Several pieces of different musical kinds were studied measuring A(t), the
output amplitude of a peak detector driven by the electric signal arriving
to the loudspeaker. Fixed a suitable threshold (A) over bar, we considered
N(A), the number of times that A(t) > (A) over bar, each of them we named e
vent and N(t), the distribution of times t between two consecutive events.
Some N(A) and N(t) distributions are displayed in the reported logarithmic
plots, showing that jazz, pop, rock and other popular rhythms have noise-di
stribution, while classical pieces of music are characterized by more compl
ex statistics. We pointed out the extraordinary case of the aria "La calunn
ia e un venticello", where the words describe an avalanche or seismic proce
ss, calumny, and the rossinian music shows N(A) and N(t) distribution typic
al of earthquakes.