jbmjr,
Your statement, taken at face value, is misguided. Of course every speaker designer is always working to improve musical accuracy and fidelity of the products. But you should understand that the NRC curves must be viewed as a family of curves, and that they require careful interpretation by experts, read in context with data from double-blind listening tests of the particular speakers in question, especially if you are making comparisons between different brands and models.
Note also the date of the measurements. For example, the curves for the Axiom M80ti, measured by Soundstage in August, 2001, do not represent the performance of current production M80s. Speaker designers are always seeking to tweak the performance of models, and changes may occur in production. Part of my role at Axiom is to occasionally suggest small tweaks or refinements to the performance of a particular speaker, if it's achieveable without making other compromises.
Looking at one set of NRC curves in isolation is not very useful because the anechoic NRC curves do not tell the full story. The curves can be a reasonably good predictor of gross spectral aberrations or general trends in response. Some small irregularities can be insignificant in the upper octaves; but a nasty spike in a portion of the audio band where our ears are most sensitive (the midrange) may be quite audible. However, if the spike isn't repeated in the off-axis (45, 60, and 75-degree) measurements, it's unlikely to show up in the listening tests.
Moreover, small dips in response are inconsequential, whereas a large hump over a broad band of frequencies will almost always be heard in listening tests.
Even two different speakers that have generally similar, and smooth, frequency-response curves both on- and off-axis, will never sound identical in listening tests. "Similarly good," is a phrase I often used in blind tests where it was impossible to rank one speaker over the other, or they were ranked by a group of listeners within a few tenths of a point, which would make any tiny differences more likely the result of the listener location rather than inherent differences in the speakers.
The acoustics of speaker performance in rooms are so complex that technical measurements, at best, should only be used as a general guide. As much as we would all like the curves to be an absolute predictor, so the more obsessive part of our personality could say, "clearly, this is the best speaker," the science of speaker measurement is not there yet.
Regards,
