Visual Fingerprints of the Acoustic Environment: The Use of Acoustic Indices to Characterise Natural Habitats

摘要: Acoustic recordings play an increasingly important role in monitoring terrestrial environments. However, due to rapid advances technology, ecologists are accumulating more audio than they can listen to. Our approach this big-data challenge is visualize the content of long-duration audio-recordings by calculating acoustic indices. These statistics which describe temporal-spectral distribution energy and reflect ecological interest. We combine spectral indices produce false-color spectrogram images. not only reveal but also facilitate navigation. An additional analytic find appropriate descriptors summarize 24-hour recordings, so that it becomes possible monitor long-term changes environment at a single location compare environments different locations. 'acoustic-fingerprint' shows some preliminary promise.