Characterization and analysis of marine mammal sounds using time-frequency and time-prony techniques

This paper presents the application of time-frequency signal processing techniques to the analysis of marine mammal sounds. In general, the analysis of marine mammal sounds are compared to a baseline in order to assess the animal´s reaction to a stimulus such as man made sound sources in the ocean environment. A particular area of interest is the effect that low frequency sound has on marine mammals because of the global warming experiments that are slated to use high energy, low frequency sound as a means of assessing the gradual changes in the ocean temperatures. Until recently, the impacts of such sound on marine mammals has not been studied. However, as modern civilization makes more demands on the environments in which we live, especially the ocean environment, it is important that we understand the impacts on the resources that exist in them. Marine mammals are a top priority resource that are receiving considerable attention from communities around the world. This has prompted the United States Government to sponsor research that will assess the effects of noise, especially low frequency sounds, on marine mammals in their natural environment. Important area of this research is the analysis of sounds that marine mammals emit in response to a stimulus such as a low frequency sound source. Accurate analysis, however requires the use of advanced signal processing techniques to do this as well as experimentation performed in an ocean environment under controlled and calibrated conditions. This paper presents an approach to achieve both of these goals