Wow...just went through this thread and I'm...uh...unpleasantly surprised at how it has gone.
First, I'm as appreciative as anyone else that Amir is now taking the time to measure speakers. However, things could have gone better here.
That's nonsense. I am trying to get you guys to stop obsessing over small differences in frequency response. I have not even bother to put in the calibration for my microphone in there because 1 dB here and there is not material.
The difference in frequency response shown in your measurements vs the others (which tend to correlate more), should be audible deviations from what the manufacturer specs imply, and from that implied by the measurements taken by others which don't have that bass droop or high end tilt.
Since when did this site stop caring about audible differences, even if minor?
If the manufacturer recommends a certain listening axis to realize the flat response specs they claim, and they have been designed to be used that way by the user, what is the point of measuring "off" that axis? Sure you'll find out how they measure off the recommended axis...but we would want to know how they perform ON the recommended axis. One may as well report how a medication performs when not used as directed.
It's like testing a medication with an attitude like:
"This medication says take one pill every 6 hours. But the other brands say take one every 12 hours. For constancy sake, I'll stick to taking all brands only once every 12 hours, including the one that recommends once every 6 hours."
Well, that's not going to be very precise to find out if the medication works as directed!
This seems - I don't know how else to put it - surprisingly sloppy for a site with "science" as part of it's title and remit. It can lead, as it appears to have here, to the results looking like the manufacturer was exaggerating it's claims. That's not a fair stain to leave on a manufacturer if it's due to the experimentor's decision to *not follow the manufacturer's own recommendations.*
What do you think your ears are? That they care about these tiny differences when assessing two different speakers?
I find this very odd. You have sweated over incredibly precise differences in electronics and DACs, even bothering to go over measured differences that are inaudible, and caring about thresholds as to when distortions and deviations *just become audible*. Why all of a sudden are audible deviations from flat in a loudspeaker (what should be per your measurements) something you or we aren't supposed to care about? Especially, again, given the point of that speaker design being they *should not* have those audible deviations if the manufacturer's claims for flat response are correct.
In fact, given your care in being precise in rendering DAC and amp measurements even when the sonic consequences are below audibility, so we can see the competence of the design and whether it's reached specs, it seems a similar concern would have carried over to testing loudspeakers. Even if the off-recomended-axis measurements you produced displayed deviations from flat that wouldn't be audible, they would still be
measurable deviations, and don't we still care about measurable deviations from an equipment's claimed spec? To try to be as precise as possible it would still make sense to measure based on manufacturer-intended axis.
I'm just as puzzled as many here by this whole thing.
Having followed this thread I can't help but be left wondering at least somewhat about the reliability of the measurements in future reviews. Simply redoing the test would have been an obvious way to mitigate the concerns raised by a significant number of readers. Instead, you have taken the worst possible reading of these concerns waving them off with a "Why bother? You'll never be satisfied!" type of response. I think you could give your audience more credit. They are all expressing appreciation while telling you the problem they are seeing in the measurements, and telling you exactly what step would gain their trust in the measurement process.