As the subject states........what does it actually measure when reporting these numbers?

SNR, I *know* the SNR being reported will be the RSSI LEVEL - NF

However, what exactly is the NF, where does it come from? (I know what the NF is, and what it means, in general RF terms, as I've been in the RF field since 1985...this pertain to Mikrotiks usage of "NF" and "SNR".)

I did a test. At one end of the bench I placed an RB411AH w/Ubiquiti XR5 radio card, with a small antenna just laying on the bench.

On the other end of the bench about 7 feet away, I placed another RB411AH w/Ubiquiti XR5 radio card, with a small antenna just laying on the bench.

These two radios are configured to "talk" to each other, and do, in the -40dBm range. I've been using this setup for developing some utilities

Now I add a third radio, in between the two radios, another RB411AH w/Ubiquiti XR5 (brand new), with a small little antenna laying on the bench, and do minimal configuration to it, enable the radio, make is a 'station', etc, being sure it WON'T "talk" with the other 2 radios.

The first set of radios now have a band-width test started between them.

The third, non-associated radio in-between them, using the serial terminal, I log on, and start the frequency monitor.

As expected, the frequency monitor shows a very high percentage of usage, 72%, on the channel the BW trest is running between the other radios.

However, the NF being reported is still -96. Since this is an un-associated radio with completely different operating parameters, I would think it would interpret the other signals (from the pair running the BW test) as environmental noise, and thus, be reported as NF.

I'm thinking I'll be told this is a Ubiquity problem.

I think it is because the station from which you are monitoring is not yet connected to any AP, so that it can measure the noise floor on the channel it is connected.
I have seen in real life links, on a very wireless crowded area, the noise floor dropping at -75 (very bad), actually jumping all the time from -75 to -85. But that was some years ago, I don't do wireless anymore that much.

If you turn both connected ends off does the monitor unit still report -96? Also, is the test radio on the same frequency as the connected pair? My best guess based on standard practice in microwave is that MT measures the NF at the frequency of operation vs measuring it across the entire band which would be almost useless to the field engineer. The standard practice use of noise floor measurement is to determine if the environment at the Frequency of Operation (Fo) that the receiver is operating in is "quiet enough" to support the anticipated performance of the receiver.