I think that what has been said earlier is that listeners don't actually prefer an in-room curve (which I think is what you are meaning by steady state i.e. not anechoic or gated to be anechoic). What they prefer is a neutral* speaker as measured anechoically regardless of whatever room it is placed in. This gives a resulting in-room steady-state curve that will be different for each type of room.
But if the speakers aren't neutral (very few if any are of course) there is an unresolved ambiguity about what the listener prefers. Thinking that it is 'the steady state in-room curve' is jumping to an unwarranted conclusion.
There may even be a correlation between apparent preferences and in-room curves, but thinking it is the curve that is causing the preference is a red herring.
*neutral meaning flat frequency response and uniform dispersion at all frequencies - but it can still be directional, not omni.
@Cosmik , instead of relying 100 percent on a first principle - which is in your case is "a flat speaker of low distortion is all you need" - I like to measure things to understand a bit more.
Genelec showed at the start of the millennium (see Mäkivirta and Anet:
https://www.genelec.com/documents/publications/SurveyofMonitoringConditions.pdf) that a flat speaker may not be all there is to sound. So instead of a one-factor model - where the speaker is the only factor - they introduced the room as a new factor.
Since then, several "room correction" products were introduced.
In 2009, Olive writes: "The sonic benefits of these room correction products are generally not well known since, to my knowledge, no one has yet published the results of a well-controlled, double-blind listening test on room correction products" (source:
http://seanolive.blogspot.no/2009/11/subjective-and-objective-evaluation-of.html ).
And he concludes that room compensation products are not equal. Some room compensation products improve sound, some harm the sound. And the uncompensated sound - the one factor model - outperforms in some cases and underperforms in other cases.
To conclude: Olive's research suggests that room compensation done right (and wrong!) deserves further investigation.
Regrettably, Olive doesn't define the measured sound. Is it steady state or is it direct sound? To my knowledge, Toole also uses the term "steady state" without defining it for other researchers to replicate and investigate his findings.
Say measured sound can be divided into two categories:
1) Direct sound (DS)
2) Steady-state (SS)
And, DS≠SS.
Further - and now I am on thin ice because audio is not my field - I assume the following:
SS=DS+R
where R is reflected sound.
Interestingly, in anechoic chamber, R is zero. So in free field - where there's only DS - the correct curve is flat, not tilted.
I don't know exactly how to define DS as opposed to sound in anechoic chamber sound (AS). What I assume, is that DS is an approximation to the sound in anechoic chamber (AS), however with some real-world complications. I assume the following:
AS=DS+P
where P is real-world problems (outside of anechoic chamber).
Modern measuring devices let us measure both DS and SS. Therefore, I would have liked to see how listeners' preferences may change if they are exposed to DS only or SS. I haven't seen research that look at listeners' curve preferences in both the DS and the SS case.
For Genelec it seems to be important to focus on both DS and SS. See the following publication:
https://www.genelec.com/sites/defau...gues/genelec_monitors_in-room_performance.pdf
They advice their customers on the use of their products; both the shortest distance from the monitors (due to the summing of drivers problem) and the critical distance (where DS still dominates). Source:
https://www.genelec.com/sites/defau...gues/genelec_monitors_in-room_performance.pdf
According to Genelec, there is a critical distance where R has a higher sound than DS. Both minimum distance (minD) and critical distance critD can be calculated.
So what confuses me, is that I haven't seen Toole or Olive define exactly what they mean by SS; they haven't defined minD, critD and R of their audio setups. The magnitude of the curve tilting (in dB) could be depending on where the measurements where made; the higher R is, the more tilting of the curve.
In other words, if DS is an approximation to AS, and the listener is exposed to DS, the "correct curve" (CC) is the flat one. As R increases, it seems like the tilted curve is CC. Therefore CC is not a fixed variable; it depends on the relationship between DS, R and SS!
So let me hypothesize a bit:
1) If exposed to pure DS, CC is flat.
2) If exposed to "pure" SS, CC is tilted.
3) If exposed to a mix of DS and SS - i.e. where R is a factor but not dominating - CC is somewhere between flat and tilted (as per point 2 above).
A more intelligent algorithm would know about both (1) and (2) and give the user a dynamic mix based on measurements of both DS and SS.
To sum up:
=> In the 1980s, progressive speaker producers solved the problem of flat and low distortion.
@Cosmik believes we are still here, but others were not so sure.
=> Around 2000 progressive speaker producers went from a one-factor model - where the speaker was the only factor - to a two-factor model where the room entered the equation.
=> Around 2005 progressive speaker producers went from a two-factor model - where the room was accounted for - to a three-factor model where the room was divided into both DS and SS (due to controlling for R).
So in order to optimize sound in the listening position, you'd need the following:
1) A flat speaker of low distortion.
2) Room compensation where both DS and SS are accounted for.
3) For reference listening, the listener needs to ensure he is exposed to primarily DS in order to avoid colourations by defining minD and critD.
As said, my field is not audio, so take my attempt at putting my confusions on paper with a fist of salt. Don't hesitate to point out where I'm "hearing voices in the air"