2aAB12. Removal of environmental effects from whale calls using coherent beamforming methods.

Session: Tuesday Morning, December 3

Time: 11:10

Abstract:

Coherent beamforming with point constraints (MPC) and matched-field processing (MFP) techniques are used to remove multipath reflections from whale vocalizations recorded on a 64-element 120-m vertical aperture array suspended from R/P FLIP. From a rough estimate of the whales's location, the beamformer response required to cancel the surface and bottom reflections is constructed. The CMPC algorithm then derives the appropriate array element weightings. The beamformed output at each frequency is summed and inverse-transformed to reconstruct the original whale call. MFP is an alternate method for removing reflections from sounds generated within well-defined acoustic environments. MFP yields an estimate for the range, azimuth, and depth of a calling whale, which in turn allows subsequent calculation of the amplitude and phase shifts experienced by the call's frequency components during propagation. The transfer function of the channel is thereby constructed and removed from the received call spectrum, producing another estimate of the whale source signal. The recovered signal spectrum are compared to spectral estimates obtained by incoherent averaging across phones and single-phone cepstral analysis to examine their utility for situations where coherent multiphone data are unavailable. [Work supported by ONR.]