My turn....
We debated this all the time in the old days. I didn't watch the video in this thread but others I've seen are pretty much BS.
Yes, if you have a single frequency detector there will be a difference in response to different targets if you change the operating frequency from 1.5khz to 100khz.
But, that is not how it works in the FBS world. They transmit square wave pulses of two frequencies, and use where they mate up to generate the 17 (methinks) frequencies spread from roughly 1.5khz to 100khz. When you do a noise cancel or change the frequency number you are changing the frequency (and maybe period) of one of the square waves by a small amount. Now maybe they may mate up from 1.6khz to 101khz going from channel 1 to 2. All numbers I'm using are pretty much made up, I don't know the exact values. Anyhow even going to channel 11 is only going to shift the frequency slightly, say perhaps 2khz at the bottom end. Going from 1.5k to 2k is not going to have much if any affect on target response.
They reason for noise cancel and channel selection has nothing to do with ground mineralization or optimizing target response. Only to try get away from EMI; in the US and Canada primarily harmonics of 60hz, the rest of the world 50hz.
Long ago I read many of the patents and if I remember correctly the FBS detectors don't even bother to process frequencies that land close 50 or 60 hz harmonics. Generally I've only been able to bump sensitivity one or two notches after a noise cancel, but occasionally you may encounter an EMI source that is not related to the power grid; say a electric dog fence of someone else's detector, then it can make a considerable difference.
Chris