnot. So one way to get around this is to make yourself a new sound reference that is known flat and try to replicate that. Let's use person B's loudspeaker system that is known flat as the reference! So you go to the seating position and record your HRTF in there, and do and then equalize the headphones using that HRTF. And in doing so will make the headphones sound like the speakers of room B. But at least now we have made it so that the headphones over your ear has a flat frequency response as heard from room B's flat loudspeakers. Getting a flat frequency response from a pair of headphones depends on a person's physiology
I do not want to continue the discussion until we at least can agree on this.
Ych, I am sorry but this is flawed logic. First, what does the issue of stereo recording playback with headphone (when mix was intended for speakers with some amount of cross-talk) has anything to do with the topic at hand (which is existence of reasonable target curve for headphone design)? For instance, just like the headphone reproduction of the recording won't be effectively flat transducer response at our ear canal, neither will the loudspeakers in room A or B measure flat at the entrance of your and my ear canal (using same open ear canal mics as the realiser for example).
Next you're saying our physiology is out of the equation because the speakers response has been measured using flat omnidirectional mics at seating position. Yes fine, but this indeed exactly what the phone equalization curve is providing for !! E.g removing the effect of the measuring head by equliazing out its HRTF for selected headings (e.g. free field response in anechoic environment or diffuse field response in more or less reverberant room). An equalized dummy head measurement gives you exactly that: same flat response as the omnidirectional mic if the head was sitting in the same room as used for the equalization curve with speakers at the same headings as the target curve. Effectively, an equalized dummy head response is like an omnidirectional mic response at the seating position.
What I am arguing is that it doesn't matter what the head physiology is because it gets out of the equation once equalized. The equlization curve is indeed different for any two individuals, but it gets filtered out before looking at the headphone response in any case. What I am saying though is that the target eq. curve is probably much more critical (since it requires to make an assumption on the typical speaker / listener placement and room chraracteristics for measuring the dummy head HRTF).
By the way, the realiser actually does not compensate for individual ears geomety either since the HPEQ effectively attempts to equalize out the pinnae effect for a given headphone (since this effect is already factored in the PRIR)...