This has, sadly, been the situation for a while now. I'm sat just 2m away from my hAP ax2 at home with near line of sight. My devices are nearly always on 2GHz (AX) not 5GHz. This, linked with the lack of a cloud controller, means I've gone back to Unifi. Yes, I'm sure there are flaws but clients like the web interface and I do seem to spend a lot less time tweaking them. But we've been over that many times before too.
There are different preferences. For me, I do not want any cloud controller to interfere with my infrastructure.
Half of my clients are connected to my own self-hosted controller. That is a plus with UniFi.
I work primarily in the SOHO space where it's clearly more usable than having to find a Mikrotik engineer in the UK. I wish it was different but I wish lots of things that don't happen, that's life.
AX3 has one of best wifi ranges from any router of those sizes i touched, but here is the thing, its nerfed heavily by regulatory. And whats ironical, some may look at HAP AX2 and AX3 and think, oh AX3 has big antennas, its gonna have better signal for sure compared to smaller AX2 without antennas? Well wrong, because regulatory in for example EU, this 2 routers will run at exactly the 100% same output power, and will have exactly the SAME range.So it doesnt matter how big antennas they put, radio will just drop down power because of total EIRP as mentioned by normis. So if you want to blame someone for bad signal its this stupid regulatory and not Mikrotik gear, or you should move to a different country like US where limits are noticeable higher if you get what i mean 
But again, as mentioned: better (bigger ears) antennas matter for Rx ... and in case of AP this means better station uploads. Now, with mostly asymmetrical data transfers (majority being towards stations) ... does this matter? Mostly not ...
But then what many AP admins forget about is that EIRP limits are not uniform over the whole band. For example, my wAP ax (running ROS 7.21.4) when selecting ETSI country (mind that by default country setting is Latvia, which is ETSI country), shows this:
5170-5250/23dBm/160MHz/indoor
5250-5330/23dBm/160MHz/indoor/dfs
5490-5730/30dBm/160MHz/dfs
5735-5875/14dBm/80MHz
So clearly the best range will give using frequencies between 5490MHz and 5730MHz, 7dB lower range when using frequencies between 5170MHz and 5330MHz and even 16dB lower range when using frequencies between 5745MHz and 5875MHz ... and newer versions of wifi-qcom driver seems to prefer the highest frequencies (which come with lowest EIRP limits and many not-so-new devices don't even support).
To compare, US limits are quite a bit higher in most of frequency range (and in some considerably higher, but in the range where ETSI allows highest EIRP US limit is lower):
5170-5250/30dBm/6gain/160MHz
5250-5330/24dBm/6gain/160MHz/dfs
5490-5730/24dBm/6gain/160MHz/dfs
5735-5835/30dBm/6gain/80MHz
Or Brazil:
5170-5250/30dBm/160MHz
5250-5330/24dBm/160MHz/dfs
5490-5730/24dBm/160MHz/dfs
5735-5835/30dBm/80MHz
5945-7125/30dBm/160MHz/indoor
Additional problem when using frequencies on the edge of supported band is that antennas likely have slighly lower gain there. All antenna patterns I've seen so far (and I've seen many) showed slight (or not so slight) drop in gain on edge of supported frequency range while the single number specified for antenna gain is usually the highest for given frequency range. And I highly doubt that's different for any of Mikrotik's antennas. Which, coupled with already lower EIRP limit in certain regions, means really shitty radiated power for the highest frequencies (and somehow lower Rx performance as well).
Exactly @mkx !
And also would like to add that those higher freqs 5700+ mhz have even lower wall penetration, so combined with big regulatory difference as per your example 5735-5875/14dBm vs US 5735-5835/30dBm/6gain/80MHz that can make absolute huge signal difference to the point of un-usability.
Which makes me wonder why anyone would use 6GHz as backhaul frequency.
Penetration through walls/objects will even be worse.
I guess that line of thought is something like this:
- we need fast backhaul, hence band which allows for highest bandwidth ... and 6GHz allows for 320MHz channels
- we need fast backhaul, hence band where interference will be lowest possible ... and 6GHz will be likely less interfered due to two reasons: low device penetration (at least initially) and higher attenuation of interferers (through walls and free-air)
- we need maximum coverage for end users ... hence we use lower bands (2.4GHz and 5GHz) to operate APs
People tend to forget that running wider channels (e.g. 320MHz v.s. 80MHz) reduces energy per OFDMA symbol which reduces SINR and which then drops number of bits per symbol (and reduces actual throughput). And sometimes drop due to low SINR is worse than potential benefit of having more OFDMA symbols (due to more bandwidth).
Another (perhaps opposite) strategy would be to use lower frequencies for backhaul (2.4GHz AX is not so bad throughput-wise) and use higher frequencies to provide service to end stations. Which reduces service coverage, possibly removing AP overlap (causing connection drops when moving from one AP to another one). Possibly reduces throughput (due to slower backhaul) as well. But it does allow for slightly less optimal relative positioning of involved APs (because lower frequency will allow backhaul connection in worse conditions).
It's not always a matter of "coverage" in regards to distance or square footage as it is in devices-per-radio.
I have customers piling on WiFi-enabled everything (cameras, doorbells, light switches, cow monitors--whatever), to where you log into the AP and see 30+ devices. They then wonder why their Teams or Zoom meetings don't work. (Fortunately most of that IoT gear is 2.4, but that ruins that band for phones and laptops.)
We end up placing a second or third AP in a home to help spread the load across the AP radios. That only works well without making the 5GHz radio busier. With either cabling or a third radio, we have a clean channel for use as backhaul. Since hardly anything is on 6GHz, it's a prime use case for that--for now.
I'm testing 6GHz in my home now and yeah, coverage in distance is pretty bad (barely makes it out of the room). But it's clean, and the only CPE that supports it in our home is my iPhone. So I'm testing a UniFi Device Bridge Switch connected back to the U7 AP and we get gigabit over that thing all day long.
If you really wanted distance between the meshing AP's on 6GHz, you'd need directional antennas to make it work.
I like the Audience's implementation: 2.4 and 5.8 for clients (should add 5.4 to that), and 5.2 for bridging nodes. To get bridging to work with Wave2 drivers, I overlaid VXLANs on top of the 5.2 PTMP links.
Ubiquity's Amplifi Alien had a ton of spatial streams. I wish they still made/sold that thing. It was a beast.