Impulsive-source active sonar systems are often plagued by false alarm echoes resulting from the presence of naturally occurring clutter objects in the environment. Sonar performance could be improved by a technique for discriminating between echoes from true targets and echoes from clutter. Motivated by anecdotal evidence that target echoes sound very different than clutter echoes when auditioned by a human operator, this paper describes the implementation of an automatic classifier for impulsive-source active sonar echoes that is based on perceptual signal features that have been previously identified in the musical acoustics literature as underlying timbre. Perceptual signal features found in this paper to be particularly useful to the problem of active sonar classification include: the centroid and peak value of the perceptual loudness function, as well as several features based on subband attack and decay times. This paper uses subsets of these perceptual signal features to train and test an automatic classifier capable of discriminating between target and clutter echoes with an equal error rate of roughly 10%; the area under the receiver operating characteristic curve corresponding to this classifier is found to be 0.975.

Human listening tests were conducted to investigate if participants could distinguish between samples of target echoes and clutter obtained from a broadband active sonar experiment. For each echo, the listeners assigned a rating based on how confident they were that it was a target echo or clutter. The measure of performance was the area under the binormal receiver-operating-characteristic (ROC) curve, A(z). The mean performance was A(z)=0.95 ± 0.04 when signals were presented with their full available acoustic bandwidth of approximately 0-2 kHz. It was A(z)=0.77 ± 0.08 when the bandwidth was reduced to 0.5-2 kHz. The error bounds are stated as 95% confidence intervals. These results show that the listeners could definitely hear differences, but their performance was significantly degraded when the low-frequency signal information was removed. The performance of an automatic aural classifier was compared against this human-performance baseline. Results of statistical tests showed that it outperformed 2 of 13 listeners and 5 of 9 human listeners in the full-bandwidth and reduced-bandwidth tests, respectively, and performed similarly to the other listeners. Given its performance, the automatic aural classifier may prove beneficial to Navy sonar systems.

Superdirective line arrays can provide a significant array gain from a structure that is relatively small in terms of acoustic wavelengths. However, system imperfections, electronic noise and acoustic scatter from the array structure can degrade their performance. An acoustic calibration of a six-element line array, 0.8 m in length, has been performed over the frequency range 1 to 4 kHz in order to investigate the performance of a real array. The data is used to identify the angular variation of the hydrophone outputs and the phase difference between hydrophone pairs. These angular responses are analyzed in terms of a modal series in order to quantify the variations and help identify the source of perturbations. The effects of imperfections are also investigated by synthesising superdirective arrays of order 1 to 5 and monitoring how the array gain varies for both deterministic signals and ambient acoustic noise. These results are compared with theoretical predictions. Further evidence of the variation in performance is gained by comparing the output of different implementations of lower order arrays, synthesized from subsets of the full array. The results indicate the influences that the array structure may have on the performance of the array.

In musical acoustics significant effort has been devoted to uncovering the physical basis of timbre perception. Most investigations into timbre rely on multidimensional scaling (MDS), in which different musical sounds are arranged as points in multidimensional space. The Euclidean distance between points corresponds to the perceptual distance between sounds and the multidimensional axes are linked to measurable properties of the sounds. MDS has identified numerous temporal and spectral features believed to be important to timbre perception. There is reason to believe that some of these features may have wider application in the disparate field of underwater acoustics, since anecdotal evidence suggests active sonar returns from metallic objects sound different than natural clutter returns when auralized by human operators. This is particularly encouraging since attempts to develop robust automatic classifiers capable of target-clutter discrimination over a wide range of operational conditions have met with limited success. Spectral features relevant to target-clutter discrimination are believed to include click-pitch and envelope irregularity; relevant temporal features are believed to include duration, sub-band attack/decay time, and time separation pitch. Preliminary results from an investigation into the role of these timbre features in target-clutter discrimination will be presented. [Work supported by NSERC and GDC.]

Unacceptably high false-alarm rates due to the inability to discriminate between target echoes and environmental clutter are an issue for existing low-frequency active sonar systems operating in coastal environments. A research project at Defence R&D Canada—Atlantic is investigating the potential use of aural cues to tackle this challenge. One aspect of the project is to evaluate the human ability to aurally discriminate between target echoes and environmental clutter. The design and preliminary results from the study are presented here. Human subjects are presented with a series of sounds containing target echoes and clutter obtained from recordings of an incoherent broadband sonar experiment. The quantitative data collected in the study are the subjects’ decisions as to whether the echo heard was a target echo or clutter and their level of confidence associated with the decisions. Receiver-operating characteristic (ROC) analysis is used to produce a statistical model of the subjects’ performance. The study also includes a questionnaire: answers may prove useful in supporting the quantitative results and in providing a better understanding of the cues and decision techniques used by the subjects.