This paper investigates methods aiming at the automatic recognition and classification of discrete environmental sounds, for the purpose of subsequently applying these methods to the recognition of soundscapes. Research in audio recognition has traditionally focused on the domains of speech and music. Comparatively little research has been done towards recognizing non-speech environmental sounds. For this reason, in this paper, we apply existing techniques that have been proved efficient in the other two domains. These techniques are comprehensively compared to determine the most appropriate one for addressing the problem of environmental sound recognition.

Temporal feature integration refers to a set of strategies attempting to capture the information conveyed in the temporal evolution of the signal. It has been extensively applied in the context of semantic audio showing performance improvements against the standard frame-based audio classification methods. This paper investigates the potential of an enhanced temporal feature integration method to classify environmental sounds. The proposed method utilizes newly introduced integration functions that capture the texture window shape in combination with standard functions like mean and standard deviation in a classification scheme of 10 environmental sound classes. The results obtained from three classification algorithms exhibit an increase in recognition accuracy against a standard temporal integration with simple statistics, which reveals the discriminative ability of the new metrics.

Beamforming using a circular array of hydrophones may be employed for the task of two-dimensional (2D) underwater sound-field visualisation. In this article, a parametric spatial post-filtering method is proposed, which is specifically intended for applications involving large circular arrays and aims to improve the spatial selectivity of traditional beamformers. In essence, the proposed method is a reformulation of the cross-pattern coherence (CroPaC) spatial post-filter, which involves calculating the normalised cross-spectral density between two signals originating from coincident beamformers. The resulting parameter may be used to sharpen another beamformer steered in the same look-direction, while attenuating ambient noise and interferers from other directions. However, while the original 2D version of the algorithm has been demonstrated to work well with second-order circular harmonic input, it becomes increasingly less suitable with increasing input order. Therefore, the proposed reformulation extends the applicability of CroPaC for much higher orders of circular harmonic input. The method is evaluated with simulated data of a 96-channel circular hydrophone array in three different passive sonar scenarios, where the proposed post-filter is shown to improve the spatial selectivity of both delay-and-sum and minimum-variance distortionless response beamformers.

This letter presents a spatial post-filter that can be employed in linear hydrophone arrays, commonly found in sonar systems, for the task of improving the bearing estimation and noise suppression capabilities of traditional beamformers. The proposed filter is computed in the time-frequency domain as the normalised cross-spectral density between two beamformed signals, which are generated by applying conventional beamforming to two adjacent non-overlapping sub-arrays. The evaluation on both simulated and real-world data demonstrates promising performance compared to other popular post-filters in some cases, especially for targets near the end-fire direction and in the presence of uncorrelated interferers or diffuse noise.

This paper presents an underwater soundfield visualisation method for passive-sonar applications employing circular hydrophone arrays. The method operates by segregating the space by means of beamforming into angular sectors scanning the whole horizontal plane and then computing acoustic parameters within each sector. The information from these directionally constrained parameters is fused in order to produce spatial spectra which depict the distribution of acoustic energy over bearing. The evaluation is performed on simulated data of circular hydrophone arrays mounted on rigid cylindrical baffles. Comparisons against baseline methods of similar computational complexity suggest that, for moderate to high signal-to-noise ratio levels, the proposed method offers improved performance in terms of background noise suppression, angular resolution, and direction-of-arrival estimation accuracy. Additionally, it is demonstrated that, with the appropriate choice of sector pattern, the proposed method can, at least in some cases, achieve superior performance to the baseline methods in the presence of interferers even at low signal-to-interference ratio levels. Last, the sector-based parameter diffuseness, which is directly related to the direct-to-diffuse ratio, may be used both as a weight function to further attenuate the background noise level and as a confidence measure of the estimation accuracy.