sounding_out_chorus

<p>This repository contains the data for the paper <strong>Towards interpretable learned representations for Ecoacoustics using variational auto-encoding</strong>.<strong> </strong>This dataset contains a series of 1 min wav files recorded across UK and Ecuadorian habitats. Each sample has 26 acoustic indices calculated and a full list of avian species and abundances. This dataset is an updated version of a <a href="https://zenodo.org/record/1255218">previous release (10.5281/zenodo.1255218)</a> including KML maps containing GPS data for each sample site and updated label metadata. A data module for use in a PyTorch machine learning pipeline is <a href="https://gitlab.com/ecolistening/sounding_out_torch">available here</a>.</p> <p><strong>Abstract</strong><br> Ecoacoustics is an emerging science that seeks to understand the role of sound in ecological processes.<br> Passive acoustic monitoring is increasingly being used to collect vast quantities of whole-soundscape<br> audio recordings in order to study variations in acoustic community activity across spatial and<br> temporal scales. However, extracting relevant information from audio recordings for ecological<br> inference is non-trivial. Recent approaches to machine-learned acoustic features appear promising<br> but are limited by inductive biases, crude temporal integration methods and few means to interpret<br> downstream inference. To address these limitations we developed and trained a self-supervised<br> representation learning algorithm - a convolutional Variational Auto-Encoder (VAE) - to embed<br> latent features from acoustic survey data collected from sites representing a gradient of habitat<br> degradation in temperate and tropical ecozones and use prediction of survey site as a test case for<br> interpreting inference. We investigate approaches to interpretability by mapping discriminative<br> descriptors back to the spectro-temporal domain to observe how soundscape components change<br> as we interpolate across a linear classification boundary traversing latent feature space; we advance<br> temporal integration methods by encoding a probabilistic soundscape descriptor capable of capturing<br> multi-modal distributions of latent features over time. Our results suggest that varying combinations<br> of soundscape components (biophony, geophony and anthrophony) are used to infer sites along a<br> degradation gradient and increased sensitivity to periodic signals improves on previous research using<br> time-averaged representations for site classification. We also find the VAE is highly sensitive to<br> differences in recorder hardware&rsquo;s frequency response and demonstrate a simple linear transformation<br> to mitigate the effect of hardware variance on the learned representation. Our work paves the way for<br> development of a new class of deep neural networks that afford more interpretable machine-learned<br> ecoacoustic representations to advance the fundamental and applied science and support global<br> conservation efforts.<br> <br> <strong>Sampling Methods (extract from paper)</strong><br> Surveys were designed to monitor the acoustic characteristics of sites across a gradient of degradation, ranging from primary forest, through secondary forest (or areas in the process of ecological restoration), to agricultural monocultures, providing a space-for-time substitution to investigate changes in soundscapes across a gradient of ecological status. Samples were taken for 1 minute in every 15 for 10 sequential days at each site. Full dawn and dusk recordings were also collected. In each site, 15 recorders were placed in a grid-like system spaced a minimum of 200m away from their neighbours in the UK - 300m in Ecuador - to mitigate acoustic overlap and avoid spatial pseudo-replication. Wildlife Acoustics Song Meters equipped with two channel omni-directional microphone were used. Seven SM2+ and eight SM3 devices were deployed. Gains were matched between recorders (analogue gains at +36dB on SM2+ and +12dB on SM3 which has inbuilt +12dB gain) and recordings made at resolution of 16 bits with a sampling rate of 48 kHz. To provide a cleaner validation data set, local weather recordings were used to select 3 days with lowest wind and rain from each site giving 4725 1 min recording in total. Sites are labelled by their quality in descending order i.e. UK1 (primary), UK2 (regenerating), UK3 (degraded).</p>