This research focused on the soundscape design of a wide range of small to medium sized water features (waterfalls, fountains with upward jet(s), and streams) which can be used in gardens or parks for promoting peacefulness and relaxation in the presence of road traffic noise. Firstly, the thesis examined the audio-visual interaction and perceptual assessment of water features, including the semantic components and the qualitative categorisation and evocation of water sounds; and secondly, the thesis investigated the effectiveness of the water features tested in promoting relaxation through sound mapping. Different laboratory tests were carried out, and these included paired comparison tests (audio-only, visual-only and audio-visual tests), semantic differential tests, as well as tests aimed at the qualitative categorisation and evocation of water features. Sound maps of the water generated sounds were developed through the use of propagation models based on either point or line sources. Three acoustic zones (‘water sounds dominant zone’, ‘optimum zone’ and ‘RTN dominant zone’ (RTN: road traffic noise)) were defined in the maps as the zones where relaxation/pleasantness can be promoted over a 20 m × 20 m area for different road traffic noise levels. Paired comparisons highlighted the interdependence between uni-modal (audio-only or visual-only) and bi-modal (audio-visual) perception, indicating that equal attention should be given to the design of both stimuli. In general, natural looking features tended to increase preference scores (compared to audio-only paired comparison scores), while manmade looking features decreased them. Semantic descriptors showed significant correlations with preferences and were found to be more reliable design criteria than physical parameters. A principal component analysis identified three components within the nine semantic attributes tested: “emotional assessment,” “sound quality,” and “envelopment and temporal variation.” The first two showed significant correlations with audio-only preferences, “emotional assessment” being the most important predictor of preferences, and its attributes naturalness, relaxation, and freshness also being significantly correlated with preferences. Categorisation results indicated that natural stream sounds are easily identifiable (unlike waterfalls and fountains), while evocation results showed no unique relationship with preferences. The results of sound maps indicated that small to medium sized water features can be used mainly in environments where road traffic noise levels are equal or lower than 65 dBA.

It is widely accepted that performing musicians adjust their technique according to the acoustic conditions they hear on stage. It is likely that a musician performing in favourable acoustic conditions will give a higher quality performance. However, preferred conditions for performers are comparatively less well understood than for audience members. This presents a significant challenge when attempting to design a successful auditorium. Stage acoustic conditions are commonly assessed in terms of the overall energy of early reflections, relative to the direct sound, and reverberation time. These parameters relate to two subjective attributes of high importance to performers. However, these parameters are independent of the spatial or temporal distribution of the reflected energy which, in auditorium acoustics, are known to influence the perception of sound. It is proposed that a similar effect is observed for soloist performers and that these aspects of the soundfield will influence the perceived quality of the acoustic conditions. This research aims to observe how the spatial and temporal distribution of early reflections varies for differing stage enclosures and to determine if these factors influence a soloist’s impression of the stage acoustics. A detailed acoustic survey of eight concert hall stages has been undertaken to characterise how the spatio-temporal distribution of early energy varies under different circumstances. This includes musician related aspects such as position on stage and orientation in addition to venue related features, such as the geometry of the stage enclosure. Spatial soundfield measurement and analysis techniques are developed to enable the spatial and temporal characteristics of early reflections to be observed. A set of objective parameters are developed to formally characterise these observations. An interactive listening test allows experienced musicians to compare a series of virtual stage enclosures by playing their instrument. Test subjects rate each hall in terms of preference and in relation to specific subjective attributes. The listening test uses a real-time auralisation system to render the acoustic conditions of a concert hall, in controlled laboratory conditions. This auralisation is based on Spatial Impulse Response Rendering (SIRR) to accurately render stage acoustic conditions over a loudspeaker array. This research proposes new methods of measuring and assessing stage acoustic conditions which will aid in the design of future auditoria. In addition, this research demonstrates the use of more recent spatial audio techniques in stage acoustic laboratory experiments.

This thesis examines speech intelligibility and multi-lingual communication, in terms of acoustics and perceptual factors. More specifically, the work focused on the impact of room acoustic conditions on the speech intelligibility of four languages representative of a wide range of linguistic properties (English, Polish, Arabic and Mandarin). Firstly, diagnostic rhyme tests (DRT), phonemically balanced (PB) word lists and phonemically balanced sentence lists have been compared under four room acoustic conditions defined by their speech transmission index (STI = 0.2, 0.4, 0.6 and 0.8). The results obtained indicated that there was a statistically significant difference between the word intelligibility scores of languages under all room acoustic conditions, apart from the STI = 0.8 condition. English was the most intelligible language under all conditions, and differences with other languages were larger when conditions were poor (maximum difference of 29% at STI = 0.2, 33% at STI = 0.4 and 14% at STI = 0.6). Results also showed that Arabic and Polish were particularly sensitive to background noise, and that Mandarin was significantly more intelligible than those languages at STI = 0.4. Consonant-to-vowel ratios and languages’ distinctive features and acoustical properties explained some of the scores obtained. Sentence intelligibility scores confirmed variations between languages, but these variations were statistically significant only at the STI = 0.4 condition (sentence tests being less sensitive to very good and very poor room acoustic conditions). Additionally, perceived speech intelligibility and soundscape perception associated to these languages was also analysed in three multi-lingual environments: an airport check-in area, a hospital reception area, and a café. Semantic differential analysis showed that perceived speech intelligibility of each language varies with the type of environment, as well as the type of background noise, reverberation time, and signal-to-noise ratio. Variations between the perceived speech intelligibility of the four languages were only marginally significant (p = 0.051), unlike objective intelligibility results. Perceived speech intelligibility of English appeared to be mostly affected negatively by the information content and distracting sounds present in the background noise. Lastly, the study investigated several standards and design guidelines and showed how adjustments could be made to recommended STI values in order to achieve consistent speech intelligibility ratings across languages.

The discomfort caused by vertical vibration depends on the magnitude and frequency of vibration, but little is known about how discomfort depends on the magnitude and frequency of mechanical shocks or on body posture. The main objectives of this thesis were to advance understanding: (i) of how the discomfort caused by a vertical mechanical shock depends on the nominal frequency, magnitude, and direction of the shock and seating dynamics, and (ii) of the effects of body posture on vibration comfort when sitting and standing. Three of the four experiments presented in this thesis investigate the discomfort caused by mechanical shocks in an upright sitting posture. The first experiment compared the frequency-dependence of discomfort caused by shocks and sinusoidal vibration in the range 0.5 to 16 Hz at vibration magnitudes less than ±9.4 ms-2. A different frequency-dependence was found for shocks and for vibration, with shocks being less uncomfortable than vibration at frequencies greater than 4 Hz. The difference is explained by shocks containing energy at frequencies other than their fundamental frequency. The rates of growth of discomfort depended on frequency, indicating an effect of magnitude on the frequency-dependence of discomfort caused by shocks and vibration. A second experiment investigated the effect of shock direction (i.e., up or down) on discomfort in the range 2 to 5 Hz with peak accelerations from 7 to 11 ms-2. Upward displacements at frequencies from 2 to 4 Hz were more uncomfortable than downward displacements when the peak acceleration approached or exceeded 1 g. This was explained by the human body leaving, and subsequently impacting with, the seat. A third experiment found that a three degree-of-freedom model is able to predict SEAT values of blocks of polyurethane foam when people are exposed to shocks in the range 1 to 16 Hz. Predicted and measured SEAT values were consistent with subjective responses at most frequencies and magnitudes investigated. A fourth experiment investigated how the discomfort caused by vertical vibration depends on the frequency and magnitude of vertical vibration (0.5 to 16 Hz at 0.3 to 3.2 ms-2 r.m.s.) in four postures. The frequency-dependence of discomfort was equivalent to the standardised frequency weighting Wb when sitting upright, sitting leaning forward, and standing with straight legs. When standing, bending the legs increased discomfort in the range 2 to 4 Hz but reduced discomfort at frequencies greater than 5 Hz, consistent with the effects of bending the legs on biodynamic responses. There are four main findings from the research reported in this thesis: (i) The same methods can be used to predict the discomfort caused by shocks and vibration but the optimum frequency weighting for evaluating shocks depends on the shock magnitude; (ii) Shocks with fundamental frequencies in the range 4 to 16 Hz cause less discomfort than vibration of the same frequency and magnitude; (iii)The SEAT value is a useful predictor of seat comfort and a three degree-of-freedom model can be used to predict SEAT values of occupied foam cushions during exposures to vertical shocks in the range 1 to 16 Hz with peak accelerations less than 1g; (iv) The frequency-dependence of discomfort caused by vertical vibration is similar in normal standing and when sitting upright or sitting leaning forward, but differs when standing with bent legs.

Locating the source of small arms fire is deemed a mission-critical auditory task by infantry personnel (Bevis et al. 2014; Semeraro et al. 2015). Little is known about the acoustic localisation cues within a gunshot and human ability to localise gunshots. Binaural recordings of ‘live’ gunshots from an SA80 rifle were obtained using a KEMAR dummy head placed 100 m from the firer, within 30 cm of the bullet trajectory and with 13 azimuth angles from 90° left to 90º right. The ‘crack’, created by the supersonic bullet passing the target, produced smaller interaural time and level differences than the ‘thump’, created by the muzzle blast, for the rifle at the same angle. Forty normal-hearing listeners (20 civilian, 20 military personnel) and 12 hearing impaired listeners (all military personnel) completed a virtual azimuthal localisation task using three stimuli created from the recordings (whole gunshot, ‘crack’ only and ‘thump’ only) plus a 50 ms broadband noise burst convolved with KEMAR impulse responses. All listeners localised all stimuli types above chance level. Average localisation error increased in the order of: noise burst < thump < gunshot < crack, for all cohorts. Military personnel (regardless of their hearing level) performed significantly worse than civilians for all stimuli; they had a higher tendency to select the extreme left and right sources, resulting in an increased lateral bias. The difference between military and civilian participants may be due to their understanding of the task or military training/experience. Mild to moderate bilateral symmetrical sensorineural hearing loss did not have a significant impact on localisation accuracy. This suggests that, providing the gunshot is clearly audible and audiometric thresholds are equal between the ears, binaural cues will still be accessible and localisation accuracy will be preserved. Further work is recommended to investigate the relationship between other hearing loss configurations and small arms gunshot localisation accuracy before considering gunshot localisation as a measure of auditory fitness for infantry personnel.

Acoustically treated, lined ducts are used in a wide range of applications, one of which is a transmission-line loudspeaker (TLL), which consists of a long, acoustically-lined, folded duct attached to the rear of the loudspeaker driver. Consequently, knowledge and understanding of sound propagation within acoustically treated ducts is essential in order to be able to create and analyse designs for the intended applications. The lowfrequency driver of a loudspeaker creates pressure fluctuations on both sides of the diaphragm. Therefore, a loudspeaker cabinet of some sort is required to control the sound radiation from the rear of the driver and to prevent the unwanted interference of those sounds with that radiated from the front of the loudspeaker. The transmission-line loudspeakers are however, designed and optimized to control this rear driver radiations by redirecting the pressure at the back of the driver and use them to extend the overall low-frequency response of the loudspeaker system. Transmission-line loudspeakers rely on the use of sound absorbing materials and, although attempts at modelling the performance of these have been reported in the literature, most transmission-line loudspeakers are designed empirically, using a combination of experience and trial-and-error. This project is concerned with creating and evaluating an engineering method of accurately modelling the sound propagating inside the transmission-line loudspeaker waveguides. Loudspeaker systems inherently suffer from an insufficient low-frequency response, due to their inefficiency at low-frequencies. Therefore, TLL rely on the use of sound absorbing materials added on their internal boundaries to extend their overall response of the loudspeaker at the lowfrequency region. The acoustic load on the driver and the sound radiated from the open end of the TLL duct both depend upon the propagation of sound through the duct; and the physical length of the duct determines the frequencies that can propagate within it. The addition of sound absorbing materials along the interior boundaries of the TLL reduces the speed of propagating sound within it, causing the TLL to respond such as having a much longer internal waveguide, consequently accommodating far lower frequencies within the TLL duct, extending the overall response of the loudspeaker system. The characteristics of sound propagation through a variety of two-dimensional and three-dimensional acoustically lined ducts at low-frequencies have been analyzed. Analytical models of straight ducts have been compared with the developed numerical models. In this research dissipative mufflers, that consist of ducts lined on the inside with an acoustically absorptive material, have been considered. Starting with the propagation of sound within hard-walled boundary condition ducts, this investigation moves to the modelling of waveguides treated with locally-reacting acoustic liners and next into the analysis of ducts treated with bulk-reacting acoustic absorbent materials; two kinds of excitations have been considered, namely pistonic and non-uniform excitation. The impedance mismatch and acoustic dissipation between the sound absorbing layer and the free propagation within the duct has been modelled numerically, and the results have been compared with the in-situ measurements conducted on a range of acoustically treated and purpose built transmission-line loudspeakers. A wide range of sound absorbing materials, namely fibrous and porous absorbers, have been characterized using their low-resistivity and acoustic impedance. Based on their individual characteristics, acoustical optimization was applied on a simple geometry U-shaped TLL duct.

Advances in computer technology and low cost cameras open up new possibilities for three dimensional (3D) sound reproduction. The problem is to update the audio signal processing scheme for a moving listener, so that the listener perceives only the intended virtual sound image. The performance of the audio signal processing scheme is limited by the condition number of the associated inversion problem. The condition number as a function of frequency for different listener positions and rotation is examined using an analytical model. The resulting size of the "operational area" with listener head tracking is illustrated for different geometries of loudspeaker configurations together with related cross-over design techniques. An objective evaluation of cross-talk cancellation effectiveness is presented for different filter lengths and for asymmetric and symmetric listener positions. The benefit of using an adaptive system compared to a static system is also illustrated. The measurement of arguably the most comprehensive KEMAR database of head related transfer functions yet available is presented. A complete database of head related transfer functions measured without the pinna is also presented. This was performed to provide a starting point for future modelling of pinna responses. The update of the audio signal processing scheme is initiated by a visual tracking system that performs head tracking without the need for the listener to wear any sensors. The solution to the problem of updating the filters without any audible change is solved by using either a very fine mesh for the inverse filters or by using commutation techniques. The filter update techniques are evaluated with subjective experiments and have proven to be e®ective both in an anechoic chamber and in a listening room, which supports the implementation of virtual sound imaging systems under realistic conditions. The design and implementation of a visually adaptive virtual sound imaging system is carried out. The system is evaluated with respect to filter update rates and cross-talk cancellation effectiveness.

Sound propagation in complex non-uniform mean flows is an important research area for transport, building and power generation industries. Unsteady flows are responsible for noise generation in rotating and pulsating machines. Sound propagates in ducts and radiates through their openings. Duct discontinuities and complex flow effects on acoustic propagation need to be investigated. Although it provides accurate results, the most commonly used Computational AeroAcoustics propagation method, the full potential theory, does not describe the whole physics. Turbofan exhaust noise radiation involves strong refraction of the sound field occurring through jet shear layer, as well as interaction between the acoustic field and the vorticity/entropy waves. The Linearised Euler Equations are able to represent these effects. Solving these equations with time-domain solvers presents shortcomings such as linear instabilities and impedance modelling, which can be avoided by solving in the frequency domain. Nevertheless the classical Finite Element Method in frequency domain suffers from dispersion error and high memory requirements. These drawbacks are particularly critical at high frequencies and with the Linearised Euler Equations, which involve up to five unknowns. To circumvent these obstacles a novel approach is developed in this thesis, using a high-order Finite Element Method to solve the Linearised Euler Equations in the frequency domain. The model involves high-order polynomial shape functions with unstructured triangular meshes, numerical stabilisation and Perfectly Matched Layers. The computational effort is further optimised by coupling the Linearised Euler Equations in the regions of complex sheared mean flow with the Linearised Potential Equation in the regions of irrotational mean flow. The numerical model is applied to aeroengine acoustic propagation either by an intake or by an exhaust. Comparisons with analytic solutions demonstrate the method accuracy which properly represents the acoustic and vorticity waves, as well as the refraction of the sound field across the jet shear layer. The benefits in terms of memory requirements and computation time are significant in comparison to the standard low-order Finite Element Method, even more so with the coupling technique.

Ride quality in vehicles can be influenced by multiple-input vibration: vibration simultaneously transmitted to the body via multiple vibrating surfaces, such as steering wheels, seats, and floors. Standardised methods of predicting the discomfort of multiple-input vibration (ISO 2631-1, 1997) do not take into account vibration of the hands or the phase between vibration input to the hands, the feet, and the seat. This research seeks to understand the mechanisms associated with discomfort caused by multiple-input vertical vibration, examining the effect of vibration frequency, vibration magnitude, input location, and phase between vibration at the hands, the seat, and the feet, at frequencies between 2 and 12.5 Hz. A total of four psychophysical laboratory experiments were conducted. Three studies were designed to expand knowledge of the absolute and relative sensitivity to vibration at the inputs, or between the three inputs, by determining: (i) absolute thresholds for perception of vibration (minimum levels of vibration that can be detected) at the hands and at the feet (Experiment 1); (ii) equivalent comfort contours for vibration (levels of vibration that produce similar discomfort across the range of frequencies) at the hands and at the feet (Experiment 2); and (iii) the equivalence of vibration sensation (vibration magnitudes required to produce equivalent discomfort) between the hands and the feet (from 2 to 12.5 Hz), between the hands and the seat (from 4 to 12.5 Hz), and between the seat and the feet (from 4 to 12.5 Hz) (Experiment 3). The findings indicate similar absolute thresholds at the hands and the feet between 2 and 5 Hz, with the thresholds having constant velocity. At frequencies between 5 and 12.5 Hz, absolute thresholds have constant acceleration at the feet while continuing to have constant velocity at the hands: an indication that different mechanisms may be involved in the detection of threshold levels of vibration at the hands and the feet between 5 and 12.5 Hz. Contours of equivalent comfort at the hands and the feet between 2 and 12.5 Hz resemble the shapes of the absolute thresholds, with little dependence in vibration magnitude. The relative magnitudes of vibration required to produce equivalent discomfort at the hands and the feet, or at the hands and the seat, depend on the frequency of vibration and indicate greater sensitivity to vertical vibration at the seat than at the hands and the feet over the frequency range investigated. A clear difference in the frequency-dependence of sensitivity to vertical vibration at the seat from those at the hands and the feet suggests a separate mechanism for detecting supra-threshold vibration of the seat. The final study (Experiment 4) concerned the perception and discomfort of multiple-input vertical vibration with various phases between vibration at pairs of inputs, so as to determine: (a) difference thresholds (just noticeable differences, JNDs) for the detection of phase differences, (b) the effect of phase on discomfort, and (c) the localisation (i.e. body location) at which phase differences are detected. The results suggest that phase differences between the hands and the feet at frequencies greater than 5 Hz are unlikely to be detected, and that any phase differences between the hands and the feet over the range 2 to 12.5 Hz are unlikely to influence discomfort. Phase differences between the seat and the hands, or between the seat and the feet, over the range 4 to 12.5 Hz were detected by some subjects and increased discomfort up to 50% (when changing phase only, with no change in vibration magnitude). Over the three inputs (i.e. the hands, the feet, and the seat) sensitivity to vertical vibration was greatest at the seat and least at the hands. The phase between the inputs can affect discomfort, but only when vibration is applied at the seat. Changes in discomfort due to vibration frequency, vibration magnitude, or phase differences between the inputs may be partly associated with changes in the body locations experiencing greatest discomfort, due to different paths for the transmission of vibration into the body. The research has increased understanding of how multiple-input vibration applied simultaneously at the hands, the seat, and the feet contribute to produce an overall sensation of discomfort.

In this research, the concept of auralization is explored taking into account a hybrid numerical approach to establish good options for calculating sound wave propagation and the application of virtual sound environments to evaluate acoustical conditions of a classroom, in order to determine the impact of acoustic variables on cognitive processes. The hybrid approach considers the combination of well-established Geometrical Acoustic (GA) techniques and the Finite Element Method (FEM), contemplating for the latter the definition of a real valued impedance boundary condition related to absorption coefficients available in GA databases. The realised virtual sound environments are verified against real environment measurements by means of objective and subjective methods. The former is based on acoustic measurements according to international standards, in order to evaluate the numerical approaches used with established acoustic indicators to assess sound propagation in rooms. The latter comprises a subjective test comparing the virtual auralizations to the reference ones, which are obtained by means of binaural impulse response measurements. The first application of the auralizations contemplates an intelligibility and listening difficulty subjective test, considering different acoustic conditions of reverberation time and background noise levels. The second application studies the impact of acoustic variables on the cognitive processes of attention, memory and executive function, by means of psychological tests.