Source characterization using recordings made in a reverberant underwater channel

The ability to accurately characterize an underwater sound source is an important prerequisite for many applications including detection, classification, monitoring and mitigation. Unfortunately, anechoic underwater recording environments required to make ideal recordings are generally not available. This paper presents a practical approach to source characterization when working in an imperfect recording environment; the source spectrum is obtained by equalizing the recording with the inverse of the chan-nel's impulse response (IR). An experiment was conducted in a diving well (depth of 5.18 m) using a logarithmic chirp to obtain the IR. IR length is estimated using methods borrowed from room acoustics and inversion of non-minimum phase IR is accomplished separately in the time and frequency domain to allow for a direct comparison. Results indicate that the energy of controlled sources can be recovered with root-mean-square error of À70 dB (10–70 kHz band). Two equations, one coherent and the other incoherent, are presented to calculate source spectral levels of an unknown source in a reverberant environment. This paper introduces a practical procedure outlining steps to obtain an anechoic estimate of an unknown source using equipment generally available in an acoustic laboratory. Underwater source characterization is important for numerous applications. For example, passive acoustic detection and classification can be improved by knowledge of the sound characteristics of the object of interest. With knowledge of the source, array configuration and specifications can be optimized for monitoring. As another example, environmental compliance laws regulate an environment by putting limits on emitted acoustic energy, so that a sound source needs to be well understood before being used in the environment. Unfortunately, anechoic underwater recording environments required to make ideal recordings are generally not available or are cost-prohibitive. An anechoic recording contains the direct arrival of acoustic energy from a source to the hydrophone with minimal noise or wall reflections. Sound levels estimated from recordings made in a reverberant environment (such as a test tank or pool) generally overestimate source levels due to additional wall reflections and noise. It was found [4,9] that the acoustic power of a source can be separated from reverberant energies by measuring the spectral pressure at one or more random locations in a reverberant enclosure (yielding spatial mean spectral levels). Recordings must be conducted in the far field of the source, e.g., the hydrophone is placed within the homogeneous and isotropic reverberant field. An estimate of the source is obtained by adjusting …