I have been lurking while this discussion has proceeded, and I have a couple of thoughts to add.
First, and most importantly, please, please, please stop trusting the "Olive" single number sound quality ratings. Even Sean Olive, and all of Harman International, where it was created, don't use it. I have said this several times in the ASR forums. It served its purpose as a statistical factor in the revealing subjective vs. objective correlations of ratings, and persuasively showing that the spinorama incorporates much important data about loudspeaker performance. These positional substitution, equal loudness, multiple-loudspeaker comparisons were designed to reveal resonances, the most audible flaws - i.e. evidence of timbral neutrality.
In these tests the program and the room were constant factors which is never true in real life. So when Sean talks about the resolution of the calculated scores, it refers to a laboratory experiment. However, knowing that the loudspeaker is free of audible resonances is an important start. Evidence of resonances is seen in the spinorama data set - looking at all curves, not just on axis. It is difficult to find loudspeakers that have NO resonances, even the Genelecs being discussed, but they need to be attenuated to below audible thresholds - and we have psychoacoustic data indicating what the thresholds are.
The results showed that bass performance - alone - accounted for about 30% of the overall ratings. In real life the bass extension of the loudspeaker is a very important factor and it is often the primary cause of single-number rating differences - a subwoofer changes that part of the rating completely. Bass smoothness is also a factor, but in real life that is dominated by adjacent boundary effects and small room resonances. These can only be evaluated and addressed after the system is set up.
So, all single number ratings are limited in value below about 400 Hz in typical rooms, and if EQ is available, and directivity is smoothly changing or constant above that frequency, the frequency response can be whatever one wishes. A slightly flawed loudspeaker can be easily improved, or degraded to satisfy whatever whims prevail. This guidance is available if one learns how to interpret spinorama data rather than relying on distracting calculated ratings.
Equalization can attenuate prominent room resonances, but not all EQ systems do it the same way. One cannot assume that all proprietary algorithms have the same potential for bass improvement. The resonances need to be identified in terms of centre frequency and Q before being attenuated using minimum-phase filters of the same parameters. This done, the resonance is truly damped and both the frequency and time-domain problems are alleviated. If the filter is not matched, and most are not, the resonance is attenuated, but not necessarily rendered inaudible. This method only works for the microphone location - a single head location. If non-minimum-phase equalization is attempted all bets are off.
Chapter 14 in the 4th edition discusses this in detail, including explaining how multiple subwoofers can reduce the energy supplied to resonances, greatly expanding the listening area by literally changing the standing wave structure. For most listening situations this is an advantage.
Near field monitoring gratifies a need while mixing, but is not the way to listen if one is hoping that the experience will "translate" to listeners at home in normally reflective rooms or on the prowl with headphones. Apart from the obvious acoustical mismatches, all the problems of stereo are displayed in their ugliness.
I spend time on this topic in the 4th edition and in the accompanying slide shows. Number 6 is interesting in this context.
Finally, from my perspective, today there are an ever increasing number of consumer and professional loudspeakers that are well past the point of diminishing returns in terms of sound quality. Owners with access to anechoic spinoramas and equalization are well situated to maximize overall system performance, including the listening room. All of them are corrupted by adjacent boundary interactions and small-room standing waves/resonances. Reviewers who maintain consistent listening situations for all products being auditioned are doing what they can, but it is still flawed because of human adaptability - we quickly normalize flaws in loudspeakers we listen to for any length of time. It is often attributed to products "breaking in" but it is really listeners adapting to the new sounds, or just imagining things - humans are good at that
Interpreting trustworthy measurements is very likely a more reliable method of identifying timbrally neutral, technically accurate, loadspeakers than the normal stereo listening situation available to most. My new slogan is:
It is important to like the art heard through neutral loudspeakers.
Liking the loudspeakers is the answer to the wrong question.