Quantifying noise in animal acoustic signals.

Abstract

Variability in noisy calls or call segments is often difficult to quantify but critical in order to understand animal acoustic communication. Such components represent a large proportion of animal acoustic signals and they are behaviorally meaningful. Such elements have also collectively been labeled as nonlinear phenomena. Three approaches are reviewed. The harmonics‐to‐noise ratio (HNR) can be related to perceptual characteristics of an acoustic signal as well as to the mechanism of its production. The HNR seems to be a useful numeric parameter to quantify vocal variability. Two problems with the procedure will be outlined. The nonlinear measure (NLM) is designed to provide an overall estimate of the strength of nonlinearity in a signal. The rationale is that if signal noise reflects low‐dimensional chaos as produced through nonlinear processes, deterministic‐nonlinear modeling will produce a relatively small error component. A stochastic linear model, on the other hand, will produce a relatively large ...