Speaker measurements can also be misleading, so many variables from the room itself to the actual microphone placement.
Many enthusiasts would glance at the two graphs below and automatically predict that the first speaker has a much more "neutral" and "correct" sound and that the second speaker has a pronounced "artificial" or "colored" top end.
The problem is that it is the same speaker (which I also owned at one time), so which graph tells the "truth" and gives a better indication of how the speaker will sound?
https://www.soundstagenetwork.com/index.php?option=com_content&view=article&id=839:nrc-measurements-polk-audio-lsim703-loudspeakers&catid=77:loudspeaker-measurements&Itemid=18
https://www.stereophile.com/content/polk-lsiiim703-loudspeaker-measurements
While I agree single speaker measurements can be misleading sometimes, and you need to take into account some variability between measurement sources, we're also lucky to live in a time when measurement availability and consistency is better now than ever. But sure, I personally don't 100% lock in any measurement until I've seen it from two or three trusted sources because I personally know how many things can go wrong when measuring a speaker. Nonetheless, I do still consider them more reliable on their own than written impressions =]
For the most part the trusted measurement sources are quite consistent with their measurements. This specific comparison is one of the biggest 'differences' I've seen, and yet I still I don't see much of a problem. Yes, it's true that speaker enthusiasts might read these individual measurements differently, but I don't think anyone has suggested there isn't
some knowledge/skill needed to interpret the measurements. Plus you only shared the listening windows (measured with different types of averages and methodology, mind you), while the directivity measurements (although presented differently), do suggest the speakers are very similar. Obviously, the more data you have to work with the better.
I would hope many people looking at these measurements would realize this is a speaker with a ring radiator that is
extremely sensitive to positioning in the last two octaves. The directivity measurements already show this last octave or two can vary hugely with small angle changes. So that brightness from 8k-20kHz -- a region that doesn't usually have a dramatic effect for most listeners anyway -- is highly position-dependent, which you could glean from both the directivity data and speaker design.
I'd guess for most people it's a mostly neutral-sounding speaker that can sound a
little bright in the top octave depending on positioning -- like I found with the similar-measuring
Reserve R200 which uses a very similar ring radiator, if not the same one. Hard to tell much more without complete vertical data, but curious what your own impressions were.
Even taking all that into account, the measurements are still very similar, especially once you factor in the limitations of each measurement methodology. The NRC measurements still show some possibility for brightness, just not as much.
Just to the point on positioning, the slight difference above 8kHz points to model variance or difference in positioning, as I saw when I measured the Polk L200 twice, which also uses a similar ring radiator. The first time I did the on-axis (in white) I did not line up the tweeter perfectly with the mic. The second time (in blue) I did a better job of it.
So I suspect that upper octave difference in these two measurements is something similar =]