Hunter, if you're going to try the experiment thing with noise channel vs. the best coin detection, i'd recommend burying the coins instead of air testing. Its important the test bed is located in an area that has no EMI interference which would skew the test results for some channels likely rendering the whole exercise useless.
As i indicated earlier, some swear certain channels are deeper seeking for specific targets and/or soil but i haven't seen that albeit never actually tested for it.
I would think you'll find little to no difference as i don't know this for a fact but i suspect the xmit frequency bandwidth doesn't shift enough to make a gigantic difference in depth or specific target ID.
The most important thing to consider with noise channels, is to use a channel that is the quietest especially in a high EMI area. I typically AUTO noise cancel in hi EMI areas and call it a day. The quietest channel in those areas will give the best depth and IDs regardless of target or soil conditions.
The reason for raising the coil during AUTO noise cancel utilizes the fact that for a lack of better words, the coil can act as a defacto 'radio' antenna to better receive and process any EMI radiation in the area to be detected.
Ground canceling on the other hand is a function of FBS technology. While i can half way see how changing xmit frequency bandwidth could affect ground canceling for some soil conditions i think the effect would be minimal, barely or not perceptable to the user. Again i never experimented with this, just guessing.
If you go ahead with some testing we'd all appreciate posting the test description and results. Who knows, maybe some of us will be surprised!
IMO, the only time you can manually pick and choose a noise channel is in a very low to no EMI area then have at it sugar pill or not.