We are discussing on ASR that phase/time alignment can be resolved with DSP/room correction. Speaker brands like Thiel, Vandersteen an others came up with hardware solutions to solve the Phase/time alignment problem using specific designed drivers, slope designed baffles an first order crossovers which are a pain to get it wright.
I don't know what to belief. Thing is i own Vandersteen who are from design Phase/time aligned on top of that they are corrected for Phase/time by my DPS. When i listen to same sort of far-field speakers who are not Phase/time aligned by design i expect after DSP correction that both sound more or less the same for instance regarding depth/staging. Still after DSP correction i hear a big difference between the two different designed speakers.
An no it is not subjective the difference is IMO day an night. Could be that I did not have the opportunity to listen to the latest speakers who could sound differently.
What is your opinion possible empirical/theoretical experience on both.
Theoretically, with first order filters combined with "wide passband" drivers that can handle the power and excursion, you can set up a two-way system, that combines acoustically to a flat response with zero phase. You then need to align the drivers physically so that there is no phase difference due to distance. I don't know how well this can be done, but do remember some old measurements, where seemingly the time domain aspects of such speakers were quite good (impulse response, square wave response).
For a general speaker, you should not "distance align" willy-nilly, without considering the phase response of the drivers and the cross-over first.
Since a speaker is generally a non-minimum phase system (it is a sum of several (generally) minimum-phase systems; each driver with its crossover part), it is not directly 'equalizable', since we run into trouble with stability and causality, when we try to invert such system. Causality issues imply that we need to be able look forward in time, and we typically think of systems as being causal. But with DSP, we can take in inputs for a while and hold them in a buffer, and thereby form a new timeline from which we start, and have the output be dependent on both previous input ("new zero time") and the recent data ("future signals"), and correct, albeit with an overall time delay as a result, as can be seen in the measurements on e.g. KEF LF50 Active and Kii Three.
Erin showed recently that the amplitude response is also affected (
https://www.erinsaudiocorner.com/loudspeakers/kii_three/), which makes it a little difficult to compare even a single speaker with itself in different settings, when you are only interested in the temporal difference. So comparing two speakers with each other, I would probably not expect them to sound the same, since they have different amplitude responses(?) and raditation patterns.
A lot of journal papers show that having optimized transient/temporal behavior is not worth pursuing, and that was what I just assumed. But listening to the Kii Three's myself when visiting Purify at some point in time where Bruno was visiting, I changed my mind, as I in a blind test could quickly pick out which setting it was in, without having ever listened to them before (and generally starting out thinking how to put in a nice way, that I could not hear a difference). Granted, the track (that I also had never heard) had a certain characteristic, where a lot of instruments had to fall in place at the same time, and somehow I picked up on that particular second or less, and right there was a difference.
Time alignment via displacing drivers will lead to big differences in diffraction and so it is best for comparison purpose to do these things with DSP.
Don't know if that answered anything...