One reference standard instrument to me is acoustic guitar. For instance, if a recording of steel-string acoustic guitar does not have the type of timbre I associate with the real thing - to me, a sense of luscious shimmering harmonics, and often the combination of the woody resonance of the soundboard and contribution of the body, which to my ears often adds a bit of a warm, woody, golden tone
Everyone is free to choose own reference points and preferred instruments. The question is, if what you hear with one acoustic guitar on one recording, will translate to thousands of different recordings. I personally have my doubts, as guitars from recording perspective tend to offer low SPL, very transient enveloping curve and comparably low level of higher harmonics. The sound on the recording is very microphone- and room-dependent, while in practice it almost never sounds unrealistic, harsh or annoying.
But I have in my house the Neumann KH120II, it is as flat as it can be on axis. The directivity is very good, although not as good as, say the Ascilab or a KEF Blade and Reference, but in the world of speakers, it's terrific
These are all examples of speakers which are indeed linear on axis, but all show a severaly colorated off-axis response in this or that way. and window of angles So if you take balanced off-axis behavior as seriously as AJ (or other developers), you would certainly not design or buy such a product.
If you find the sound terrific with all recordings, well, who am I to tell you that it sounds wrong? For others who wonder why concert hall reverb sounds so lower midrange-heavy and dull, instruments lack enveloping ´air´ in the treble reverb, phantom sources of singers are detached from their reverb, almost like monaural sources of a pop recording singing in a perceived vacuum, well, I can offer an explanation and maybe an alternative.
Every single track I throw at it, it sounds terrific unless the source material is trash.
How do you define that the source material is trash, is you have not heard it on speakers with a balanced indirect soundfield in the room? And is this ´trash´ applicable to just a limited number of recordings?
I don't know how you can "voice" one singular speaker for two music genres with opposing subjectively desired EQ curves.
It is a long and iterative process. The key is not to EQ a speaker for one specific recording to make it meet exactly your taste of tonal balance on axis - that is the job of the mastering engineer. And would certainly lead to other recordings, in which the mastering engineer shares your personal vision of tonality, sounding horrible, as the EQ curve is so to speak applied twofold.
The key is to listen to a variety of recordings sounding pretty differently, but all enjoyable and free from speaker-induced flaws (in case this is a hi-fi product) or sufficiently transparent for the mixing engineer (in case of a studio monitor).
I understand if a speaker was "voiced" to compensate off axis non-linearity response, although I would ask why not fix the directivity (yes, I know easy said than done)?
You cannot ´fix´ directivity in the sense of achieving perfectly constant one. Directivity behavior always involved heavy compromises, and I personally find the ones you have listed previously, with increasing directivity index, steps in the most critical band and (lower) midrange heavy reflection tonally, to be subjectively the worst of all. I couldn't stand it even in my bedroom listening to podcasts.
Or KEF saying, their research shows a more aggressive downward tilt is preferable based on the study of multiple test subjects.
Heavy imbalance of the reverb tonality + room properties being far off from usual living room conditions + circle of confusion at play with the choice of recordings? It is difficult for me to imagine an explanation why someone would apply additional downward-tilted EQ on a speaker which already results in like heavy downward-tilted reflection tonally, particularly lacking brilliance and lower treble.
all of these types of "voicing" are still trying to achieve a linear in-room response, regardless of the tilt, not adding any subjective or artistic EQ bumps here and there.
Sorry, but a tilt with broad-banded deviations from the linear curve, is the ultimate form of coloration. Like a tone control you cannot deactivate. And adding downward-tilted EQ to a soundfield which is already heavily downward-tilted, I would not see as a measure to achieve linear in-room response, not at all. Could you elaborate on that goal of linear in-room response further pls?
How do you define which ´bump here and there´ is ´subjective/artistic´, and which is meant to aim for a more balanced tonality?