Thanks, I'm glad it was helpful!
The version of the calibration screen shown in Devlog #10 is an old one, for what it's worth (see https://rhythmquestgame.com/devlog/41.html)
For your question about measuring a player's taps, there is unfortunately no best answer :o) There are arguments for both; I think my current approach does a sort of hybrid where it attempts to use each of the players taps as a "base" and then see which one of their taps results in the lowest average discrepancy.
You should definitely experiment to see what works best! You can also do things like take the average, but filter out any values that don't lie within a certain expected range of inputs. You can visually show those inputs as being discarded, as well.