For maximal flatness, I ran AR3a with HF at Max, and MF at Mid, substantially in accordance with the original factory recommendation. Still, it achieved that only at the -5° axis. Allison later provided the requisite fixed-resistor balances for "standard" operation:
Best I have figured thus far, based upon earlier clues from Steve F, Kloss distinguished his fundamental "Octave-to-octave balance" design principle as having flat frequency response as opposed to Villchur and Allison's pursuing flat power response. It's all interrelated, of course -- flat frequency response in combination with wide broadband dispersion implicitly produces flat power response.
[Well, in THEORY, at least.... ]
Still floundering trying to draw a correlation between what you hear and what you measure? That seems to be a universal problem, you are not alone. Leo Beranek published two of his papers on his web site comparing 20 measurments for 59 well known concert halls in one and 20 measurements for 29 opera houses in the other in which he attempts to correlate which measurements correspond to the preferences of golden eared "experts" such as conductors of symphony orchestras. There is a difference between that and doing the same for sound reproducing systems. In the evaluation of acoustics of a performance venue, that factor of acoustics is not only known to be but is supposed to alter the sound to make it most pleasing to the most people. It becomes in effect part of the musical instrument itself altering the way in which musicians perform in it. It does not have an objective goal, but an entirely subjective one. The sound recording/reproducing system on the other hand is supposed to introduce nothing of its own, at least if it is defined as high fidelity according to our one time understanding of the term.
Here's a hint. It's not what comes out of the speakers that matters, it's what reaches your ears. I'd heard Snell say it around 1982 or 1983 but the same thought occurred to me in 1974. This of course has to take whatever influence the acoustics of the room that the equipment is installed in plays. Here's an added wrinkle, it also has to take into account the variables of the recordings themselves. They are made under a bewildering variety of variables. Why do sound systems that sound "flattest" sound best? Because people who make musical instruments try to create them with the most pleasing tonalities and that is what confers much of their value to them. Therefore accurately reproducing something that is pleasing to begin with should give pleasing results. How do you do this? Now there's a tough engineering question. Which sound systems do this well? I'm sorry to say that the real and honest answer despite the anger and insults the truth invariably provokes, is that none of them do, not consistently, in fact almost never. It is still beyond the state of the art. That is why the LvR demos were so valuable. It proved at least within the limits of highly contrived conditions it was possible. But those conditions were very specific. It was expected at the time that technology would advance to where that became more and more common under a greater range of condidtions. It didn't. Quite the opposite happened, the technology for all its advances effectively went nowhere, in fact if anything it went backwards. And in refusing to admit it, those who work in that industry redefined the goal of their industry to exclude what they couldn't accomplish. Therefore with no objective goal of accurately reproducing musical sounds whatsoever, you can sell anything you want, come out with an endless parade of products, and claim that each one is the best. You can also charge whatever you want for them and claim it is the fruits of your valuable "research." So wires costing 1000 dollars or a pair of 2 way 6" speaker systems costing 4000 dollars exists and doesn't even seem to raise an eyebrow.
BTW, the two factors that Beranek found correlated best with concert hall preference were BQI (Binaural Quality Index) which = 1 minus IACC (Inter Aural Cross Correlation) and bass response. In a recent paper he also explains the importance of Listener Envelopment Factor (LEV) and how it is now discovered that reflections from the ceiling and back of the room play a critical roll. All of these factors are among those the current speaker design philosophy does not do well at reproducing. It's also nice if you can enjoy the best sound without having to sit whre X marks the only really good spot. Otherwise you'll have to limit yourself to listening alone while others either suffer lesser sound or do something else.