They are very closely related, not that I see people acknowledge that.>>59990418
Rudolf mentioned a few things offhandedly, anecdotes about Sonarworks. Hopefully I remember them correctly.
One of them was about developing the DSP. The engineers would keep screwing up the implementation, so they eventually gave up on developing it themselves and hired a mathematician to fix it.
In any case, one sample is not representative of all units made. There is always deviation from the reference model that must be accepted in a finished product. An exact calibration for other models of the headphone is not possible. Some margin of error will be permitted in generalizing the EQ target.
Transducer position itself doesn't matter so much, as it creates distinct shifts in the magnitude of the headphone transfer function. The result will vary as the headphone is reseated and replaced on the head.
As for >>59990515
and the interpretation of coupler results:
-Most couplers are poor representations of the human ear, average or otherwise. Material stiffness and anatomical accuracy are some more notable errors. Simulated drum impedance is ill controlled and defined, although this will have more obvious relevance to measurements of earphones.
-The recorded result at the eardrum is not so important compared to how it relates to the subject's natural transfer function.
-Individual results may not matter as much as one might first assume. These will be concentrated to high frequency deviations within the top one and one-half octaves or so.