You can safely treat a situation, where sources are <1/4wl apart, as 1d. Eg a 3-way with steep slopes between low-mid. It needs a simple phase unwrap so the resulting group-delay would be below the audibility threshold. Even with wider spacing, the arriving reflections can be mangled, but the direct sound needs to be half-decent.
Yep, keeping all driver center to center spacing within 1/4wl would be best case for successful global correction.....
But realistically, how often does that happens? Particularly maintaining 1/4wl throughout the full critical crossover summation range? (which I'd call at least thru -20dB response)
And
how often are passive crossovers steep enough to confine the full critical summation range to stay within 1/4wl? I think kinda never...
I also think both low-mid, and mid-high, need to keep c2c within 1/4wl. Not easy, especially on any kind of flat baffle.
So yeah, on paper a 1D correction could be used for a 3D requirement, if the 3D space/requirement is compacted enough....I just don't see that happening very often, if at all.
Which means most likely the global correction is correct only to the specific spot of measurement.
Admittedly, I strive for the best results possible out of a speaker's acoustic design. I personally can't see that being anything other that multi-amped, driver sections individually corrected prior to crossover, and then tied together with linear-phase xovers.
Which means for me at the speaker design level, FIR needs to be available on all channels, and not just for global input..