it's measurement methods. in the nearfield (which is done to mitigate any room influences and achieve a very long 'window' of measurement to measure distortion) there is a high frequency cutoff where the soundwaves are not given enough time to be represented accurately. thats when you have to move to the far-field (ie; 1/2m, 1m, 2m, etc). this NF high frequency cutoff point is determined by driver diameter. for this driver, that cutoff is about 2.5khz or so, IIRC. so that basically means the waveform accuracy of the response above this frequency should be evaluated in the far field. but for distortion, it's still OK because you're only comparing the original fundamental vs the distortion artifacts; measuring further away doesn't alter the result (I did a thread on this yeeaaaars back).

You can read more info about NF/FF (nearfield/farfield) measurements here:

IOW, don't try to match up the HD fundamental to the FF response because the measurement methods are different and that results in different response attributes. A good tester will know this. So be careful of test results you see. I've seen a couple people measure drivers in the nearfield and post that up as frequency response... their results are wrong.

*note: nearfield is typically within inches of the driver, while farfield is feet. the above linked pdf provides measurements that specify, but that's just a general point that may help you better understand what I mean.

Hope that helps,