Crossover circuits are not inherently a problem. All that really matters is how closely the compression waves of air follow the voltage signal over an area in front of the speaker. It doesn't really matter how you arrive at that goal, at least if your priority is fidelity.
If you want to talk about polar responses, what in-ceiling speaker is going to do that better than one that uses beamforming? The Perlisten speakers do not have the radiation pattern of typical two or three-way speakers, and I am sure you have seen
the polar maps. A 5 1/4" full-range driver can not have a consistent polar response through the audible frequency range. It will narrow its dispersion, probably around 2kHz to 2.5kHz. You would be lucky not to have it beam like a laser by 5kHz. Now compare that to any polar map of any Perlisten loudspeaker.
I will concede that the simplicity of using a MarkAudio full-range driver is probably the better way to go for the majority of people on account of its cost. Not many people can or should spend $2k+ on an in-ceiling speaker, but if you really want high-performance, Perlisten is the way to go. Just because something is complex doesn't mean it is worse. The only other one that looks like a good fit for that application is KEF's THX in-ceilings, but there is no way that the KEF speakers will have the dynamic range of the Perlistens.
Crossovers are inherently a problem. For two reasons. The first is that they do significantly tend to alter the polar response, which takes great skill to minimize. Secondly they cause inherent phase and therefore time discrepancies. The latter in the analog domain, which can now be ameliorated in the digital domain in active designs.
I know many have claimed that these time discrepancies are of no or little consequence. However many, and I include myself believe this is not true, and over time more are coming to our point of view.
In a design with spaced drivers and an analog crossover the polar response will always be skewed in some fashion. For a ceiling Atmos speaker, I maintain this is deleterious.
Now I agree that uncontrolled cone break up narrows the polar response drastically. However the work of Ted Jordan and the work we did on on the Jordan Watts module, changed that paradigm, with the unique design of the cone, that avoided uncontrolled break up. I feel highly privileged to have taken part in this development.
This was all covered in patents.
From the late fifties to the early seventies there was no other driver that had an FR anywhere close to as good as the JW driver. The problem was limited power handling and above all reliability.
Now the successor making use of these patents is in the hands of Mark Audio. and
E.J. Jordan Designs.
The driver the CHN 110 is the closest successor to this technology.
Now I agree that an Atmos speaker has to have wide and even dispersion, and preferably a circular distribution and devoid of lobing issues that crossovers introduce.
I agree that most full range drivers have serious uncontrolled break up in the midband, that rule them out due to narrowing dispersion due to this break up and also a rising response.
This is not the case with the drivers I have selected.
This is the FR of my Atmos speakers.
The Blue line is 1 meter in axis, the orange line is 1 meter 30 degrees off axis.
The purple line is at the MLP!
Now I maintain that result is more than adequate for an Atmos speaker, and an excellent superior result.
I highly doubt those Perlisten speakers will be free of lobing issues, but you are welcome to prove me wrong.
I maintain the drivers I have selected are close to ideal in the ceiling Atmos application and at very reasonable cost.
In my view only a good full range driver or a good coaxial speaker can really perform well as a ceiling Atmos speaker.
Lastly I would point out, that I have around 70 years experience with speakers and their design, and I am not in the habit of developing or proposing anything other than optimally engineered designs.