That's an explanation of the automated sub optimizer called SFM (Soundfiled Management). The previous reference was someone doing manual optimization.
I have never done "foundational blind tests on phase audibility". You must be thinking of someone else. My experiments revealed that the dominant audible differences between loudspeakers were attributable to resonances. Resonances have an associated phase anomaly, but it is the magnitude anomaly, not the phase or the time-domain ringing that reliably relate to audibility. Phase shift, per se, in a channel is not a significant audible factor according to several serious investigations. Finally, I don't see a logical connection to playback sound level.
Ah, now I see where you are coming from - bass reflex problems. Actually, when the air flow in a port becomes non-laminar (not non-linear) the effective mass changes - it is reduced - and the reflex tuning frequency changes, so it is more than just a change in time-domain performance, the parameters of the reflex design have changed. Some years ago Harman engineers investigated optimized port profiles to minimize this effect for various applications - it is in AES publications.
The stapedius reflex, also called the acoustic reflex, is real, of course, and because it is a muscle it fatigues with time and the bass response changes. Some people can manually activate the middle-ear muscle.
BTW it was the Harman International lab, not the Harman-Kardon lab. Harman-Kardon was a small brand among the many that Harman International (now part of Samsung) own. I simply don't understand what "acoustic flaws" you are talking about that cause these alarming sounding perceptual effects when listening at home.
ITD differences are binaural - between the ears - effects that determine localization. I did my PhD on binaural localization and know these effects well. They have little or no connection with the perception of phase shifts in the signal itself, which are single-ear perceptions. I point out in my books that there is "hidden hearing loss", a degradation in binaural hearing that can exist with or even without conventional threshold shift. Even in young people.
This sounds like an opinion without references. Clearly you don't enjoy loud sounds. Hyperacusis is a possible side effect of hearing loss, which itself is aggravated by loud sounds.
Thank you so much for taking the time to continue educating us on the science of audio! There is so much garbage out there in audiophile land that it’s hard not to fall for the hype!
It turns out that resonances are MORE audible when listening at a distance than when listening in the near field - room reflections are repeated versions of the direct sound and they include sound of the widely dispersed resonances. These repeated "looks" at resonances provide the ear/brain with more information. This,of course, applies to the desirable resonances in the music as well - music is more enjoyable, timbrally richer, when heard in rooms than it is outdoors. Concert halls are designed to deliver many, many reflections and are really enormously enjoyable comb filters.
If you have been following my posts and writings you will have learned not to trust subjective "reviews". Trust only comprehensive anechoic measurements, such as the spinoramas. Few reviewers have access to accurate anechoic data, and many don't believe in measurements - they are selling and propagating opinions, not facts.
An acclerometer measures motion at a point. Panels resonate in complex patterns, some of which exhibit equal areas of motion in opposite directions, so little or no sound is radiated, yet vibration is present. That is why scanning laser vibrometer measurements are used. The "knuckle" test is not reliable because it stimulates vibration at a point, while it is distributed acoustic pressure inside the enclosure that is the stimulus in reality.
You are measuring the loudspeaker in a room, and the room resonances are strong. This is not a mystery. It takes two chapters in my books to explain adjacent boundary and room resonance problems, and ways to address them. I am not going to repeat it here. Sorry. The spinorama describes the performance of the loudspeaker in an anechoic situation. In-room measurements add boundary and resonance effects which require separate analysis and treatment. BTW, transducer and prominent room resonances are minimum-phase phenomena, so the time domain performance (ringing) is predictable from the amplitude vs frequency response. Parametric EQ can alleviate both.
I commented at some length earlier in this forum (which thread?) on this topic, and it is discussed in my book. Putting it succinctly, all the indicators in Sean's model pointed to a flat and smooth direct sound with smooth spatially averaged curves (listening window, early reflections, PIR and sound power), having low narrow-band variations. - i.e. smoothness. All this points to an absence of resonances more than it does to the broadband shapes of curves other than the direct sound. You can see from post 14 in this thread that resonances are revealed as bumps in the curves that are repeated in all the curves. OK? Preferences have correlated with the absence of resonances from the very earliest tests which were summarized in the 1985/86 JAES papers and my books. It is not new, and predates the model.