We typically think of seismic waves as continuous functions of space and time. Yet, we record, process, and interpret these waves as series of numbers (“samples”) of the underlying wavefield. The question of how these samples should be chosen to minimize loss of information lies at the heart of digital signal processing. The cornerstone in this field is the sampling theorem. It guarantees that a certain class of continuous functions (so-called “band-limited”) can be perfectly recovered from sampled values if certain sampling criteria are obeyed.