We discuss a basic human-machine interface for transmitting human intention to a machine via impulsive sound, using wavelet analysis to determine whether impulsive sound is recognized. We found that tapping transmits indicates human intention to a machine, if we choose an appropriate analysis wavelet and its scaling parameter. We found that Daubechies wavelet function parameters were feasible for our purpose at N=3 (order) and j=8 (scaling parameter). This interface should prove useful for orders in regular work and emergency stoppage.