The V3 doesn't use the three frequencies as another layer of discrimination. It uses one of two modes to process the three freqs: best data or correlate. In best data, the best signal of the three is used to display the VDI. In correlate, it uses all three in a manner that I haven't studied in detail. I haven't used correlate extensively as it is clear that depth is lost: on deep coins, the 22.5 kHz signal is faint or nonexistent and correlate seems to reject targets in this case. Correlate does, however, work rather well in iron fields where most deep targets are masked anyway.
In general, it's up to the operator to view the analyze screen and note the relative amplitudes of the three signals to determine the likely target.
I mentioned the idea of using the amplitudes and non-normalized VDI from the three frequencies as another layer of discrimination to Carl a few weeks ago and I got the impression that White's is already working on it and may employ three frequency type discrimination in a future model.
Oddly enough, despite owning the most advanced detector on the market, I find myself using the VDI display and analyze less and less as I get more adept at using the detector. Tone ID, and the sound of the audio tell me most everything I need to know. I mainly use analyze at this point to discriminate between bottlecaps and coins when using DD coils.
I will say this: a Dankowski style test garden with deep objects, close trash, and iron masking thrown in goes a long way in getting the most from the V3. As we well know, there are a multitude of settings and adjustments; experimentation is the rule of the day to finding what works and what doesn't.