Comparative Bioacoustics: An Overview

There is no difference in principle between the infrasonic and ultrasonic sounds which are inaudible to humans (or other animals) and the sounds that we can hear. In all cases, sound is a wave of pressure and particle oscillations propagating through an elastic medium, such as air. This chapter is about the physical laws that govern how animals produce sound signals and how physical principles determine the signals’ frequency content and sound level, the nature of the sound field (sound pressure versus particle vibrations) as well as directional properties of the emitted signal. Many of these properties are dictated by simple physical relationships between the size of the sound emitter and the wavelength of emitted sound. The wavelengths of the signals need to be sufficiently short in relation to the size of the emitter to allow for the efficient production of propagating sound pressure waves. To produce directional sounds, even higher frequencies and shorter wavelengths are needed. In this context ‘short’ is measured relative to the size of the sound source. Some sound sources, such as dipoles and pistons, are inherently directional, whereas others, such as monopoles, are inherently omnidirectional.

As an acoustic signal travels from the source to a receiver, it is affected by a variety of physical processes, all dictated by properties of the signal and the environment. The signal energy is weakened by geometric attenuation as well as absorption by the medium. The temporal and spectral properties can be modified by sound absorption, refraction, and interference from multi paths caused by reflections. The path from the source to the receiver may be bent due to refraction. Besides geometrical attenuation, the ground effect and turbulence are the most important mechanisms to influence communication sounds for airborne acoustics and bottom and surface effects for underwater sounds. Refraction becomes very important close to shadow zones. For echolocation signals, geometric attenuation and sound absorption have the largest effects on the signals.

Laryngeal sounds in most frogs, in reptiles and most mammals are produced by the interaction between an airstream through the larynx and soft tissue vocal folds positioned laterally in the larynx. This produces a sound characterized by a fundamental frequency (F0), a spectrum of higher frequencies, amplitude and duration. The vibrating vocal folds disturb the airstream so that acoustic waves are generated which travel along the vocal tract from which a small portion of sound energy is radiated from mouth or nostrils. Laryngeal muscles are used for posturing of vocal folds, they adduct and abduct, or elongate and shorten them. Not only posturing and length changes of vocal folds affect the acoustic properties of a voice, but their morphology is also an important determinant of the vocal output. Vocal folds in frogs, reptiles and mammals are composed of several layers of tissue. An epithelial layer covers a lamina propria. The lamina propria itself can be composed of more than one layer, and the number of layers varies by species. The thyroarytenoid muscle, the third distinct structural part of a mammalian vocal fold, is located lateral to the lamina propria. The cellular and acellular morphology of vocal folds determines their viscoelastic properties, and therefore are critical in determining how the tissue responds to changes in airflow, posturing, and tension. The effect of multiple aspects on sound output, for example, (a) active movements facilitated by laryngeal muscles, (b) vocal fold morphology, (c) vocal fold viscoelastic properties and (d) vibration characteristics, can be studied in isolation, but the full picture of laryngeal biomechanics requires the investigation of the whole organ in action. One approach which we discuss here is the excised larynx experiment. Although this approach cannot reproduce natural vocal behavior, it helps reveal important aspects of the vocal fold functional morphology. The use of perfused in situ larynx preparations and the differentiated stimulation of motor efferent fibers of intrinsic laryngeal muscles, has helped characterize the acoustic space available to the vocal organ. We also describe the production of ultrasonic vocalization in rodents, which do not rely on tissue oscillation but on a purely aerodynamic process. For ultrasonic vocalization, the vocal folds are used as a dynamic obstruction of the airway to produce sound by a whistling mechanism.

Elaborate and diverse sounds are an important aspect of many bird behaviors, and these sounds are generated by sophisticated multiple motor systems regulating respiration, vocal organ and upper vocal tract structures. The avian vocal organ, the syrinx, is a unique sound generator among vertebrates, and its morphology varies substantially between different taxa. In this review an introduction to our current knowledge of the peripheral mechanisms of sound production and modification is presented in light of the methodologies that have been used to study various aspects of phonatory mechanisms. Limitations of these methods are also identified and areas for future study and needed information are discussed for each participating motor system. Respiratory control determines the coarse temporal aspects of vocalizations. Rapid switching between expiration and inspiration enables birds to take mini-breaths during inter-syllable intervals up to syllable repetition rates of approx. 30 Hz. At higher rates, pressure modulation of a sustained expiratory pulse still indicates detailed respiratory involvement in the fine control of sound production. Even during rapid sequences of expiration and inspiration, gas exchange is maintained, allowing birds to sing very long songs. Although syringeal morphology has been studied for centuries, the functional aspects of this morphological variation have only recently become subject of investigation to complement efforts focused on neural control of acoustic features. The interplay of morphology, biomechanics and neural control remains a fertile ground for future investigation of song production mechanisms and differences between avian taxa. The neuromuscular control of sound production is best understood in doves and oscine songbirds. Syringeal muscles contribute to the regulation of airflow and tension of the vibrating tissues (membranes or labia), but complex biomechanical interactions make complete understanding of the control of acoustic parameters difficult. For example, the control of sound frequency in oscines arises from a complex interplay of muscle action, physical parameters (flow and pressure gradients) and morphological specializations (extracellular matrix design of the labia). The presence of two independently controlled sound generators in some bird taxa also creates the potential for an enhanced vocal range and for complex acoustic interactions. Once sound is generated in the vocal organ, it is modified as it exits the bird through trachea and oropharyngeal spaces. This modification can be highly sophisticated, as birds can dynamically adjust resonances to track the fundamental frequency of rapidly modulated song syllables to generate tonal sounds or give rise to complex harmonic content with formant-like quality. At each motor level, many details remain to be discovered, and a thorough understanding of the peripheral mechanisms will be required for decoding the central motor program of song generation. In addition, the morphological variation in syrinx structure across different bird taxa provides a rich source for studying functional aspects of sound generation, but also for investigating evolutionary aspects of this unique and elaborate sound producing organ among vertebrates.

Sources of sound exist all-around us. Sound sources are often accompanied with a filter, and changes in the shape of the filter, like changes in the slide of a trombone, has a strong impact on the emitted sound. We first describe two archetypes of filters, the Helmholtz resonator and the pipe resonator. The Helmholtz resonator consists of a larger cavity with a narrow opening, and the pipe resonator consists of a uniform tube.

We also describe an experimental approach that allows researchers to estimate the resonance characteristic of both types of filters. Three types of sound sources are used to test a resonator: a swept sine wave, a broadband noise, and an impulse. They can be played as an input to “excite” the resonator, and the output can be recorded. The ratio of the output over the input sound provides an image of the filter’s resonance characteristics. A computational approach permits researchers to numerically predict the resonance properties of the filter based on the geometrical dimensions of the filter. The computational approach provides a reasonably accurate prediction of the resonance characteristics of both types of filters. Finally, we apply these concepts to biological systems focusing on human speech production.

Acoustic mating signals play an important role as mate attracting signals in many vertebrate and invertebrate species. Often, individuals within one population vary in the quantity and quality of their signaling effort. To test whether a signal indeed functions in mate attraction and whether variation in the signal influences this process, preference tests have been established as important research tools. This chapter reviews and explains contemporary methodology for acoustic preference testing. Special attention is given to general conceptual issues, experimental design, potential (but avoidable) experimental confounds and good testing practices.

Working in bioacoustics requires knowledge of filtering, which is the application of frequency-dependent energy attenuation. General filter types include low-pass, high-pass, band-pass, and band-stop versions, each of which involves selecting a target frequency range, corresponding corner frequencies, and an optimized combination of attenuation slope and pass-band ripple. Filters can be constructed in either analog (hardware) or digital (software) forms, the former being necessary when converting signals between these two kinds of representations. However, the latter are more flexible, less expensive, and the more common when working with digital signals. Readily available programs allow even novice users to easily design and use digital filters. Filtering applications include removing various kinds of noise, simulating environmental degradation effects, and searching for signals embedded in noise. While easily performed, each of these applications requires some background knowledge. There is also good reason to avoid unnecessary use of filtering, as it is easy to create unintended effects. This chapter discusses these and other issues in the context of the everyday work of bioacoustics.

Animal vocalizations range from tonal sounds produced by almost periodic vocal fold vibrations to completely aperiodic sounds generating noisy signals. Between these two extremes, a variety of nonlinear phenomena such as limit cycles, subharmonics, biphonation, chaos, and bifurcations have been found. This chapter introduces a concept of nonlinear dynamics and its methodology applicable to bioacoustic data. Since conventional spectral analysis is not sufficient to characterize nonlinear properties of the recorded sound signals, a temporal analysis based upon the method of nonlinear dynamics is developed. First, using a mathematical model of the vocal folds, basics of nonlinear dynamics and bifurcations are illustrated. The temporal analysis is then applied to acoustic data from real animal vocalizations. Our focus is on extracting low-dimensional nonlinear dynamics from several samples of vocalizations ranging from tonal sounds to irregular atonal sounds. We demonstrate that nonlinear analysis is a profitable approach for analyzing mammalian vocalizations with a harmonic composition or low-dimensional chaos.

After many decades of slow incremental growth, computer-based automatic recognition of human speech has recently gone through a much more rapid transition from the research lab to mainstream application, available on most of the 1+ trillion smartphones on the planet. (Worldwide, there are almost as many mobile phones as people, and about 1 in 5 of these are smartphones.) This growth has been largely fueled by the growth of raw computational power, rather than fundamental changes in speech recognition technology itself. The methods used in nearly every state-of-the-art automatic speech recognition system are based on the same statistical model that was first used for speech more than 30 years ago, the Hidden Markov Model. Hidden Markov Models are in many ways straightforward models, simple state machines that take input sequences and identify the most likely corresponding state sequences. The main strength of the approach is in its flexibility – flexibility to match sequences in a non-linear temporal pattern, flexibility to learn more detailed models if more training data is available, flexibility to connect multiple models together into longer continuous patterns, and flexibility to incorporate whatever data features and probabilistic models are best suited to the task. Nearly all of these benefits also carry over to the domain of bioacoustics, specifically to the classification of animal vocalizations. Although there are limits to this – human speech is better understood than animal communication – there is also much to gain, and many improvements that are possible by taking advantage of the large body of knowledge available through the long history of human speech processing and recognition technology. Agreeing with this idea, this chapter presents an overview of the use of Hidden Markov Models for classification, detection, and clustering of bioacoustics signals.

Humans naturally classify the sounds they hear into different categories, including sounds produced by animals. Bioacousticians have supplemented this type of subjective sorting with quantitative analyses of acoustic features of animal sounds. Using neural networks to classify animal sounds extends this process one step further by not only facilitating objective descriptive analyses of animal sounds, but also by making it possible to simulate auditory classification processes. Critical aspects of developing a neural network include choosing a particular architecture, converting measurements into input representations, and training the network to recognize inputs. When the goal is to sort vocalizations into specific types, supervised learning algorithms make it possible for a neural network to do so with high accuracy and speed. When the goal is to sort vocalizations based on similarities between measured properties, unsupervised learning algorithms can be used to create neural networks that objectively sort sounds or that quantify sequential properties of sequences of sounds. Neural networks can also provide insights into how animals might themselves classify the sounds they hear, and be useful in developing specific testable hypotheses about the functions of different sounds. The current chapter illustrates each of these applications of neural networks in studies of the sounds produced by chickadees (Poecile atricapillus), false killer whales (Pseudoorca crassidens), and humpback whales (Megaptera novaeangliae).

Audio recordings of birds and other animals, and also other forms of ‘biodiversity media’ (e.g., video recordings), capture the behavioral phenotype in ways that traditional museum specimens cannot, and natural history audio/media archives hold collections of recordings that span geography, time, and taxonomy. As such, these recordings can be used for a broad range of studies in ecology, evolution, and animal behavior, and newly developed tools for collecting and analyzing these recordings promise to further increase that research potential. Moreover, the digital revolution has made it easier than ever for high quality recordings to be collected and deposited in an archive, opening the door for large-scale citizen science efforts. But this potential also brings new challenges that must be met by the research community with regard to digital standards and accessibility. We recommend that researchers and other recordists deposit their materials in a suitable archive, that sound/media archives build strong partnerships with other types of natural history collections, that these archives also embrace technological advances to make their assets more accessible, and that archives and acoustic researchers harness “the power of the crowd” through crowd-sourcing and similar approaches. In doing so, sound archives and bioacoustic research will play an ever-increasing role in understanding our natural world, including responses of natural systems to human activities, in the 21st century.