Which could still be implemented as a single sensor per key + interpolation. I which case it really depends how good the software is in edge cases. Like two fin

Which could still be implemented as a single sensor per key + interpolation. I which case it really depends how good the software is in edge cases. Like two fingers close together in untypical location. Noting enough machine learning cannot solve :)