IMFs and RF energy distribution characteristic analysis of blue whale vocalizations

Abstract

This paper presents the intrinsic mode functions (IMFs) and residue function (RF) energy distribution characteristic analysis technology for blue whale vocalizations. Blue whale vocalization were downloaded from the Scripps Whale Acoustics Lab [7] of the University of California San Diego website; blue whale type B-calls with six blue whale vocalization samples were analyzed. Four IMFs and one RF were disassembled from the samples using the empirical mode decomposition (EMD) method. The average energy distribution ratio characteristics of IMFs and RF were presented, and the Hilbert spectrum (HS) characteristics with time-frequency-energy distribution were obtained. The analysis results show that the average energy distribution ratios of the blue whale vocalizations of IMF1, IMF2, IMF3, and IMF4, were 3.58% in the 250-310Hz frequency band during the intervals of 3.5–3.53ms, 2.69% in the 60-100Hz frequency band during the intervals of 2.54–2.57ms, 4.30% in the 40-50Hz frequency band during the intervals of 3.14–3.17ms, and 4.04% in the 20-30Hz frequency band during the intervals of 3.05–3.08ms, respectively. This study helps us understand the Hilbert spectrum (HS) features of blue whale vocalizations.