Ken Stevens' influence on aerodynamic models of fricatives
Christine H. Shadle · The Journal of the Acoustical Society of America · 2015
In the late 1950s, John Heinz, under Ken Stevens’ supervision, used a constriction in a duct to model fricatives, recording the far-field sound spectra. Ken elaborated the model over the next 15 years: The constriction’s location corresponded to place. The flowrate and constriction area were used to compute the Reynolds number, indicating the degree of turbulence of the flow; the pressure drop predicted the far-field sound. However, as I showed when Ken supervised my thesis in the 1980s, an obstacle downstream of the constriction created more turbulence noise, and was a better approximation to sibilants. It could still be modeled by a circuit analog with independent source and filter. Similar models of interdentals and palatals were not as successful; geometric details not included in the area function affected the acoustic output. Other limitations of early models include: circuit analogs use only the speed of sound, but the much-slower convection velocity plays a role in noise modulation in voiced fricatives; noise source changes during transitions are oversimplified; quasi-static modeling of fricatives excludes the noise produced by articulator movement. Ways of incorporating these mechanisms of fricative production in simple, conceptually powerful models are needed.