Thesis (Ph. D.)--University of Canterbury, 2006. / Typescript (photocopy). "June 2006." Includes bibliographical references (p. [169]-182) and index. Also available via the World Wide Web.

The field of biological acoustics has witnessed a steady increase in the research into overtone singing, or “throat-singing,” in which a singer utilizes resonance throughout the vocal tract to sing melodies with the overtones created by a vocal drone. Recent research has explored both how a singer vocalizes in order to obtain rich harmonics from a vocal drone, as well as how further manipulations of the vocal apparatus function to filter and amplify selected harmonics. In the field of phonetics, vowel production is quantified by measuring the frequencies of vocal tract resonances, or formants, which a speaker manipulates to voice a particular vowel. Thus, an investigation of throat singing is closely linked to human speech production. Formants are usually detected in vowel spectra obtained using Fast Fourier Transform algorithms (FFTs). An alternative method that provides much higher frequency resolution is external excitation of the vocal tract and measurement of the pressure response signal at the mouth’s opening, which can be used to calculate the acoustic impedance spectrum. We demonstrate the use of such an “acoustic impedance meter” to measure the formant frequencies of common vowels as well as the oscillatory modes of simple resonant pipe systems. The impedance meter accurately measures fundamental pipe modes and a variety of formant frequencies with an uncertainty of 1 Hz. Finally, we assess how the impedance meter may be used to measure the unique resonances achieved by qualified throat singers.

Thesis (Ph. D.)--State University of New York at Binghamton, Department of Mechanical Engineering, 2007. / Includes bibliographical references (leaves 177-179).

- / -

The mechanical properties of resilient cellular materials, such as dynamic stiffness and damping, depend on several physical parameters characteristic of the material and the conditions of use, eg permeability, elastic modulus, cellular structure, static pre-strain. In many end use situations the pre-compressions and dynamic amplitudes are large and the material operates in a non-linear regime. The effects of non-linear material behaviour on the performance of systems employing cushion foams has not previously been reported on. In this work the influence of non-linear material behaviour on the vibration isolation characteristics of the material is examined. Previous theoretical and experimental studies have been confined to small strain conditions where the material behaves in a linear fashion and the properties are independent of deformation. This work extends the theoretical analysis to allow the study of the variation of the mechanical properties and vibration isolation performance with pre-strain. The fluid flow model proposed by Gent and Rusch is shown to be inadequate and an alternative proposed which conforms closely to experiment. This is extended to non-Newtonian fluids and incorporated in a model for fluid flow damping in the non-linear regime. The response of cushion foams in transportation situations is studied for small and large amplitude dynamic excitations. A multi-degree of freedom model of the person-seat system is presented and used to reproduce the responses of real vehicle seats measured in the field. The model is capable of being used to predict the optimum cushion behaviour, such as stiffness, viscoelastic damping and fluidic damping, required to enhance the ride comfort provided by a particular seat system. At higher vibration amplitudes experimental determinations show that the cushion foam behaves in a non-linear manner with strain dependent properties and several degrees of freedom. Above a certain critical excitation amplitude the classical theories of vibration isolation are shown to break down with the appearence of subharmonic frequencies in the power spectrum of the motion of the isolated mass. The resulting period doubling bifurcation cascade is similar to that found by workers in other fields. The motion of the isolated mass is complex and has not been reported previously. The behaviour is interpreted as a manifestation of chaos.

Multi-layered noise control systems are used worldwide in the automotive industry to control vehicles interior noise quality. These composites, which include moulded, and slabstock, polyurethane foam cored carpet systems, are intended to attenuate the ingress of airborne noise and suppress the radiation of structure-borne noise. However, little information has been published on their performance characteristics and the optimisation of their design. This thesis reviews previous theoretical work on the dynamics of panel vibration and the airborne acoustic insulation and the structure-borne isolation provided by composite systems. Mathematical models are developed for the acoustic behaviour of unbonded multilayer foam cored carpet composites as experienced on the experimental test rig. The models identify the important material and system parameters governing behaviour. These models, together with experimental evidence are used to optimise the design of the polyurethane foam core and rank the foam materials in order of performance. The experimental and theoretical studies are not intended for vehicle interior noise prediction purposes. The experimental facility uses a horizontal steel test panel, 1mm thick, about lmxlm, and provides acoustic and vibration excitation. Data are obtained for the effective damping (loss factor) of the 4C panel and the insertion loss (IL) of foam cored composites for both vibration and airborne excitation. The intensity method was used to measure the transmitted noise. The precision of the insertion loss measurements was shown to be better than 1.7dB for frequencies below 4kHz. For airborne excitation reproducibility was better than 1.5dB (f < 2kHz). For vibration excitation the reproducibility was less. This was attributed to the coupling method used. Experimental and theoretical studies are divided into three sections: (i) loss factors, (ii) vibration insertion loss and (iii) airborne insertion loss. Particular attention is given to airborne insertion loss since the precision of measurement allows a detailed analysis to be made. It is shown that IL passes through a minimum governed primarily by the modulus and thickness of the foam core and the surface density of the septum and steel substrate. The level of the IL depends in a complex way on material and design parameters, including the damping of the foam core. The damping of the septum mass is shown to have little effect upon behaviour. The design optimisation procedure described in the thesis takes account of the loudness of vehicle interior noise at relatively high frequencies (f > 500Hz) and the annoyance of discrete tonal noise at relatively low frequencies. It is shown that for a particular incident noise spectrum the carpet composite can be designed to provide the most acceptable noise quality in a vehicle. For a typical large volume production vehicle the optimum resonance frequency is predicted to be about 300Hz. Ways of achieving this frequency with different combinations of design parameters are described.

With axially symmetrical vibrating systems like rings, cones and bells most of the normal modes occur as degenerate doublets with the two components having equal Q values. If a slight asymmetry is introduced into the system - either structural or metallurgical - these doublets split slightly and often interest centres on the amount of this splitting. The easiest method to measure this is both to drive and detect at one point midway between the two nodal/antinodal meridians of the two components. By driving the system at this "symmetry" radian frequency of the doublet response a beating decay is observed if the drive is switched off after attaining the steady state.

<p> This study investigated the acoustic characteristics that contribute to the perception of foreign-accentedness of English spoken by native Ewe speakers. Forty monosyllabic words spoken by four speakers were rated on accentedness by 109 participants; 51 with exposure to Ghanaian-accented-American English and 58 with none. The ratings and measurements of F1 and F2 values of the vowels were analyzed and compared. The results suggest that the perception of accentedness was influenced by the acoustic properties of vowels. Listeners rated L2 speakers as more accented than they did L1 speakers. Accentedness ratings did not correlate with comprehension of words, and finally, listeners with previous exposure to Ewe accented American English rated the non-native tokens as less accented than listeners without previous exposure. Previous exposure did not influence comprehension of tokens.</p><p>

No description available.

No description available.