Calibration method in CorrectADCResponse.ino seems inadequate

The calibration method and sketch [CorrectADCResponse.ino](https://github.com/arduino/ArduinoCore-samd/blob/master/libraries/SAMD_AnalogCorrection/examples/CorrectADCResponse/CorrectADCResponse.ino) did not do the job for me on a Seeeduino Xiao - so I started looking at how it works.

The problem with the method used is that it measures only the end points (0.0 and 3.3V) of the ADC input range, and those two measurements will often not be representative for a line through a number of points inside the range. It would be far better to have measurement points away from the range end points.

I realize that this is is the best that can be done with a simple two-point calibration, and that more elaborate methods must be used to get a better result.

I propose solving this by renaming this sketch something like "CorrectADCResponse_simple.ino", and then add a second sketch ("CorrectADCResponse_elaborate.ino" ?) with a more elaborate method - which could be implementing voltage dividers connected to multiple ADC inputs.

Resistors would have to be measured to take their production tolerance into account, and measured values inserted in the code.

Start by reading libraries/SAMD_AnalogCorrection/examples/CorrectADCResponse/CorrectADCResponse.ino and tracing its two-point calibration method. Define the scope of the proposed simple-sketch rename and the additional multi-point calibration sketch; the work is done when both sketches and the resistor-tolerance requirements are clearly implemented and documented.

Written by the indexing model from the issue text.