Comparative bioacoustics: a roadmap for quantifying and comparing animal sounds across diverse taxa

Gharial acoustic signaling: Novel underwater pops are temporally based, context-dependent, seasonally stable, male-specific, and individually distinctive

Gharials (Gavialis gangeticus) produce a sudden, high amplitude, pulsatile, underwater sound called a POP. In this study, gharial POPs ranged from 9 to 55 ms, and were clearly audible on land and water, at ≥500 m. POPs were only performed underwater by adult males possessing a sex-specific, cartilaginous narial excrescence, termed the ghara. We recorded 130 POP events of seven wild adult males in 115 km stretch of the Chambal River during 2017-2019, using hydrophones and aerial mics. A POP event occurs when a male produces a single or double or triple POP, each with a specific duration and timing. A POP event was incorporated into a complex, multi-modal breathing display, typically performed by each male during the breeding season. Key features of this novel gharial POP signal are documented here for the first time. These include its incorporation into a complex breathing display, its reliance on temporal rather than spectral elements, its dependence on a specific social context, its stability within an individual, and its individually distinctive patterning specific to a particular male. The breathing display consisted of sub-audible vibrations (SAV) preceding each POP, then a stereotyped exhalation-inhalation-exhalation sequence, concluding with bubbling and submergence. In our study, 96% of the variation in POP signal parameters was explained by POP signal timings (92%) and number of POPs (4%), and only 2% was related to spectral features. Each POP event was performed in a specific social setting. Two behavioral contexts were examined: ALERT and PATROL. In each context, male identities were examined using Discriminant Function Analysis (DFA). Within each context, each of the seven males exhibited distinctive POP patterns that were context-specific and denoted a male's identity and his location. POP signal features were stable for individual males, from 1 year to the next. Overall, the seven males showed POP patterns that were individually specific, with minimal overlap amongst males, yet these were remarkably diverse. The stereotypy of POP patterns, based on temporal versus frequency difference was best characterized statistically using DFA metrics, rather than Beecher's Information Statistic, MANOVA, or Discriminant Score computations. Our field observations indicated that audiences of gharial, located nearby, and/or in the distance, responded immediately to POPs by orienting in the signal direction. Extensive auditory studies of crocodylians indicate that their capacity for auditory temporal discrimination and neural processing in relation to locating a sound target is on par with that of birds. How the POP sound is produced and broadcast loudly in both water and air has received little study to date. We briefly summarize existing reports on ghara anatomy, ontogeny, and paleontology. Finally, preliminary observations made in a clear underwater zoo enclosure indicate that jaw claps performed entirely underwater produce POP sounds. Simultaneous bubble clouds emanating from the base of the ghara are suggestive of cavitation phenomena associated with loud high volume sounds such as shrimp snaps and seal/walrus claps. We discuss the likelihood that the adult male's ghara plays an essential role in the production of the non-vocal underwater POP, a sexually dimorphic acoustic signal unique to gharial.

BioSense: An automated sensing node for organismal and environmental biology

Automated remote sensing has revolutionized the fields of wildlife ecology and environmental science. Yet, a cost-effective and flexible approach for large scale monitoring has not been fully developed, resulting in a limited collection of high-resolution data. Here, we describe BioSense, a low-cost and fully programmable automated sensing platform for applications in bioacoustics and environmental studies. Our design offers customization and flexibility to address a broad array of research goals and field conditions. Each BioSense is programmed through an integrated Raspberry Pi computer board and designed to collect and analyze avian vocalizations while simultaneously collecting temperature, humidity, and soil moisture data. We illustrate the different steps involved in manufacturing this sensor including hardware and software design and present the results of our laboratory and field testing in southwestern United States.

Automatic detection for bioacoustic research: a practical guide from and for biologists and computer scientists

Recent years have seen a dramatic rise in the use of passive acoustic monitoring (PAM) for biological and ecological applications, and a corresponding increase in the volume of data generated. However, data sets are often becoming so sizable that analysing them manually is increasingly burdensome and unrealistic. Fortunately, we have also seen a corresponding rise in computing power and the capability of machine learning algorithms, which offer the possibility of performing some of the analysis required for PAM automatically. Nonetheless, the field of automatic detection of acoustic events is still in its infancy in biology and ecology. In this review, we examine the trends in bioacoustic PAM applications, and their implications for the burgeoning amount of data that needs to be analysed. We explore the different methods of machine learning and other tools for scanning, analysing, and extracting acoustic events automatically from large volumes of recordings. We then provide a step-by-step practical guide for using automatic detection in bioacoustics. One of the biggest challenges for the greater use of automatic detection in bioacoustics is that there is often a gulf in expertise between the biological sciences and the field of machine learning and computer science. Therefore, this review first presents an overview of the requirements for automatic detection in bioacoustics, intended to familiarise those from a computer science background with the needs of the bioacoustics community, followed by an introduction to the key elements of machine learning and artificial intelligence that a biologist needs to understand to incorporate automatic detection into their research. We then provide a practical guide to building an automatic detection pipeline for bioacoustic data, and conclude with a discussion of possible future directions in this field.

BASSA: New software tool reveals hidden details in visualisation of low-frequency animal sounds

The study of animal sounds in biology and ecology relies heavily upon time-frequency (TF) visualisation, most commonly using the short-time Fourier transform (STFT) spectrogram. This method, however, has inherent bias towards either temporal or spectral details that can lead to misinterpretation of complex animal sounds. An ideal TF visualisation should accurately convey the structure of the sound in terms of both frequency and time, however, the STFT often cannot meet this requirement. We evaluate the accuracy of four TF visualisation methods (superlet transform [SLT], continuous wavelet transform [CWT] and two STFTs) using a synthetic test signal. We then apply these methods to visualise sounds of the Chagos blue whale, Asian elephant, southern cassowary, eastern whipbird, mulloway fish and the American crocodile. We show that the SLT visualises the test signal with 18.48%-28.08% less error than the other methods. A comparison between our visualisations of animal sounds and their literature descriptions indicates that the STFT's bias may have caused misinterpretations in describing pygmy blue whale songs and elephant rumbles. We suggest that use of the SLT to visualise low-frequency animal sounds may prevent such misinterpretations. Finally, we employ the SLT to develop 'BASSA', an open-source, GUI software application that offers a no-code, user-friendly tool for analysing short-duration recordings of low-frequency animal sounds for the Windows platform. The SLT visualises low-frequency animal sounds with improved accuracy, in a user-friendly format, minimising the risk of misinterpretation while requiring less technical expertise than the STFT. Using this method could propel advances in acoustics-driven studies of animal communication, vocal production methods, phonation and species identification.

Vocal complexity in the long calls of Bornean orangutans

Vocal complexity is central to many evolutionary hypotheses about animal communication. Yet, quantifying and comparing complexity remains a challenge, particularly when vocal types are highly graded. Male Bornean orangutans (Pongo pygmaeus wurmbii) produce complex and variable "long call" vocalizations comprising multiple sound types that vary within and among individuals. Previous studies described six distinct call (or pulse) types within these complex vocalizations, but none quantified their discreteness or the ability of human observers to reliably classify them. We studied the long calls of 13 individuals to: (1) evaluate and quantify the reliability of audio-visual classification by three well-trained observers, (2) distinguish among call types using supervised classification and unsupervised clustering, and (3) compare the performance of different feature sets. Using 46 acoustic features, we used machine learning (i.e., support vector machines, affinity propagation, and fuzzy c-means) to identify call types and assess their discreteness. We additionally used Uniform Manifold Approximation and Projection (UMAP) to visualize the separation of pulses using both extracted features and spectrogram representations. Supervised approaches showed low inter-observer reliability and poor classification accuracy, indicating that pulse types were not discrete. We propose an updated pulse classification approach that is highly reproducible across observers and exhibits strong classification accuracy using support vector machines. Although the low number of call types suggests long calls are fairly simple, the continuous gradation of sounds seems to greatly boost the complexity of this system. This work responds to calls for more quantitative research to define call types and quantify gradedness in animal vocal systems and highlights the need for a more comprehensive framework for studying vocal complexity vis-à-vis graded repertoires.

DISCO: A deep learning ensemble for uncertainty-aware segmentation of acoustic signals

Recordings of animal sounds enable a wide range of observational inquiries into animal communication, behavior, and diversity. Automated labeling of sound events in such recordings can improve both throughput and reproducibility of analysis. Here, we describe our software package for labeling elements in recordings of animal sounds, and demonstrate its utility on recordings of beetle courtships and whale songs. The software, DISCO, computes sensible confidence estimates and produces labels with high precision and accuracy. In addition to the core labeling software, it provides a simple tool for labeling training data, and a visual system for analysis of resulting labels. DISCO is open-source and easy to install, it works with standard file formats, and it presents a low barrier of entry to use.

Individual identity information persists in learned calls of introduced parrot populations

Animals can actively encode different types of identity information in learned communication signals, such as group membership or individual identity. The social environments in which animals interact may favor different types of information, but whether identity information conveyed in learned signals is robust or responsive to social disruption over short evolutionary timescales is not well understood. We inferred the type of identity information that was most salient in vocal signals by combining computational tools, including supervised machine learning, with a conceptual framework of "hierarchical mapping", or patterns of relative acoustic convergence across social scales. We used populations of a vocal learning species as a natural experiment to test whether the type of identity information emphasized in learned vocalizations changed in populations that experienced the social disruption of introduction into new parts of the world. We compared the social scales with the most salient identity information among native and introduced range monk parakeet (Myiopsitta monachus) calls recorded in Uruguay and the United States, respectively. We also evaluated whether the identity information emphasized in introduced range calls changed over time. To place our findings in an evolutionary context, we compared our results with another parrot species that exhibits well-established and distinctive regional vocal dialects that are consistent with signaling group identity. We found that both native and introduced range monk parakeet calls displayed the strongest convergence at the individual scale and minimal convergence within sites. We did not identify changes in the strength of acoustic convergence within sites over time in the introduced range calls. These results indicate that the individual identity information in learned vocalizations did not change over short evolutionary timescales in populations that experienced the social disruption of introduction. Our findings point to exciting new research directions about the robustness or responsiveness of communication systems over different evolutionary timescales.

Two pup vocalization types are genetically and functionally separable in deer mice

Vocalization is a widespread social behavior in vertebrates that can affect fitness in the wild. Although many vocal behaviors are highly conserved, heritable features of specific vocalization types can vary both within and between species, raising the questions of why and how some vocal behaviors evolve. Here, using new computational tools to automatically detect and cluster vocalizations into distinct acoustic categories, we compare pup isolation calls across neonatal development in eight taxa of deer mice (genus Peromyscus) and compare them with laboratory mice (C57BL6/J strain) and free-living, wild house mice (Mus musculus domesticus). Whereas both Peromyscus and Mus pups produce ultrasonic vocalizations (USVs), Peromyscus pups also produce a second call type with acoustic features, temporal rhythms, and developmental trajectories that are distinct from those of USVs. In deer mice, these lower frequency "cries" are predominantly emitted in postnatal days one through nine, whereas USVs are primarily made after day 9. Using playback assays, we show that cries result in a more rapid approach by Peromyscus mothers than USVs, suggesting a role for cries in eliciting parental care early in neonatal development. Using a genetic cross between two sister species of deer mice exhibiting large, innate differences in the acoustic structure of cries and USVs, we find that variation in vocalization rate, duration, and pitch displays different degrees of genetic dominance and that cry and USV features can be uncoupled in second-generation hybrids. Taken together, this work shows that vocal behavior can evolve quickly between closely related rodent species in which vocalization types, likely serving distinct functions in communication, are controlled by distinct genetic loci.

DISCO: A deep learning ensemble for uncertainty-aware segmentation of acoustic signals

Recordings of animal sounds enable a wide range of observational inquiries into animal communication, behavior, and diversity. Automated labeling of sound events in such recordings can improve both throughput and reproducibility of analysis. Here, we describe our software package for labeling sound elements in recordings of animal sounds and demonstrate its utility on recordings of beetle courtships and whale songs. The software, DISCO, computes sensible confidence estimates and produces labels with high precision and accuracy. In addition to the core labeling software, it provides a simple tool for labeling training data, and a visual system for analysis of resulting labels. DISCO is open-source and easy to install, it works with standard file formats, and it presents a low barrier of entry to use.

1/f laws found in non-human music

A compelling question at the intersection of physics, neuroscience, and evolutionary biology concerns the extent to which the brains of various species evolved to encode regularities of the physical world. It would be parsimonious and adaptive, for example, for brains to evolve an innate understanding of gravity and the laws of motion, and to be able to detect, auditorily, those patterns of noises that ambulatory creatures make when moving about the world. One such physical regularity of the world is fractal structure, generally characterized by power-law correlations or 1/f ^β spectral distributions. Such laws are found broadly in nature and human artifacts, from noise in physical systems, to coastline topography (e.g., the Richardson effect), to neuronal spike patterns. These distributions have also been found to hold for the rhythm and power spectral density of a wide array of human music, suggesting that human music incorporates regularities of the physical world that our species evolved to recognize and produce. Here we show for the first time that 1/f^β laws also govern the spectral density of a wide range of animal vocalizations (music), from songbirds, to whales, to howling wolves. We discovered this 1/f^β power-law distribution in the vocalizations within all of the 17 diverse species examined. Our results demonstrate that such power laws are prevalent in the animal kingdom, evidence that their brains have evolved a sensitivity to them as an aid in processing sensory features of the natural world.

The impacts of fine-tuning, phylogenetic distance, and sample size on big-data bioacoustics

Vocalizations in animals, particularly birds, are critically important behaviors that influence their reproductive fitness. While recordings of bioacoustic data have been captured and stored in collections for decades, the automated extraction of data from these recordings has only recently been facilitated by artificial intelligence methods. These have yet to be evaluated with respect to accuracy of different automation strategies and features. Here, we use a recently published machine learning framework to extract syllables from ten bird species ranging in their phylogenetic relatedness from 1 to 85 million years, to compare how phylogenetic relatedness influences accuracy. We also evaluate the utility of applying trained models to novel species. Our results indicate that model performance is best on conspecifics, with accuracy progressively decreasing as phylogenetic distance increases between taxa. However, we also find that the application of models trained on multiple distantly related species can improve the overall accuracy to levels near that of training and analyzing a model on the same species. When planning big-data bioacoustics studies, care must be taken in sample design to maximize sample size and minimize human labor without sacrificing accuracy.

A female mimic signal type in the vibrational repertoire of male Enchenopa treehoppers

Animals vary in the complexity and size of the signal repertoire used in communication. Often, these behavioural repertoires include multiple signal types for the same process, for instance, courtship. In Enchenopa treehoppers (Hemiptera: Membracidae) mate-searching males produce plant-borne vibrational advertisement signals. Receptive females then respond to males with their own signals. Here we describe an additional signal type in the repertoire of these males. We collected nymphs in Wisconsin, USA, and recorded the spontaneous signalling bouts of adult males and duetting signals of females using laser vibrometry. Two-thirds of males produced the additional signal type, which differed in temporal and spectral features from the main male advertisement signals, whilst resembling female duetting signals in placement and acoustic features. Our findings suggest that this might be a female mimic signal. Overall, our findings highlight the diversity in the behavioural repertoire that animals may use for reproduction.

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.

Higher-order sequences of vocal mimicry performed by male Albert's lyrebirds are socially transmitted and enhance acoustic contrast

Most studies of acoustic communication focus on short units of vocalization such as songs, yet these units are often hierarchically organized into higher-order sequences and, outside human language, little is known about the drivers of sequence structure. Here, we investigate the organization, transmission and function of vocal sequences sung by male Albert's lyrebirds (Menura alberti), a species renowned for vocal imitations of other species. We quantified the organization of mimetic units into sequences, and examined the extent to which these sequences are repeated within and between individuals and shared among populations. We found that individual males organized their mimetic units into stereotyped sequences. Sequence structures were shared within and to a lesser extent among populations, implying that sequences were socially transmitted. Across the entire species range, mimetic units were sung with immediate variety and a high acoustic contrast between consecutive units, suggesting that sequence structure is a means to enhance receiver perceptions of repertoire complexity. Our results provide evidence that higher-order sequences of vocalizations can be socially transmitted, and that the order of vocal units can be functionally significant. We conclude that, to fully understand vocal behaviours, we must study both the individual vocal units and their higher-order temporal organization.

The singing question: re-conceptualizing birdsong

Birdsong has been the subject of broad research from a variety of sub-disciplines and has taught us much about the evolution, function, and mechanisms driving animal communication and cognition. Typically, birdsong refers to the specialized vocalizations produced by oscines. Historically, much of the research on birdsong was conducted in north temperate regions (specifically in Europe and North America) leading to multiple biases. Due to these historic biases these vocalizations are generally considered to be highly sexually dimorphic, heavily shaped by sexual selection and essential for courtship and territoriality. Song is also typically defined as a learned trait shaped by cultural evolution. Together, this framework focuses research specifically on males, particularly during the north temperate breeding season - reflecting and thereby reinforcing this framework. The physiological underpinnings of song often emphasize the role of the hypothalamic-pituitary-gonadal axis (associated with breeding changes) and the song control system (underlying vocal learning). Over the years there has been great debate over which features of song are essential to the definition of birdsong, which features apply broadly to contexts outside males in the north temperate region, and over the importance of having a definition at all. Importantly, the definitions we use can both guide and limit the progress of research. Here, we describe the history of these definitions, and how these definitions have directed and restricted research to focus on male song in sexually selected contexts. Additionally, we highlight the gaps in our scientific knowledge, especially with respect to the function and physiological mechanisms underlying song in females and in winter, as well as in non-seasonally breeding species. Furthermore, we highlight the problems with using complexity and learning as dichotomous variables to categorize songs and calls. Across species, no one characteristic of song - sexual dimorphism, seasonality, complexity, sexual selection, learning - consistently delineates song from other songbird vocal communication. We provide recommendations for next steps to build an inclusive information framework that will allow researchers to explore nuances in animal communication and promote comparative research. Specifically, we recommend that researchers should operationalize the axis of variation most relevant to their study/species by identifying their specific question and the variable(s) of focus (e.g. seasonality). Researchers should also identify the axis (axes) of variation (e.g. degree of control by testosterone) most relevant to their study and use language consistent with the question and axis (axes) of variation (e.g. control by testosterone in the seasonal vocal production of birds).