@Thunderlips I think we can disagree with Amir's methodology and philosophy without disagreeing with his intentions and the value of what he's doing here. Come on folks. That's not going to get us anywhere. Evidence and discussion might.
It should be obvious that Neumann's measurement is self referent.
It looks flat because they made it flat, not necessarily because the KH-80 is flat in itself.
If their measurement microphone was broken and had a +3 dB peak at 3000 Hz, for example, they would still trust it, and make their speaker flat, creating a -3 dB dip in its frequency response. And they would publish a completely flat frequency response for it.
The frequency responses that they publish about their own products are flat by design.
This is still a weak argument mr Pio.
Fist of all, the response they publish is not 100 percent flat. It is flat within 0.7dB from 100 10Khz. Forget the "interpolated" graph, which appears to be an average of several measurements. Look at this one instead, which appears to be of a single measurement:
It's fantastic, but not
perfect. I'd say just about +/- 0.7dB without needing any smoothing. If they were just going for perfectly flat, why not go all the way?
Second: Two independent measurements show broad agreement with Neumann's own measurements, the ones in this thread are the outlier. (And we know the likely reasons why!).
I don't know how Neumann and S&R are doing their measurements, but unless they are both using a cheapo Umik-1 calibrated by the same lab, there is absolutely
no reason for our measurements to agree as closely as they do(especially in the highs where my gated technique has more resolution). Other than that the speakers are just
that flat.
You underestimate how repeatable measurements can be with different gear. Want more proof?
Well, the KH80s are packaged separately. I've measured both of my speakers. So here is the measurement of my KH80's I've shared before in white , next to a measurement of my other KH80 80 in blue, measured two months earlier when I was just starting out, measured on a different stand, measured with a
different microprone with different calibration.
This was before I'd gotten another Umik-1 from CSL labs, so this is just the regular Umik-1 with the standard MiniDSP calibration, which leaves a dip above 16K and some unnecessary squiggles. At the bottom is the Klippel's results. Despite the bargain-bin eqipment, mine still show broad agreement with Neumann's and S&Rs up to 16K.
Third and more importantly,
almost nobody here cares about the speaker being 100 percent flat in the sense of being free of squiggles. Such squiggles are likely not audible. The bass tilt, slight crossover dip, and tilt in the treble in the OP
are. Remember, we are more sensitive to wide low Q deviations than narrow ones.
I know measurements aren't perfect, even with a state of the art system. I'm okay with variations. The deviations shown are significantly larger than necessary and could have readily been avoided, but I'm even mostly okay with that too.
Ultimately my biggest personal issue is that the when you open this thread and just read the original review, as most people will do in the months and years to come, you see broad deviations from flat -small ones, but certainly audible ones - without an acknowledgment of why. When we almost certainly know why. The speaker was measured too loud, the stated SPL scaling is incorrect, the microphone calibartion was not used, and the speaker was not measured from the intended acoustical axis. But when people read this review, they won't see that information - they'll see performance that's not fully representative of a speaker.
Yet again, I very much appreciate all the work Amir is doing - it's an incredible effort and potentially industry-changing - but I do think these are fair criticisms. I think I'll mostly recuse myself from this thread as my evidence and arguments are out there. Luckily Amir has taken more care with level setting and I know the speaker measurements will only get better as we go - they are an invaluable contribution to the audio community.