Source separation in ecoacoustics: a roadmap towards versatile soundscape information retrieval

Abstract

A comprehensive assessment of ecosystem dynamics requires the monitoring of biological, physical and social changes. Changes that cannot be observed visually may be trackable acoustically through soundscape analysis. Soundscapes vary greatly depending on geophysical events, biodiversity and human activities. However, retrieving source-specific information from geophony, biophony and anthropophony remains a challenging task, due to interference by simultaneous sound sources. Audio source separation is a technique that aims to recover individual sound sources when only mixtures are accessible. Here, we review techniques of monoaural audio source separation with the fundamental theories and assumptions behind them. Depending on the availability of prior information about the source signals, the task can be approached as a blind source separation or a model-based source separation. Most blind source separation techniques depend on assumptions about the behaviour of the source signals, and their performance may deteriorate when the assumptions fail. Model-based techniques generally do not require specific assumptions, and the models are directly learned from labelled data. With the recent advances of deep learning, the model-based techniques can yield state-of-the-art separation performance, accordingly facilitate content-based audio information retrieval. Source separation techniques have been adopted in several ecoacoustic applications to evaluate the contributions from biodiversity and anthropogenic disturbance to soundscape dynamics. They can also be employed as nonlinear filters to improve the recognition of bioacoustic signals. To effectively retrieve ecological information from soundscapes, source separation is a crucial tool. We believe that the future integrations of ecological hypotheses and deep learning can realize a high-performance source separation for ecoacoustics, and accordingly improve soundscape-based ecosystem monitoring. Therefore, we outline a roadmap for applying source separation to assist in soundscape information retrieval and hope to promote cross-disciplinary collaboration. (biophony), their habitats (geophony) and human development (anthropophony) (Dumyahn and Pijanowski 2011; Pijanowski et al. 2011; Sueur and Farina 2015; Krause and Farina 2016). By listening to biotic sounds, the composition of soniferous animals can be characterized and applied to assess wildlife biodiversity (Andr e et al. 2011; Saito et al. 2015; Lin et al. 2017a; Ross et al. 2018). On the other hand, abiotic sounds can be used to evaluate the ecosystem dynamics associated with

Sensing ecosystem dynamics via audio source separation: A case study of marine soundscapes off northeastern Taiwan

“…Audio source separation aims to reconstruct independent sound sources from a mixture. Therefore, the separation result not only reduces the interference among sound sources but also facilitates further applications, such as variability assessment of source behaviors [ 23 ]. State-of-the-art speech separation models based on deep learning were trained to map noisy recordings to clean speech.…”

“…Investigating sounds associated with geophysical events and human activities can also be used to evaluate the impacts of anthropogenic activities such as fishing and shipping on marine ecosystems [ 19 – 22 ]. The increase of environmental and anthropogenic sounds represents an indicator of habitat change, but it interferes with acoustic analysis by masking the received biological sounds [ 23 ]. To date, the data and metrics generated from soundscapes have not been effectively converted into tools that can be used by managers and stakeholders.…”

“…The issue of simultaneous source interference may be addressed by applying blind source separation (BSS). A BSS model aims to separate a set of sound sources from a mixture, without prior information specific to individual sources or the mixing process [ 23 ]. Unlike clustering, effective BSS can reconstruct individual sound sources and add value to noisy recordings.…”

“…With the assumption of source-specific periodicity, biological chorus and noise can be effectively separated in an unsupervised manner [ 11 ]. The PC-NMF can also separate sound sources with varied patterns of periodical occurrence, revealing a more delicate temporal pattern of acoustic diversity [ 23 ].…”

Ecoacoustics: acoustic sensing for biodiversity monitoring at scale

“…In either case, there will usually be multiple animals audible on any given audio track. Lin and Tsao (2019) provide a review and roadmap of source‐separation methods including recent techniques that may help to disentangle overlapping sounds in monophonic recordings. Sumitani et al (2020) demonstrate that interaction patterns among vocalizing individuals can be characterized with the aid of a dimension reduction algorithm coupled to a new compact microphone array, leading to automatic source localization.…”