Detecting bird sounds in a complex acoustic environment and application to bioacoustic monitoring

Ecology & computer audition: Applications of audio technology to monitor organisms and environment

Among the 17 Sustainable Development Goals (SDGs) proposed within the 2030 Agenda and adopted by all the United Nations member states, the 13th SDG is a call for action to combat climate change. Moreover, SDGs 14 and 15 claim the protection and conservation of life below water and life on land, respectively. In this work, we provide a literature-founded overview of application areas, in which computer audition - a powerful but in this context so far hardly considered technology, combining audio signal processing and machine intelligence - is employed to monitor our ecosystem with the potential to identify ecologically critical processes or states. We distinguish between applications related to organisms, such as species richness analysis and plant health monitoring, and applications related to the environment, such as melting ice monitoring or wildfire detection. This work positions computer audition in relation to alternative approaches by discussing methodological strengths and limitations, as well as ethical aspects. We conclude with an urgent call to action to the research community for a greater involvement of audio intelligence methodology in future ecosystem monitoring approaches.

Detection ranges of forest bird vocalisations: guidelines for passive acoustic monitoring

Passive acoustic monitoring has proven to have many advantages for monitoring efforts and research activities. However, there are considerations to be taken into account regarding the placement of autonomous sound recorders. Detection ranges differ among species and in response to variable conditions such as weather or the location of vocalising animals. It is thus important to the success of a research project to understand, with a certain degree of confidence, the distances at which birds might be detected. In two types of forests in Poland, we played back the vocalisations of 31 species of European forest birds exemplifying different singing characteristics. Based on recordings obtained along a 500-m transect, we estimated the probability of detection and maximum detection distance of each vocalisation. We broadcasted the recording at three heights of singing and repeated playbacks three times during the breeding season to evaluate the effect of vegetation growth. Our results revealed that environmental and meteorological factors had a significant influence on both detection probability and maximum detection distances. This work provides comprehensive measurements of detection distance for 31 bird species and can be used to plan passive acoustic monitoring research in Europe, taking into account species traits and individual characteristics of the study area.

Bioacoustic IoT Sensors as Next-Generation Tools for Monitoring: Counting Flying Insects through Buzz

The global loss of biodiversity is an urgent concern requiring the implementation of effective monitoring. Flying insects, such as pollinators, are vital for ecosystems, and establishing their population dynamics has become essential in conservation biology. Traditional monitoring methods are labour-intensive and show time constraints. In this work, we explore the use of bioacoustic sensors for monitoring flying insects. Data collected at four Italian farms using traditional monitoring methods, such as hand netting and pan traps, and bioacoustic sensors were compared. The results showed a positive correlation between the average number of buzzes per hour and insect abundance measured by traditional methods, primarily by pan traps. Intraday and long-term analysis performed on buzzes revealed temperature-related patterns of insect activity. Passive acoustic monitoring proved to be effective in estimating flying insect abundance, while further development of the algorithm is required to correctly identify insect taxa. Overall, innovative technologies, such as bioacoustic sensors, do not replace the expertise and data quality provided by professionals, but they offer unprecedented opportunities to ease insect monitoring to support conservation biodiversity efforts.

EnViTSA: Ensemble of Vision Transformer with SpecAugment for Acoustic Event Classification

Recent successes in deep learning have inspired researchers to apply deep neural networks to Acoustic Event Classification (AEC). While deep learning methods can train effective AEC models, they are susceptible to overfitting due to the models' high complexity. In this paper, we introduce EnViTSA, an innovative approach that tackles key challenges in AEC. EnViTSA combines an ensemble of Vision Transformers with SpecAugment, a novel data augmentation technique, to significantly enhance AEC performance. Raw acoustic signals are transformed into Log Mel-spectrograms using Short-Time Fourier Transform, resulting in a fixed-size spectrogram representation. To address data scarcity and overfitting issues, we employ SpecAugment to generate additional training samples through time masking and frequency masking. The core of EnViTSA resides in its ensemble of pre-trained Vision Transformers, harnessing the unique strengths of the Vision Transformer architecture. This ensemble approach not only reduces inductive biases but also effectively mitigates overfitting. In this study, we evaluate the EnViTSA method on three benchmark datasets: ESC-10, ESC-50, and UrbanSound8K. The experimental results underscore the efficacy of our approach, achieving impressive accuracy scores of 93.50%, 85.85%, and 83.20% on ESC-10, ESC-50, and UrbanSound8K, respectively. EnViTSA represents a substantial advancement in AEC, demonstrating the potential of Vision Transformers and SpecAugment in the acoustic domain.

A Novel Bird Sound Recognition Method Based on Multifeature Fusion and a Transformer Encoder

Birds play a vital role in the study of ecosystems and biodiversity. Accurate bird identification helps monitor biodiversity, understand the functions of ecosystems, and develop effective conservation strategies. However, previous bird sound recognition methods often relied on single features and overlooked the spatial information associated with these features, leading to low accuracy. Recognizing this gap, the present study proposed a bird sound recognition method that employs multiple convolutional neural-based networks and a transformer encoder to provide a reliable solution for identifying and classifying birds based on their unique sounds. We manually extracted various acoustic features as model inputs, and feature fusion was applied to obtain the final set of feature vectors. Feature fusion combines the deep features extracted by various networks, resulting in a more comprehensive feature set, thereby improving recognition accuracy. The multiple integrated acoustic features, such as mel frequency cepstral coefficients (MFCC), chroma features (Chroma) and Tonnetz features, were encoded by a transformer encoder. The transformer encoder effectively extracted the positional relationships between bird sound features, resulting in enhanced recognition accuracy. The experimental results demonstrated the exceptional performance of our method with an accuracy of 97.99%, a recall of 96.14%, an F1 score of 96.88% and a precision of 97.97% on the Birdsdata dataset. Furthermore, our method achieved an accuracy of 93.18%, a recall of 92.43%, an F1 score of 93.14% and a precision of 93.25% on the Cornell Bird Challenge 2020 (CBC) dataset.

Acoustic indices as proxies for bird species richness in an urban green space in Metro Manila

We assessed eight acoustic indices as proxies for bird species richness in the National Science Complex (NSC), University of the Philippines Diliman. The acoustic indices were the normalized Acoustic Complexity Index (nACI), Acoustic Diversity Index (ADI), inverse Acoustic Evenness Index (1-AEI), Bioacoustic Index (BI), Acoustic Entropy Index (H), Temporal Entropy Index (Ht), Spectral Entropy Index (Hf), and Acoustic Richness Index (AR). Low-cost, automated sound recorders using a Raspberry Pi were placed in three sites at the NSC to continuously collect 5-min sound samples from July 2020 to January 2022. We selected 840 5-min sound samples, equivalent to 70 hours, through stratified sampling and pre-processed them before conducting acoustic index analysis on the raw and pre-processed data. We measured Spearman's correlation between each acoustic index and bird species richness obtained from manual spectrogram scanning and listening to recordings. We compared the correlation coefficients between the raw and pre-processed.wav files to assess the robustness of the indices using Fisher's z-transformation. Additionally, we used GLMMs to determine how acoustic indices predict bird species richness based on season and time of day. The Spearman's rank correlation and GLMM analysis showed significant, weak negative correlations between the nACI, 1-AEI, Ht, and AR with bird species richness. The weak correlations suggest that the performance of acoustic indices are dependent on various factors, such as the local noise conditions, bird species composition, season, and time of day. Thus, ground-truthing of the acoustic indices should be done before applying them in studies. Among the eight indices, the nACI was the best-performing index, performing consistently across sites and independently of season and time of day. We highlight the importance of pre-processing sound data from urban settings and other noisy environments before acoustic index analysis, as this strengthens the correlation between index values and bird species richness.

A workflow for the automated detection and classification of female gibbon calls from long-term acoustic recordings

Passive acoustic monitoring (PAM) allows for the study of vocal animals on temporal and spatial scales difficult to achieve using only human observers. Recent improvements in recording technology, data storage, and battery capacity have led to increased use of PAM. One of the main obstacles in implementing wide-scale PAM programs is the lack of open-source programs that efficiently process terabytes of sound recordings and do not require large amounts of training data. Here we describe a workflow for detecting, classifying, and visualizing female Northern grey gibbon calls in Sabah, Malaysia. Our approach detects sound events using band-limited energy summation and does binary classification of these events (gibbon female or not) using machine learning algorithms (support vector machine and random forest). We then applied an unsupervised approach (affinity propagation clustering) to see if we could further differentiate between true and false positives or the number of gibbon females in our dataset. We used this workflow to address three questions: (1) does this automated approach provide reliable estimates of temporal patterns of gibbon calling activity; (2) can unsupervised approaches be applied as a post-processing step to improve the performance of the system; and (3) can unsupervised approaches be used to estimate how many female individuals (or clusters) there are in our study area? We found that performance plateaued with >160 clips of training data for each of our two classes. Using optimized settings, our automated approach achieved a satisfactory performance (F1 score ~ 80%). The unsupervised approach did not effectively differentiate between true and false positives or return clusters that appear to correspond to the number of females in our study area. Our results indicate that more work needs to be done before unsupervised approaches can be reliably used to estimate the number of individual animals occupying an area from PAM data. Future work applying these methods across sites and different gibbon species and comparisons to deep learning approaches will be crucial for future gibbon conservation initiatives across Southeast Asia.

Eyebirds: Enabling the Public to Recognize Water Birds at Hand

Enabling the public to easily recognize water birds has a positive effect on wetland bird conservation. However, classifying water birds requires advanced ornithological knowledge, which makes it very difficult for the public to recognize water bird species in daily life. To break the knowledge barrier of water bird recognition for the public, we construct a water bird recognition system (Eyebirds) by using deep learning, which is implemented as a smartphone app. Eyebirds consists of three main modules: (1) a water bird image dataset; (2) an attention mechanism-based deep convolution neural network for water bird recognition (AM-CNN); (3) an app for smartphone users. The waterbird image dataset currently covers 48 families, 203 genera and 548 species of water birds worldwide, which is used to train our water bird recognition model. The AM-CNN model employs attention mechanism to enhance the shallow features of bird images for boosting image classification performance. Experimental results on the North American bird dataset (CUB200-2011) show that the AM-CNN model achieves an average classification accuracy of 85%. On our self-built water bird image dataset, the AM-CNN model also works well with classification accuracies of 94.0%, 93.6% and 86.4% at three levels: family, genus and species, respectively. The user-side app is a WeChat applet deployed in smartphones. With the app, users can easily recognize water birds in expeditions, camping, sightseeing, or even daily life. In summary, our system can bring not only fun, but also water bird knowledge to the public, thus inspiring their interests and further promoting their participation in bird ecological conservation.

A Review of Automated Bioacoustics and General Acoustics Classification Research

Automated bioacoustics classification has received increasing attention from the research community in recent years due its cross-disciplinary nature and its diverse application. Applications in bioacoustics classification range from smart acoustic sensor networks that investigate the effects of acoustic vocalizations on species to context-aware edge devices that anticipate changes in their environment adapt their sensing and processing accordingly. The research described here is an in-depth survey of the current state of bioacoustics classification and monitoring. The survey examines bioacoustics classification alongside general acoustics to provide a representative picture of the research landscape. The survey reviewed 124 studies spanning eight years of research. The survey identifies the key application areas in bioacoustics research and the techniques used in audio transformation and feature extraction. The survey also examines the classification algorithms used in bioacoustics systems. Lastly, the survey examines current challenges, possible opportunities, and future directions in bioacoustics.

Introducing the Software CASE (Cluster and Analyze Sound Events) by Comparing Different Clustering Methods and Audio Transformation Techniques Using Animal Vocalizations

Simple Summary

Unsupervised clustering algorithms are widely used in ecology and conservation to classify animal vocalizations, but also offer various advantages in basic research, contributing to the understanding of acoustic communication. Nevertheless, there are still some challenges to overcome. For instance, the quality of the clustering result depends on the audio transformation technique previously used to adjust the audio data. Moreover, it is difficult to verify the reliability of the clustering result. To analyze bioacoustic data using a clustering algorithm, it is, therefore, essential to select a reasonable algorithm from the many existing algorithms and prepare the recorded vocalizations so that the resulting values characterize a vocalization as accurately as possible. Frequency-modulated vocalizations, whose frequencies change over time, pose a particular problem. In this paper, we present the software CASE, which includes various clustering methods and provides an overview of their strengths and weaknesses concerning the classification of bioacoustic data. This software uses a multidimensional feature-extraction method to achieve better clustering results, especially for frequency-modulated vocalizations.

Abstract

Unsupervised clustering algorithms are widely used in ecology and conservation to classify animal sounds, but also offer several advantages in basic bioacoustics research. Consequently, it is important to overcome the existing challenges. A common practice is extracting the acoustic features of vocalizations one-dimensionally, only extracting an average value for a given feature for the entire vocalization. With frequency-modulated vocalizations, whose acoustic features can change over time, this can lead to insufficient characterization. Whether the necessary parameters have been set correctly and the obtained clustering result reliably classifies the vocalizations subsequently often remains unclear. The presented software, CASE, is intended to overcome these challenges. Established and new unsupervised clustering methods (community detection, affinity propagation, HDBSCAN, and fuzzy clustering) are tested in combination with various classifiers (k-nearest neighbor, dynamic time-warping, and cross-correlation) using differently transformed animal vocalizations. These methods are compared with predefined clusters to determine their strengths and weaknesses. In addition, a multidimensional data transformation procedure is presented that better represents the course of multiple acoustic features. The results suggest that, especially with frequency-modulated vocalizations, clustering is more applicable with multidimensional feature extraction compared with one-dimensional feature extraction. The characterization and clustering of vocalizations in multidimensional space offer great potential for future bioacoustic studies. The software CASE includes the developed method of multidimensional feature extraction, as well as all used clustering methods. It allows quickly applying several clustering algorithms to one data set to compare their results and to verify their reliability based on their consistency. Moreover, the software CASE determines the optimal values of most of the necessary parameters automatically. To take advantage of these benefits, the software CASE is provided for free download.

From Soundwave to Soundscape: A Guide to Acoustic Research in Captive Animal Environments

Sound is a complex feature of all environments, but captive animals' soundscapes (acoustic scenes) have been studied far less than those of wild animals. Furthermore, research across farms, laboratories, pet shelters, and zoos tends to focus on just one aspect of environmental sound measurement: its pressure level or intensity (in decibels). We review the state of the art of captive animal acoustic research and contrast this to the wild, highlighting new opportunities for the former to learn from the latter. We begin with a primer on sound, aimed at captive researchers and animal caregivers with an interest (rather than specific expertise) in acoustics. Then, we summarize animal acoustic research broadly split into measuring sound from animals, or their environment. We guide readers from soundwave to soundscape and through the burgeoning field of conservation technology, which offers new methods to capture multiple features of complex, gestalt soundscapes. Our review ends with suggestions for future research, and a practical guide to sound measurement in captive environments.

Acoustic Sensor-Based Soundscape Analysis and Acoustic Assessment of Bird Species Richness in Shennongjia National Park, China

Passive acoustic sensor-based soundscape analysis has become an increasingly important ecological method for evaluation of ecosystem conditions using acoustic indices. Understanding the soundscape composition and correlations between acoustic indices and species richness of birds, the most important sound source in the ecosystem, are of great importance for measuring biodiversity and the level of anthropogenic disturbance. In this study, based on yearlong sound data obtained from five acoustic sensors deployed in Dalongtan, Shennongjia National Park, we analyzed the soundscape composition by comparing the distributions of the soundscape power in different frequency ranges, and examined the correlations between acoustic indices and bird species richness by means of the Spearman rank correlation coefficient method. The diurnal dynamic characteristics of acoustic indices in different seasons were also described. Results showed that the majority of sounds were in the frequency of 2-8 kHz, in which over 50% sounds were in 2-6 kHz, commonly considered the bioacoustic frequency range. The Acoustics Complexity Index, Bioacoustic Index, and Normalized Difference Soundscape Index were significantly correlated with bird species richness, suggesting that these indices can be used for evaluation of bird species richness; Apparent diurnal dynamic patterns of bird acoustic activities were observed in spring, summer, and autumn; however, the intensity and duration of bird acoustic activities in summer is larger/longer than in spring and autumn.

Utilizing DeepSqueak for automatic detection and classification of mammalian vocalizations: a case study on primate vocalizations

Bioacoustic analyses of animal vocalizations are predominantly accomplished through manual scanning, a highly subjective and time-consuming process. Thus, validated automated analyses are needed that are usable for a variety of animal species and easy to handle by non-programing specialists. This study tested and validated whether DeepSqueak, a user-friendly software, developed for rodent ultrasonic vocalizations, can be generalized to automate the detection/segmentation, clustering and classification of high-frequency/ultrasonic vocalizations of a primate species. Our validation procedure showed that the trained detectors for vocalizations of the gray mouse lemur (Microcebus murinus) can deal with different call types, individual variation and different recording quality. Implementing additional filters drastically reduced noise signals (4225 events) and call fragments (637 events), resulting in 91% correct detections (N_total = 3040). Additionally, the detectors could be used to detect the vocalizations of an evolutionary closely related species, the Goodman’s mouse lemur (M. lehilahytsara). An integrated supervised classifier classified 93% of the 2683 calls correctly to the respective call type, and the unsupervised clustering model grouped the calls into clusters matching the published human-made categories. This study shows that DeepSqueak can be successfully utilized to detect, cluster and classify high-frequency/ultrasonic vocalizations of other taxa than rodents, and suggests a validation procedure usable to evaluate further bioacoustics software.

Bioacoustic classification of avian calls from raw sound waveforms with an open-source deep learning architecture

The use of autonomous recordings of animal sounds to detect species is a popular conservation tool, constantly improving in fidelity as audio hardware and software evolves. Current classification algorithms utilise sound features extracted from the recording rather than the sound itself, with varying degrees of success. Neural networks that learn directly from the raw sound waveforms have been implemented in human speech recognition but the requirements of detailed labelled data have limited their use in bioacoustics. Here we test SincNet, an efficient neural network architecture that learns from the raw waveform using sinc-based filters. Results using an off-the-shelf implementation of SincNet on a publicly available bird sound dataset (NIPS4Bplus) show that the neural network rapidly converged reaching accuracies of over 65% with limited data. Their performance is comparable with traditional methods after hyperparameter tuning but they are more efficient. Learning directly from the raw waveform allows the algorithm to select automatically those elements of the sound that are best suited for the task, bypassing the onerous task of selecting feature extraction techniques and reducing possible biases. We use publicly released code and datasets to encourage others to replicate our results and to apply SincNet to their own datasets; and we review possible enhancements in the hope that algorithms that learn from the raw waveform will become useful bioacoustic tools.

High microphone signal-to-noise ratio enhances acoustic sampling of wildlife

BACKGROUND

Automated sound recorders are a popular sampling tool in ecology. However, the microphones themselves received little attention so far, and specifications that determine the recordings' sound quality are seldom mentioned. Here, we demonstrate the importance of microphone signal-to-noise ratio for sampling sonant animals.

METHODS

We tested 12 different microphone models in the field and measured their signal-to-noise ratios and detection ranges. We also measured the vocalisation activity of birds and bats that they recorded, the bird species richness, the bat call types richness, as well as the performance of automated detection of bird and bat calls. We tested the relationship of each one of these measures with signal-to-noise ratio in statistical models.

RESULTS

Microphone signal-to-noise ratio positively affects the sound detection space areas, which increased by a factor of 1.7 for audible sound, and 10 for ultrasound, from the lowest to the highest signal-to-noise ratio microphone. Consequently, the sampled vocalisation activity increased by a factor of 1.6 for birds, and 9.7 for bats. Correspondingly, the species pool of birds and bats could not be completely detected by the microphones with lower signal-to-noise ratio. The performance of automated detection of bird and bat calls, as measured by its precision and recall, increased significantly with microphone signal-to-noise ratio.

DISCUSSION

Microphone signal-to-noise ratio is a crucial characteristic of a sound recording system, positively affecting the acoustic sampling performance of birds and bats. It should be maximised by choosing appropriate microphones, and be quantified independently, especially in the ultrasound range.

Bioacoustics in cognitive research: Applications, considerations, and recommendations

The multifaceted ability to produce, transmit, receive, and respond to acoustic signals is widespread in animals and forms the basis of the interdisciplinary science of bioacoustics. Bioacoustics research methods, including sound recording and playback experiments, are applicable in cognitive research that centers around the processing of information from the acoustic environment. We provide an overview of bioacoustics techniques in the context of cognitive studies and make the case for the importance of bioacoustics in the study of cognition by outlining some of the major cognitive processes in which acoustic signals are involved. We also describe key considerations associated with the recording of sound and its use in cognitive applications. Based on these considerations, we provide a set of recommendations for best practices in the recording and use of acoustic signals in cognitive studies. Our aim is to demonstrate that acoustic recordings and stimuli are valuable tools for cognitive researchers when used appropriately. In doing so, we hope to stimulate opportunities for innovative cognitive research that incorporates robust recording protocols. This article is categorized under: Neuroscience > Cognition Psychology > Theory and Methods Neuroscience > Behavior Neuroscience > Cognition.

Evaluation of acoustic pattern recognition of nightingale (Luscinia megarhynchos) recordings by citizens

Acoustic pattern recognition methods introduce new perspectives for species identification, biodiversity monitoring and data validation in citizen science but are rarely evaluated in real world scenarios. In this case study we analysed the performance of a machine learning algorithm for automated bird identification to reliably identify common nightingales (Luscinia megarhynchos) in field recordings taken by users of the smartphone app Naturblick. We found that the performance of the automated identification tool was overall robust in our selected recordings. Although most of the recordings had a relatively low confidence score, a large proportion of the recordings were identified correctly.

Autonomous sound recording outperforms human observation for sampling birds: a systematic map and user guide

Autonomous sound recording techniques have gained considerable traction in the last decade, but the question remains whether they can replace human observation surveys to sample sonant animals. For birds in particular, survey methods have been tested extensively using point counts and sound recording surveys. Here, we review the latest evidence for this taxon within the frame of a systematic map. We compare sampling effectiveness of these two survey methods, the output variables they produce, and their practicality. When assessed against the standard of point counts, autonomous sound recording proves to be a powerful tool that samples at least as many species. This technology can monitor birds in an exhaustive, standardized, and verifiable way. Moreover, sound recorders give access to entire soundscapes from which new data types can be derived (vocal activity, acoustic indices). Variables such as abundance, density, occupancy, or species richness can be obtained to yield data sets that are comparable to and compatible with point counts. Finally, autonomous sound recorders allow investigations at high temporal and spatial resolution and coverage, which are more cost effective and cannot be achieved by human observations alone, even though small-scale studies might be more cost effective when carried out with point counts. Sound recorders can be deployed in many places, they are more scalable and reliable, making them the better choice for bird surveys in an increasingly data-driven time. We provide an overview of currently available recorders and discuss their specifications to guide future study designs.

ORCA-SPOT: An Automatic Killer Whale Sound Detection Toolkit Using Deep Learning

Large bioacoustic archives of wild animals are an important source to identify reappearing communication patterns, which can then be related to recurring behavioral patterns to advance the current understanding of intra-specific communication of non-human animals. A main challenge remains that most large-scale bioacoustic archives contain only a small percentage of animal vocalizations and a large amount of environmental noise, which makes it extremely difficult to manually retrieve sufficient vocalizations for further analysis - particularly important for species with advanced social systems and complex vocalizations. In this study deep neural networks were trained on 11,509 killer whale (Orcinus orca) signals and 34,848 noise segments. The resulting toolkit ORCA-SPOT was tested on a large-scale bioacoustic repository - the Orchive - comprising roughly 19,000 hours of killer whale underwater recordings. An automated segmentation of the entire Orchive recordings (about 2.2 years) took approximately 8 days. It achieved a time-based precision or positive-predictive-value (PPV) of 93.2% and an area-under-the-curve (AUC) of 0.9523. This approach enables an automated annotation procedure of large bioacoustics databases to extract killer whale sounds, which are essential for subsequent identification of significant communication patterns. The code will be publicly available in October 2019 to support the application of deep learning to bioaoucstic research. ORCA-SPOT can be adapted to other animal species.

Automated bioacoustics: methods in ecology and conservation and their potential for animal welfare monitoring

Vocalizations carry emotional, physiological and individual information. This suggests that they may serve as potentially useful indicators for inferring animal welfare. At the same time, automated methods for analysing and classifying sound have developed rapidly, particularly in the fields of ecology, conservation and sound scene classification. These methods are already used to automatically classify animal vocalizations, for example, in identifying animal species and estimating numbers of individuals. Despite this potential, they have not yet found widespread application in animal welfare monitoring. In this review, we first discuss current trends in sound analysis for ecology, conservation and sound classification. Following this, we detail the vocalizations produced by three of the most important farm livestock species: chickens ( Gallus gallus domesticus), pigs ( Sus scrofa domesticus) and cattle ( Bos taurus). Finally, we describe how these methods can be applied to monitor animal welfare with new potential for developing automated methods for large-scale farming.

Acoustic recordings provide detailed information regarding the behavior of cryptic wildlife to support conservation translocations

For translocated animals, behavioral competence may be key to post-release survival. However, monitoring behavior is typically limited to tracking movements or inferring behavior at a gross scale via collar-mounted sensors. Animal-bourne acoustic monitoring may provide a unique opportunity to monitor behavior at a finer scale. The giant panda is an elusive species of Ursid that is vulnerable to extinction. Translocation is an important aspect of the species’ recovery, and survival and recruitment for pandas likely hinge on behavioral competence. Here we tested the efficacy of a collar-mounted acoustic recording unit (ARU) to remotely monitor the behavior of panda mothers and their dependent young. We found that trained human listeners could reliably identify 10 behaviors from acoustic recordings. Through visual inspection of spectrograms we further identified 5 behavioral categories that may be detectable by automated pattern recognition, an approach that is essential for the practical application of ARU. These results suggest that ARU are a viable method for remotely observing behaviors, including feeding. With targeted effort directed towards instrumentation and computing advances, ARU could be used to document how behavioral competence supports or challenges post-release survival and recruitment, and allow for research findings to be adaptively integrated into future translocation efforts.

Quantitative acoustic differentiation of cryptic species illustrated with King and Clapper rails

Reliable species identification is vital for survey and monitoring programs. Recently, the development of digital technology for recording and analyzing vocalizations has assisted in acoustic surveying for cryptic, rare, or elusive species. However, the quantitative tools that exist for species differentiation are still being refined. Using vocalizations recorded in the course of ecological studies of a King Rail (Rallus elegans) and a Clapper Rail (Rallus crepitans) population, we assessed the accuracy and effectiveness of three parametric (logistic regression, discriminant function analysis, quadratic discriminant function analysis) and six nonparametric (support vector machine, CART, Random Forest, k-nearest neighbor, weighted k-nearest neighbor, and neural networks) statistical classification methods for differentiating these species by their kek mating call. We identified 480 kek notes of each species and quantitatively characterized them with five standardized acoustic parameters. Overall, nonparametric classification methods outperformed parametric classification methods for species differentiation (nonparametric tools were between 57% and 81% accurate, parametric tools were between 57% and 60% accurate). Of the nine classification methods, Random Forest was the most accurate and precise, resulting in 81.1% correct classification of kek notes to species. This suggests that the mating calls of these sister species are likely difficult for human observers to tell apart. However, it also implies that appropriate statistical tools may allow reasonable species-level classification accuracy of recorded calls and provide an alternative to species classification where other capture- or genotype-based survey techniques are not possible.

Body size information in large-scale acoustic bat databases

Background

Citizen monitoring programs using acoustic data have been useful for detecting population and community patterns. However, they have rarely been used to study broad scale patterns of species traits. We assessed the potential of acoustic data to detect broad scale patterns in body size. We compared geographical patterns in body size with acoustic signals in the bat species Pipistrellus pipistrellus. Given the correlation between body size and acoustic characteristics, we expected to see similar results when analyzing the relationships of body size and acoustic signals with climatic variables.

Methods

We assessed body size using forearm length measurements of 1,359 bats, captured by mist nets in France. For acoustic analyses, we used an extensive dataset collected through the French citizen bat survey. We isolated each bat echolocation call (n = 4,783) and performed automatic measures of signals, including the frequency of the flattest part of the calls (characteristic frequency). We then examined the relationship between forearm length, characteristic frequencies, and two components resulting from principal component analysis for geographic (latitude, longitude) and climatic variables.

Results

Forearm length was positively correlated with higher precipitation, lower seasonality, and lower temperatures. Lower characteristic frequencies (i.e., larger body size) were mostly related to lower temperatures and northern latitudes. While conducted on different datasets, the two analyses provided congruent results.

Discussion

Acoustic data from citizen science programs can thus be useful for the detection of large-scale patterns in body size. This first analysis offers a new perspective for the use of large acoustic databases to explore biological patterns and to address both theoretical and applied questions.

Scalable preprocessing of high volume environmental acoustic data for bioacoustic monitoring

In this work, we examine the problem of efficiently preprocessing and denoising high volume environmental acoustic data, which is a necessary step in many bird monitoring tasks. Preprocessing is typically made up of multiple steps which are considered separately from each other. These are often resource intensive, particularly because the volume of data involved is high. We focus on addressing two challenges within this problem: how to combine existing preprocessing tasks while maximising the effectiveness of each step, and how to process this pipeline quickly and efficiently, so that it can be used to process high volumes of acoustic data. We describe a distributed system designed specifically for this problem, utilising a master-slave model with data parallelisation. By investigating the impact of individual preprocessing tasks on each other, and their execution times, we determine an efficient and accurate order for preprocessing tasks within the distributed system. We find that, using a single core, our pipeline executes 1.40 times faster compared to manually executing all preprocessing tasks. We then apply our pipeline in the distributed system and evaluate its performance. We find that our system is capable of preprocessing bird acoustic recordings at a rate of 174.8 seconds of audio per second of real time with 32 cores over 8 virtual machines, which is 21.76 times faster than a serial process.

Improving Classification Algorithms by Considering Score Series in Wireless Acoustic Sensor Networks

The reduction in size, power consumption and price of many sensor devices has enabled the deployment of many sensor networks that can be used to monitor and control several aspects of various habitats. More specifically, the analysis of sounds has attracted a huge interest in urban and wildlife environments where the classification of the different signals has become a major issue. Various algorithms have been described for this purpose, a number of which frame the sound and classify these frames, while others take advantage of the sequential information embedded in a sound signal. In the paper, a new algorithm is proposed that, while maintaining the frame-classification advantages, adds a new phase that considers and classifies the score series derived after frame labelling. These score series are represented using cepstral coefficients and classified using standard machine-learning classifiers. The proposed algorithm has been applied to a dataset of anuran calls and its results compared to the performance obtained in previous experiments on sensor networks. The main outcome of our research is that the consideration of score series strongly outperforms other algorithms and attains outstanding performance despite the noisy background commonly encountered in this kind of application.

Cost-benefit analysis of acoustic recorders as a solution to sampling challenges experienced monitoring cryptic species

The inferences that can be made from any study are limited by the quality of the sampling design. By bad luck, when monitoring species that are difficult to detect (cryptic), sampling designs become dictated by what is feasible rather than what is desired. We calibrated and conducted a cost-benefit analysis of four acoustic recorder options that were being considered as potential solutions to several sampling restrictions experienced while monitoring the Australasian bittern, a cryptic wetland bird. Such sampling restrictions are commonly experienced while monitoring many different endangered species, particularly those that are cryptic. The recorder options included mono and stereo devices, with two sound file processing options (visual and audible analysis). Recording devices provided call-count data similar to those collected by field observers but at a fraction of the cost, which meant that "idealistic" sampling regimes, previously thought to be too expensive, became feasible for bitterns. Our study is one of the few to assess the monetary value of recording devices in the context of data quality, allowing trade-offs (and potential solutions) commonly experienced while monitoring cryptic endangered species to be shown and compared more clearly. The ability to overcome challenges of monitoring cryptic species in this way increases research possibilities for data deficient species and is applicable to any species with similar monitoring challenges.

Eavesdropping on the Arctic: Automated bioacoustics reveal dynamics in songbird breeding phenology

Bioacoustic networks could vastly expand the coverage of wildlife monitoring to complement satellite observations of climate and vegetation. This approach would enable global-scale understanding of how climate change influences phenomena such as migratory timing of avian species. The enormous data sets that autonomous recorders typically generate demand automated analyses that remain largely undeveloped. We devised automated signal processing and machine learning approaches to estimate dates on which songbird communities arrived at arctic breeding grounds. Acoustically estimated dates agreed well with those determined via traditional surveys and were strongly related to the landscape's snow-free dates. We found that environmental conditions heavily influenced daily variation in songbird vocal activity, especially before egg laying. Our novel approaches demonstrate that variation in avian migratory arrival can be detected autonomously. Large-scale deployment of this innovation in wildlife monitoring would enable the coverage necessary to assess and forecast changes in bird migration in the face of climate change.

Sound analysis to model weight of broiler chickens

The pattern of body weight gain during the commercial growing of broiler chickens is important to understand growth and feed conversion ratio of each flock.The application of sound analysis techniques has been widely studied to measure and analyze the amplitude and frequency of animal sounds. Previous studies have shown a significant correlation (P ≤ 0.001) between the frequency of vocalization and the age and weight of broilers. Therefore, the aim of this study was to identify and validate a model that describes the growth rate of broiler chickens based on the peak frequency of their vocalizations and to explore the possibility to develop a tool capable of automatically detecting the growth of the chickens based on the frequency of their vocalizations during the production cycle. It is part of an overall goal to develop a Precision Livestock Farming tool that assists farmers in monitoring the growth of broiler chickens during the production cycle. In the present study, sounds and body weight were continuously recorded in an intensive broiler farm during 5 production cycles. For each cycle the peak frequencies of the chicken vocalizations were used to estimate the weight and then they were compared with the observed weight of the birds automatically measured using on farm automated weighing devices. No significant difference is shown between expected and observed weights along the entire production cycles; this trend was confirmed by the correlation coefficient between expected and observed weights (r = 96%, P value ≤ 0.001).The identified model used to predict the weight as a function of the peak frequency confirmed that bird weight might be predicted by the frequency analysis of the sounds emitted at farm level. Even if the precision of the weighing method based on sounds investigated in this study has to be improved, it gives a reasonable indication regarding the growth of broilers opening a new scenario in monitoring systems in broiler houses.

The impact of environmental factors in birdsong acquisition using automated recorders

The use of automatic acoustic recorders is becoming a principal method to survey birds in their natural habitats, as it is relatively noninvasive while still being informative. As with any other sound, birdsong degrades in amplitude, frequency, and temporal structure as it propagates to the recorder through the environment. Knowing how different birdsongs attenuate under different conditions is useful to, for example, develop protocols for deploying acoustic recorders and improve automated detection methods, an essential part of the research field that is becoming known as ecoacoustics. This article presents playback and recapture (record) experiments carried out under different environmental conditions using twenty bird calls from eleven New Zealand bird species in a native forest and an open area, answering five research questions: (1) How does birdsong attenuation differ between forest and open space? (2) What is the relationship between transmission height and birdsong attenuation? (3) How does frequency of birdsong impact the degradation of sound with distance? (4) Is birdsong attenuation different during the night compared to the day? and (5) what is the impact of wind on attenuation? Bird calls are complex sounds; therefore, we have chosen to use them rather than simple tones to ensure that this complexity is not missed in the analysis. The results demonstrate that birdsong transmission was significantly better in the forest than in the open site. During the night, the attenuation was at a minimum in both experimental sites. Transmission height affected the propagation of the songs of many species, particularly the flightless ones. The effect of wind was severe in the open site and attenuated lower frequencies. The reverberations due to reflective surfaces masked higher frequencies (8 kHz) in the forest even at moderate distances. The findings presented here can be applied to develop protocols for passive acoustic monitoring. Even though the attenuation can be generalized to frequency bands, the structure of the birdsong is also important. Selecting a reasonable sampling frequency avoids unnecessary data accumulation because higher frequencies attenuate more in the forest. Even at moderate distances, recorders capture significantly attenuated birdsong, and hence, automated analysis methods for field recordings need to be able to detect and recognize faint birdsong.

Acoustic indices applied to biodiversity monitoring in a Costa Rica dry tropical forest

Standardized methods for biodiversity monitoring are needed to evaluate conservation efforts. Acoustic indices are used in biodiversity assessments, but need to be compared to traditional wildlife methods. This work was conducted in the Santa Rosa National Park between June and November, 2015. We installed recorders and conducted bird point counts in twelve sampling sites. We compared acoustic indices (Acoustic Evenness Index [AEI], Acoustic Diversity Index [ADI], Acoustic Complexity Index [ACI], Bioacoustic Index [BIO], Normalized Difference Soundscape Index [NDSI], Total Entropy [TE], Median Amplitude Envelope [MAE], Number of peaks [NP]) with indices from bird point counts (Bird Abundance, Bird Richness, Bird Diversity and Bird Evenness), and discuss the utility of acoustic indices as indicators for biodiversity monitoring in tropical forests. ADI, ACI, BIO and TE presented a similar temporal pattern peaking between 5 am and 6 am; and an additional peak at 5 pm, except for ACI. These patterns were consistent with the daily biological rhythms. AEI, ACI, BIO and Bird Abundance were related to characteristics of younger forests (lower percentage of canopy cover) but NP, ADI, TE, Bird Diversity and Bird Evenness were related to characteristics of older forests (higher percentage of canopy cover and a lower number of patches). ACI was positively correlated to Bird Abundance and NP was positively correlated to Bird Diversity. ACI reflects biological activity, but not necessarily a more diverse bird community in this study area. This might be an indication of a strong acoustic competition, or several highly dominant bird species in younger forests. Furthermore, acoustic communities in tropical forests commonly include insects (cicadas) and frogs, which might affect resulting acoustic indices. A variety of methods are probably needed to thoroughly assess biodiversity. However, a combination of indices such as ACI and NP might be considered to monitor trends in abundance and diversity of birds in dry forests.

Acoustic localization at large scales: a promising method for grey wolf monitoring

Background

The grey wolf (Canis lupus) is naturally recolonizing its former habitats in Europe where it was extirpated during the previous two centuries. The management of this protected species is often controversial and its monitoring is a challenge for conservation purposes. However, this elusive carnivore can disperse over long distances in various natural contexts, making its monitoring difficult. Moreover, methods used for collecting signs of presence are usually time-consuming and/or costly. Currently, new acoustic recording tools are contributing to the development of passive acoustic methods as alternative approaches for detecting, monitoring, or identifying species that produce sounds in nature, such as the grey wolf. In the present study, we conducted field experiments to investigate the possibility of using a low-density microphone array to localize wolves at a large scale in two contrasting natural environments in north-eastern France. For scientific and social reasons, the experiments were based on a synthetic sound with similar acoustic properties to howls. This sound was broadcast at several sites. Then, localization estimates and the accuracy were calculated. Finally, linear mixed-effects models were used to identify the factors that influenced the localization accuracy.

Results

Among 354 nocturnal broadcasts in total, 269 were recorded by at least one autonomous recorder, thereby demonstrating the potential of this tool. Besides, 59 broadcasts were recorded by at least four microphones and used for acoustic localization. The broadcast sites were localized with an overall mean accuracy of 315 ± 617 (standard deviation) m. After setting a threshold for the temporal error value associated with the estimated coordinates, some unreliable values were excluded and the mean accuracy decreased to 167 ± 308 m. The number of broadcasts recorded was higher in the lowland environment, but the localization accuracy was similar in both environments, although it varied significantly among different nights in each study area.

Conclusions

Our results confirm the potential of using acoustic methods to localize wolves with high accuracy, in different natural environments and at large spatial scales. Passive acoustic methods are suitable for monitoring the dynamics of grey wolf recolonization and so, will contribute to enhance conservation and management plans.

Detection of ground parrot vocalisation: A multiple instance learning approach

Ground parrot vocalisation can be considered as an audio event. Test-based diverse density multiple instance learning (TB-DD-MIL) is proposed for detecting this event in audio files recorded in the field. The proposed method is motivated by the advantages of multiple instance learning from incomplete training data. Spectral features suitable for encoding the vocal source information of the ground parrot vocalization are also investigated. The proposed method was benchmarked against a dataset collected in various environmental conditions and an audio detection evaluation scheme is proposed. The evaluation includes a study on performance of the various vocal source features and comparison with other classification techniques. Experimental results indicated that the most appropriate feature to encode ground parrot calls is the spectral bandwidth and the proposed TB-DD-MIL method outperformed other existing classification methods.

An FPGA-Based WASN for Remote Real-Time Monitoring of Endangered Species: A Case Study on the Birdsong Recognition of Botaurus stellaris

Fast environmental variations due to climate change can cause mass decline or even extinctions of species, having a dramatic impact on the future of biodiversity. During the last decade, different approaches have been proposed to track and monitor endangered species, generally based on costly semi-automatic systems that require human supervision adding limitations in coverage and time. However, the recent emergence of Wireless Acoustic Sensor Networks (WASN) has allowed non-intrusive remote monitoring of endangered species in real time through the automatic identification of the sound they emit. In this work, an FPGA-based WASN centralized architecture is proposed and validated on a simulated operation environment. The feasibility of the architecture is evaluated in a case study designed to detect the threatened Botaurus stellaris among other 19 cohabiting birds species in The Parc Natural dels Aiguamolls de l’Empordà, showing an averaged recognition accuracy of 91% over 2h 55’ of representative data. The FPGA-based feature extraction implementation allows the system to process data from 30 acoustic sensors in real time with an affordable cost. Finally, several open questions derived from this research are discussed to be considered for future works.