The following is based on an article I wrote for Physics Bulletin (Vol. 39, No. 7, July 1988). It is a short summary of the research I have done into the physic of the didgeridoo. It assumes some knowledge of physics, so it may be too technical for some people, but it is a lot more accessible than my original thesis! If there is enough interest in this subject I may write a simpler explanation of these concepts sometime in the future.

The Physics of the Didgeridoo

The didgeridoo is a traditional instrument of the Aborigines of northern Australia. It is absurdly simple, consisting of a thin tree trunk 4-5 feet long with a hole just over one inch in diameter down it's length, hollowed by insects. It is played by blowing into it while buzzing the lips, as with the more familiar brass instruments, and generally produces a single note, so it may come as a surprise that a study of the physics of the didgeridoo should shed light on an important issue in musical acoustics.

The didgeridoo is played using the technique of circular breathing, which enables the player to sustain a continuous drone. The tonal quality (timbre) of the sound can be controlled to a very great extent by altering the size of the cavity in the mouth- that is, by controlling the vocal tract resonance. In the established literature on musical acoustics, however, such resonances are considered to have a very small effect on sound production in wind instruments. Over the past seven years I have taught myself how to play the didgeridoo, and in a final-year undergraduate project at Boston University (supervised by Lee Roberts) I set out to investigate why the vocal tract resonance has such a large effect on the tonal quality of this instrument.

One of the main characteristics of the changing timbre of the didgeridoo is a high frequency 'nasal" quality tone which shifts in frequency as the size of the mouth cavity is altered. An examination of the acoustic spectra of these different sounds reveals that a range of higher harmonics are selectively emphasised (figure 4). When speech sounds are analysed in this way, similar humps are observed in the frequency spectrum. They are caused by resonances in the vocal tract selectively amplifying some of the harmonics of the voice and are called formants. With the didgeridoo the formant frequency varies from about 750 to 2750 Hz; this is caused by a vocal tract resonance similar to that used in whistling.