Discrimination of frequency-modulated Baleen whale downsweep calls with overlapping frequencies

Automatic classification of fin, sei, and blue whale frequency modulated downsweeps has been a challenging task for bioacousticians. These calls overlap in frequency range and have similar time durations. The traditional spectrogram methodology, the Short Time Fourier Transform, tends to be ineffective because of the large temporal ambiguities needed to achieve the necessary frequency resolution to study the fine time-frequency (TF) structures. Spectrograms generated with the Pseudo Wigner-Ville Distribution (PWVD) provide much higher simultaneous TF resolution compared with the traditional method. The PWVD allows bioacousticians to study the fine TF structures of the sound, such as the instantaneous frequency, instantaneous bandwidth, contour slope, etc. These features set the foundation of identifying sounds that are usually considered difficult to discriminate using the traditional method. Wigner-Ville distribution of the baleen whale downsweeps showed distinguishable characteristics; for example, the TF contour of fin and sei whales exhibited concave and convex shapes, which have never been reported in the literature. A Support Vector Machine classifier was trained and tested based on the parameters extracted from the PWVD.