On-topic: I acknowledge
@gnarly 's
measurements, and have engaged in private conversation with him just to clear up on the details - need a bit more documentation for complete interpretation. So far, I will say these seem like the best nonlinear distortion results I have seen from an MEH up to now.
Like said in private message, here's the mdat for the measurements I posted...for any/all who care. Zoom in to see ripple increase, and pls tell me if you think there is evidence of chronic, systematic non-linearity...cause i sure don't. (change zip to mdat of course)
So, as of now, I think I got most of the info necessary to properly interpret these measurements. I did ask for the exact measurement setup (box placement, microphone placement, ..) and did not get an answer on that, so I'll assume you measured with the box in place, as pictured
here, which might result in some baffle extension along the subwoofer it is placed on, but which should be negligible.
As for the rest, the noisefloor and higher harmonic levels in your measurements indicate excellent SNR, while no added microphone distortion in the large signal runs can be observed, so you obviously have good measurement hardware, and I don't need more info on that - and I'll just assume the SPL calibration is correct, since, well, that's really not too hard to achieve. Finally, thank you for uploading the source files, which makes evaluation easier.
Compression:
Excitation was a standard REW sine sweep; which is important info in regards to energy density (crest factor). A sine is a much denser signal than for example stochastic noise, and as such will expose compression more.
I exported the 86 and 116 dB curves, level matched them, and subtracted the 86 dB curve from the 116 dB curve, to isolate compression. This is the result from 100 to 10000 Hz (everything above and below is negligible imo):
On its own, there's really no issue here, with max. compression (or amplification) at around 1 dB at 116 dB. This result is also much better than the SH-50's.
A direct comparison to the same drivers, appropriately installed as direct radiators, would be the gold standard here, but we don't have that. So, just to have some reference, I'll throw in a comparison by proxy - the same TW Audio M12 I already used in
this comparison, a 12/1,4" PA speaker with about half the LF membrane area and half the baffle size of this Syn 10. Same setup, methodics, and hardware (GPM on hard flat surface free field far field, Isemcon EMX-7150 & SC-1, etc.) as listed in most of my threads and videos. I only had 95, 110 and 120 dB GPM runs saved for this one, so the comparison is between 95 and 120 dB; excitation was also a sine sweep:
This is a harder (higher SPL) compression test on a smaller speaker, and yet, while either's results are acceptable imo, the direct radiator doesn't display the multiple peaks and troughs of the MEH (~130 Hz, ~500 Hz).
Harmonic distortion:
Again, for lack of a direct comparison, one by proxy; avg. 116 dB both; M12 in 4pi this time; H4 mostly in the noisefloor in my measurement; graphs with identical scales:
Very comparable distortion results of the smaller direct radiator vs. the larger MEH. And the smaller direct radiator obviously radiates its SPL to a wider beam as well.
All in all, as I already said, these are the best nonlinear distortion results I've seen an MEH achieve, and I don't see an outright issue here either. However, there's also nothing that would invalidate the conclusion of my original work of increased distortion and compression in MEHs vs direct radiators.
That's it for me here. If something terribly important should arise, PM me.
. I've also attended many unamplified performances by my local symphony orchestra and choir. Good compression drivers on good horns/waveguides sound just fine in my experience.