Acoustics of the avian vocal tract a) N. H. Fletcher b) and A. Tarnopolsky School of Aerospace and Mechanical Engineering, Australian Defence Force Academy, Canberra 2600, Australia Received 3 January 1998; revised 8 July 1998; accepted 22 August 1998 The general principles underlying the acoustic performance of the avian vocal tract are examined both theoretically and experimentally. The formant resonances produced both by the total vocal tract and by the bronchial tubes are evaluated quantitatively, and their dependence upon anatomical parameters is investigated. A simplified cylindrical model for the beak is examined theoretically, and experimental results are presented that confirm the predictions of the theory for this model. A similar theoretical and experimental investigation using a more realistic conical beak model is also reported, and behaves as predicted. Finally a theoretical study of the effect of mouth volume, as influenced by tongue position, is integrated with these other studies to produce a complete analysis. The implications of these studies for understanding the acoustical behavior of avian vocal tracts are discussed. © 1999 Acoustical Society of America. S0001-49669800912-6 PACS numbers: 43.10.Ln, 43.80.Ka, 43.64.Tk FD INTRODUCTION Despite the existence of a very large literature on behav- ioral, acoustic and physiological aspects of bird song see, for example, the classic book by Greenewalt 1968 and many more recent papers there has been very little attempt to analyze or synthesize the behavior of the avian syrinx and vocal tract in quantitative terms. This contrasts with the situ- ation for the human larynx and vocal tract, for which there exists a multitude of aerodynamic and acoustic studies, both analytical and synthetic. There are, of course, obvious rea- sons for this difference, the prime one of which is that, as humans, we have a personal interest in the functioning, or malfunctioning, of our bodies, while studies of birds are mo- tivated largely by simple interest. A further difficulty in the case of birds is the great va- riety of their anatomical dimensions and the equally large variety in the form of their vocal utterances in comparison with those of humans. For these reasons the attention paid to any particular bird species is likely to be small, and it makes sense to seek out principles and approaches that can be ap- plied fairly generally. It is the purpose of the present paper to examine the passive acoustics of the avian vocal tract, to exhibit the new techniques that must be applied, and to present some results that are of fairly general applicability. Because the treatment is properly quantitative and formal- ized, it can be applied to birds with quite different anatomi- cal dimensions simply by inserting the appropriate numbers in the calculation. This is, however, only half the story, and the acoustics of the passive vocal tract must then be consid- ered quantitatively in relation to the active sound generation mechanism of the syrinx in the manner discussed in an ear- lier paper Fletcher, 1988. Papers on sound production in birds often emphasize the difference between birds and humans, but it is important to recognize also the similarities. In both cases the air flow is modulated by some kind of vibrating valve—a pair of vocal folds in humans and a membrane valve in the syrinx of birds —and the acoustic output of this valve passes through the vocal tract where it is inevitably modified by the reson- ances of this tract, which can be controlled to some extent by changing the configuration of tongue, jaw and lips in the case of humans, and of tongue and beak in the case of birds. Some human languages use whistles interpolated among voiced sounds, and all use aerodynamic noise to produce consonants; birds similarly produce voiced harmonic sounds and also more or less pure-tone whistles. A recent survey of human voice production is given in papers in the volume edited by Davis and Fletcher 1996 while Brackenbury 1982 and Casey and Gaunt 1985 have discussed the situ- ation for birds. The anatomy of psittacine birds resembles that of hu- mans in that there is one vocal valve located at the junction of the bronchi and trachea, though instead of consisting of two cartilaginous fleshy folds it comprises two membranes that are brought into opposition by inflating an external air sac. The actual motion of the vocal valve is complex in each case, and involves multiple degrees of freedom, leading to wavelike motion, or at least to a phase difference between different parts of the valve structure Ishizaka and Flanagan, 1972; Sundberg, 1987. Nevertheless, it is possible to model the behavior to a reasonable approximation by assuming a much simpler single-mass-and-spring type of motion and considering the forces exerted upon it during air flow Fletcher, 1988. In song birds Oscines, in contrast, the syrinx consists of two separate syringeal valves, one in each bronchus below its junction with the trachea. Each of these is a membrane valve and can be modeled as above. Some birds use only one of these valves for singing, but some use both simulta- neously, producing either the same or a different frequency Suthers, 1990, 1994. In either case a characteristic bron- chial resonance can be observed in the song. In both birds and humans the vocal valve is character- a ‘‘Selected research articles’’ are ones chosen occasionally by the Editor- in-Chief that are judged a to have a subject of wide acoustical interest, and b to be written for understanding by broad acoustical readership. b Permanent address: Research School of Physical Sciences and Engineer- ing, Australian National University, Canberra 0200, Australia. 35 35 J. Acoust. Soc. Am. 105 (1), January 1999 0001-4966/99/105(1)/35/15/$15.00 © 1999 Acoustical Society of America