The study of birdsong has implications in a num-ber of biological and conservational applications. However, theanalysis of bird vocalisations in the natural habitat is still largelya laborious task. One of the bottle necks is the segmentation ofbird vocalisations into individual syllables. Simple segmentationin time domain proves difficult because of overlapping signalsover different frequency bands. The common approach is toconvert audio recordings into a spectrogram and apply imageprocessing techniques to pick out the signal of interest. Weexamine several methods that have been proposed recently to dojust this and find that they are inadequate to deal with harmonicvocalisations. We propose a method that segments syllables bylooking for the fundamental frequency first then works its wayup in the frequency axis to find other harmonics if they exist. Weevaluate our method against another popular method and findthat the proposed method can segment correctly more than 70%the number of syllables, more than twice that of the method weare comparing to

If you like this work, please consider sharing with , , , , or leave me a comment below.