Chirp is a set of related tools for analyzing and comparing
bioacoustic recordings.  The central component is a pitch-detection
algorithm that will track the pitch of a signal as a function of time.
A number of features enable the algorithm to make stable estimates
even from very noisy signals.  Performance is greatly improved by
masking out noise in the spectrotemporal domain.  A graphical user
interface (/chirp/) is provided for examining recordings and creating
masks.  The GUI can run batch pitch analyses and perform pairwise
comparisons using pitch and other methods for large sets of
recordings.  Command-line scripts are also provided for these batch
analyses.

For more information, visit the project website at
http://dmeliza.github.com/chirp/