I agree, but then the revel Salon2 are not correct measuring speakers, from what I understand they employ higher order x-overs that just cannot recreate a step response or a square wave, they will not even resemble what goes into a speaker.
This is a step response from a Revel Ultima Salon 2:
Looks horrible....
Compare this to a step response from Dunlavy SC-VI
99% textbook perfect....
So if we say measurements are importane we need to look into all things, also phase coherence and time distortion.... how can a speaker be well measuring if it has a 360 degree phase shift?
Of course the ideal would be no phase shift, but there's no perfect solution, so you have to weigh these things on their audibility relative to other things. In reverberant spaces phase shift is certainly less audible than in anechoic spaces. So if people "enjoy" listening in anechoic spaces I suppose they might prefer designs with less phase shift.
For those that prefer listening in typical or live acoustic spaces, phase shift becomes much less significant, arguably inaudible on the majority of actual source content. A speaker with minimum phase shift, will have other issues. The off-axis response (not only vertical but horizontal), pistonic behaviour, and power handling are all compromised.
I'll say this though, as far as driver interference patterns, as I said earlier I think I would take 360 degrees out of phase over 270 or 90 deg out of phase.
Of course, there exist speakers that avoid these phase shifts for a single microphone (btw, we have two ears though), while drawing many of the benefits of higher order crossovers. The NHT XD comes to mind; it uses DEQX circuitry to get what is akin to 110db/octave slopes, without the accompanying phase shift. Like anything using FIR filters is not a 100% perfect solution, but they're arguably a better one. But as long as the hi fi community is using passive loudspeakers, I don't see that as an option.
Don't forget, cone n dome drivers aren't free from phase shifts either. Or anything with a vent or passive radiator? Even electronics.
Ultimately there's no perfect loudspeaker that i'm aware of, but there are many that are extremely good. You have to weigh each measurement.
Does the step response really matter as much as the off-axis response? Not IMO, in ~95% of listening spaces, to ~90% of source content.
That's where auditions DO come in. You can't decide what a speaker absolute sounds like without hearing it, but you can often get a VERY good idea. Accurate speakers sound more similar than they sound different.
Marketing departments of companies (often with anechoic chambers) making so-called time coherent speakers might want you to think elsewise, but there's a lot more to sound in real rooms than a step response and a 0 degree frequency response. That doesn't mean the measurements don't matter, only that you need to weigh them correctly for their final listening place.
