The voices of my children keep changing all the time. Does the system have to be taught continuously in this case?

It is correct that our voice features change considerably during a time span of 6 months already. This has been taken into account with the VOCAL SCW 1 in order to ensure comfortable use over long periods. The system is trainable, e.g. following certain criteria the reference sample in the memory is matched up to the current data and updated if need be.
Thus, if in regular use, the VOCAL SCW 1 is self-updating.