Standing on the shoulders of a giant – the inception of singing voice acoustics by Hermann von Helmholtz
The formulation of the linear source-filter theory of voice production is attributed to Chiba & Kajiyama (1941) and Fant (1960). While these works undoubtedly are major landmark publications, introducing the mathematical foundations and the electrical analog formulation of voice production acoustics, the origins of this concept reach back one more century. Here, we review Hermann von Helmholtz’ “Die Lehre von den Tonempfindungen” (1860/1872), focusing on his formulation of concepts pertaining to vowel production and singing voice acoustics.
Helmholtz hypothesizes that the voice source is generated by discontinuous, consecutive air pulses which set up a harmonic series (“Obertöne”) with a steadily decaying amplitude envelope. The strength of these harmonics is dependent on vocal fold adduction. The radiated sound is the sum of the overtones, and the amplitude of each overtone is dependent on its coincidence – or lack thereof – with resonances (“Eigentöne”) of the oral cavity. Based on sound excitation with tuning forks, Helmholtz found two resonances for neutral and front vowels and one for back vowels. These resonance frequencies are dependent on articulation. The oscillatory frequency of the vocal folds is independent of vocal tract resonances, but Helmholtz also provides a speculation about non-linear interaction between source and tract. He rudimentarily describes the singers’ formant cluster and its perceptual relevance. He further suggests that harmonic spacing is larger for higher-pitched vocalization, affecting vowel intelligibility in the treble range.
Helmholtz determined the vocal tract resonance frequencies of a number of vowels. For neutral and front vowels, he indicated two resonance frequencies, but for back vowels only one. We show here that these resonance frequencies approximately coincide with contemporary data (Hillenbrand et al., 1995) with an average absolute error of 183.8 (±142.9) Hz.
The inaccuracy of Helmholtz’ vowel formant data and the indication of only one resonance frequency for back vowels can be explained by the lack of modern measurement and analysis methodology. Despite this, Helmholtz’ theoretical formulations are surprisingly accurate and in line with modern singing voice acoustics. This essentially highlights that Helmholtz is not only the founding father of modern acoustics, but that he specifically is the initiator of contemporary (singing) voice acoustics more than 150 years ago.