The present work highlights how, in order to characterize the motion of a variable mass pendulum,
wavelet analysis can be an effective tool in furnishing information on the time evolution of the
oscillation spectral content. In particular, the wavelet transform is applied to process the motion
of a hung funnel that loses fine sand at an exponential rate; it is shown how, in contrast to the
Fourier transform which furnishes only an average frequency value for the motion, the wavelet
approach makes it possible to perform a joint time–frequency analysis. The work is addressed at
undergraduate and graduate students.