Influence of Higher Order Acoustical Propagation Modes on Variable Section Waveguide Directivity: Application to Vowel [α]

Abstract

The radiation of sound from the mouth of aspeaker is often described using aplane piston assumption. This is satisfactory as long as the wavelength is longer than the largest transverse dimension of the vocal tract. Forshorter wavelengths, higher order acoustical modes can propagate and the particle velocity field in the mouth exit plane can be nonuniform. As aconsequence, the plane piston assumption does not hold and amore accurate description of the particle velocity distribution at the mouth exit is necessary.T his can be achievedu sing the multimodal theory which allows one to account for higher order mode propagation. In order to study the influence of higher order modes on the sound radiated from the mouth, four approximations of the vocal tract shape for the vowel [A] are studied. The eccentricity of the junctions between the different portions of these geometries is used to control the propagation of ahigher order mode. The radiated sound is simulated and compared with the measurements performed on mechanical replicas at various angles in the horizontal plane. It is observed that the propagation of ah igher order mode can strongly modify the directivity.V arying the degree of eccentricity of the junction between the different sections shows that avery small eccentricity can induce the propagation of ahigher order mode. Reducing the size of the open end shows that higher order modes can affect the radiated sound even if the radiating surface is small compared to the largest transverse dimension of the geometry.

Influence of the Lip Horn on Acoustic Pressure Distribution Pattern of Sibilant /s/

“…Directivity measurements conducted with an eccentric twot ube simplification of /A/w ithout lips showed that complexd irectivity patterns, with changes in directivity within short frequencya nd angle intervals, occur in the frequencyrange above 6.5 kHz [10]. The influence of the lips on the transfer function of the /A/v ocal tract was mainly observed above 5.5 kHz [9].…”

“…This increase waso bserved for all measurement conditions (i.e.,n ear-field and far-field with the acoustic source or with flows upply). Far-field pressure fields of a concentric rigid two-tube simplification of vowel /A/v ocal tract with an infinite baffle(no lips) [ 10], showed that the amplitude at 90 • waslarger than those at 0 • and 180 • above 6.5 kHz, and the maximum difference of amplitudes between the center (90 • )a nd side regions (0 • and 180 • ) wasa pproximately 10 dB in the frequencyr ange from 2 to 10 kHz. The maximum difference in the amplitude for the replica of /s/ was15dBfor this frequencyrange (2 to 10 kHz), and these results indicate that the lip horn plays arole to enhance the directivity at the center compared to the model of /A/.…”

“…In addition, using areplica instead of ahuman speaker allows measurement of the spatial pressure pattern in the near field with ahigher spatial accuracythan the angle 15 • [3]. Indeed ahigher spatial accuracyismotivated since higher order modes associated with higher frequencies [7,8] potentially result in asignificant variation of the pressure pattern overasmall spatial interval [10].…”

Computational aeroacoustics to identify sound sources in the generation of sibilant /s/

“…To end this introductory section, we would like to remark that aside from static vowel sounds, for which an extensive literature is available (see, eg, literature [38][39][40][41][42][43][44][45][46][47][48][49] ), few papers can be found addressing the numerical simulation of other speech sounds. The reason for that is probably the complex physics beneath their generation and the associated high computational cost.…”

A Semi-Polar Grid Strategy for the Three-Dimensional Finite Element Simulation of Vowel-Vowel Sequences

“…The production of vowel sounds has been largely studied using either the Finite Element Method (FEM) in the time domain [1,2,3,4] or in the frequency domain [5,6,7], finite differences [8,9,10], multimodal approaches [11,12], or digital wave guide models [13,14]. All these models require very detailed 3D vocal tract geometries to achieve high quality vowel sounds.…”