[ADTech]

I don't know what you fed into your calculations, but almost 22 dBi of gain is implausible. You'll need to either model it in NEC (or similar) or use a calibrated reference antenna to make comparative readings against in order to determine gain.

I wouldn't call an antenna with a 20 dB difference in "gain" at different azimuths to be omnidirectional although with your scaling that was selected, it would appear to be so. If you adjust the scaling from 10 db/div to 3 or 5 dB/div, the directionality will be more pronounced. Additionally, since the antenna under test isn't highly directional, the readings from a "live" TV signal may also included components of reflected signals which can add or subtract from the field strength at any given azimuth. Also, when measuring the power of a single frequency within the 6 MHz bandwidth of the ATSC signal, you're going to get variations in the power readings due to the effects of signal multi-path and the bandwidth of your power measurement tool. That's one of the vagaries of testing in a location that is not tightly controlled and it's exacerbated by using a live, OTA signal that may or may not be of the signal power you believe it to be.

As for the CEA classifications, my opinion is irrelevant as I don't have any input to their decision making and my employer is not a participant in their rating scheme. I'm just an observer.