This article compares methods for the conversion of time-series into a spectro-temporal representation. These methods are designed based on a resemblance with the auditory processing of sound in the inner ear, or on mathematical principles related to, for example, Fourier analysis. This study provides a comparison between several of these methods. Two tests were devised for this comparison: one based on susceptibility to noise and one on the expression of spectro-temporal detail. While some methods produced good results on one of the two tests, others produced good results on both. Overall the transmission line model using an impedance function suggested by Zweig [1] provided the best results, though not significantly. Also a larger computational load may hinder application in some domains. The gammatone filterbank and straightforward spectrogram provide good alternatives with less computational load.

This article compares methods for the conversion of timeseries into a spectro-temporal representation. These methods are designed based on a resemblance with the auditory processing of sound in the mammalian inner ear, or on mathematical principles related to, for example, Fourier analysis. This study provides a comparison between several of these methods. Two tests were devised for this comparison: one based on susceptibility to noise and one on the expression of spectro-temporal detail. These two aspects were considered of importance for real world applications. While some methods produced good results on one of the two tests, others produced good results on both. Overall the transmission line model using an impedance function suggested by Zweig (“Finding the impedance of the organ of Corti,” J. Acoust. Soc. Amer., vol. 89, no. 3, pp. 1229-1254, 1991) provided the best results, though not significantly. Also a larger computational load may hinder application in some domains. The gammatone filterbank and straightforward spectrogram provide good alternatives with less computational load. The introduction of nonlinearity was shown to deteriorate performance on both tests, in both the filterbank and in the transmission line model.

The paper presents a knowledge-based system designed to detect evidence of aggression by means of audio analysis. The detection is based on the way sounds are analyzed and how they attract attention in the human auditory system. The performance achieved is comparable to human performance in complex social environments. The SIgard system has been deployed in a number of different real-life situations and was tested extensively in the inner city of Groningen. Experienced police observers have annotated ~1400 recordings with various degrees of shouting, which were used for optimization. All essential events and a small number of nonessential aggressive events were detected. The system produces only a few false alarms (non-shouts) per microphone per year and misses no incidents. This makes it the first successful detection system for a non-trivial target in an unconstrained environment.