Sound absorption estimation of finite porous samples with deep residual learninga).

This work proposes a method to predict the sound absorption coefficient of finite porous absorbers using a residual neural network and a single-layer microphone array. The goal is to mitigate the discrepancies between predicted and measured data due to the finite-size effect for a wide range of rectangular absorbers with varying dimensions and flow resistivity and for various source-receiver locations. Data for training, validation, and testing are generated with a boundary element model consisting of a baffled porous layer on a rigid backing using the Delany-Bazley-Miki model. In effect, the network learns relevant features from the array pressure amplitude to predict the sound absorption as if the porous material were infinite. The method's performance is quantified with the error between the predicted and theoretical sound absorption coefficients and compared with the two-microphone method. For array distances close to the porous sample, the proposed method performs at least as well as the two-microphone method and significantly better than it for frequencies below 400 Hz and small absorber sizes (e.g., 20 × 20 cm2). The significance of the study lies in the possibility of measuring sound absorption on-site in the presence of strong edge diffraction.

Developing experimental skills: A hands-on course in acoustical measurement techniques at the Norwegian University of Science and Technology.

The course "Acoustical Measurement Techniques TTT4250," offered by the Acoustics Group at the Department of Electronic Systems, Norwegian University of Science and Technology, is a fourth-year course in the specialization of acoustics in the five-year master program "Electronics Systems Design and Innovation" or MTELSYS, and the two-year international master program "Electronic Systems Design" or MSELSYS. It is one of the four required courses for MTELSYS and one of the two required courses for MSELSYS. It offers a hands-on approach to acoustics. This paper outlines the topics covered in this course and the involvement of several academic staff members, as well as invited industry and research institute guest speakers, as teachers. The assessment of laboratory reports is described, and general lecture topics, including measurement uncertainty and statistics, the introduction of standards, and programming, are also described. All aspects of the course aim to maximize students' experience with a broad range of acoustic measurements and their interest in acoustics.

Acoustic signals in air and water generated by very shallow marine seismic sources: An experimental study.

When a marine seismic source, like an airgun, is fired close to the water surface the oscillating bubble interacts with the water-air interface. The main interest for seismic applications is how this effect impacts the acoustic signal propagating into the water. It is known that the sound transmission into air is abnormally strong when the sound source is very close to the sea surface relative to the emitted wavelength. Detailed insight into how the acoustic signal changes when the source depth is changed is useful in seismic data analysis and processing. Two experiments are conducted in a water tank with two different types of seismic sources. In experiment A the source is a small cavity that is sufficiently far away from the water-air interface so that it can be assumed that no interaction between the cavity and water surface occurs. In experiment B the source is a larger air bubble that is very close to the water-air interface, and hence interaction between the bubble and water surface occurs. The effects on the water surface, oscillating bubble, and emitted acoustic pressure into air are discussed. It is demonstrated that the moving surface contributes significantly to the acoustic signal measured in air.

Machine-learning-based estimation and rendering of scattering in virtual reality.

In this work, a technique to render the acoustic effect of scattering from finite objects in virtual reality is proposed, which aims to provide a perceptually plausible response for the listener, rather than a physically accurate response. The effect is implemented using parametric filter structures and the parameters for the filters are estimated using artificial neural networks. The networks may be trained with modeled or measured data. The input data consist of a set of geometric features describing a large quantity of source-object-receiver configurations, and the target data consist of the filter parameters computed using measured or modeled data. A proof-of-concept implementation is presented, where the geometric descriptions and computationally modeled responses of three-dimensional plate objects are used for training. In a dynamic test scenario, with a single source and plate, the approach is shown to provide a similar spectrogram when compared with a reference case, although some spectral differences remain present. Nevertheless, it is shown with a perceptual test that the technique produces only a slightly lower degree of plausibility than the state-of-the-art acoustic scattering model that accounts for diffraction, and also that the proposed technique yields a prominently higher degree of plausibility than a model that omits diffraction.

Modeling sound scattering using a combination of the edge source integral equation and the boundary element method.

A hybrid method for sound scattering calculations is presented in this paper. The boundary element method (BEM) is combined with a recently developed edge source integral equation (ESIE) [J. Acoust. Soc. Am. 133, 3681-3691 (2013)]. Although the ESIE provides accurate results for convex, rigid polyhedra, it has several numerical challenges, one of which applies to certain radiation directions. The proposed method, denoted ESIEBEM, overcomes this problem with certain radiation directions by applying a similar approach as BEM. First, the sound pressure is calculated on the surface of the scattering object using the ESIE, and then second, the scattered sound is obtained at the receiver point using the Kirchhoff-Helmholtz boundary integral equation, as BEM does. The three methods have been compared for the scattering by a rigid cube. Based on results from several discretizations, ESIE and ESIEBEM results are typically (90% quartile) within 3-4 · 10-4 for a kL-value of 1.83 and 2 · 10-3 for kL = 9.15, L being the cube length, of reference results computed with the BEM. The computational cost of ESIEBEM appears to be lower than BEM.

Overview of geometrical room acoustic modeling techniques.

Computerized room acoustics modeling has been practiced for almost 50 years up to date. These modeling techniques play an important role in room acoustic design nowadays, often including auralization, but can also help in the construction of virtual environments for such applications as computer games, cognitive research, and training. This overview describes the main principles, landmarks in the development, and state-of-the-art for techniques that are based on geometrical acoustics principles. A focus is given to their capabilities to model the different aspects of sound propagation: specular vs diffuse reflections, and diffraction.

Room acoustical parameters of two electronically connected rooms.

The acoustical properties of two rooms that are one-way connected electroacoustically, e.g., in a telephone/video conference, can be analyzed through the total impulse response from a source in one room to the receiver in the other room. The total impulse response is a convolution of the two involved room impulse responses, and such a model is analyzed in this paper. The room impulse response model used here facilitates convolution analysis as the model is quite simple and composed of two terms only, a direct sound term and an exponentially decaying random Gaussian noise term. Analytical expressions have been derived for the energy decay function, leading to estimates of room acoustical parameters like clarity and the modulation transfer functions for such convolved impulse responses. Background noise expressions are also introduced to allow signal-to-noise ratio studies. Estimates of acoustic parameter values have been compared with measurements to evaluate the model used and verify the results achieved.

Estimation of aircraft angular coordinates using a directional-microphone array--An experimental study.

Ground reflections cause problems when estimating the direction of arrival of aircraft noise. In traditional methods, based on the time differences between the microphones of a compact array, they may cause a significant loss of accuracy in the vertical direction. This study evaluates the use of first-order directional microphones, instead of omnidirectional, with the aim of reducing the amplitude of the reflected sound. Such a modification allows the problem to be treated as in free field conditions. Although further tests are needed for a complete evaluation of the method, the experimental results presented here show that under the particular conditions tested the vertical angle error is reduced ∼10° for both jet and propeller aircraft by selecting an appropriate directivity pattern. It is also shown that the final level of error depends on the vertical angle of arrival of the sound, and that the estimates of the horizontal angle of arrival are not influenced by the directivity pattern of the microphones nor by the reflective properties of the ground.

An integral equation formulation for the diffraction from convex plates and polyhedra.

A formulation of the problem of scattering from obstacles with edges is presented. The formulation is based on decomposing the field into geometrical acoustics, first-order, and multiple-order edge diffraction components. An existing secondary-source model for edge diffraction from finite edges is extended to handle multiple diffraction of all orders. It is shown that the multiple-order diffraction component can be found via the solution to an integral equation formulated on pairs of edge points. This gives what can be called an edge source signal. In a subsequent step, this edge source signal is propagated to yield a multiple-order diffracted field, taking all diffraction orders into account. Numerical experiments demonstrate accurate response for frequencies down to 0 for thin plates and a cube. No problems with irregular frequencies, as happen with the Kirchhoff-Helmholtz integral equation, are observed for this formulation. For the axisymmetric scattering from a circular disc, a highly effective symmetric formulation results, and results agree with reference solutions across the entire frequency range.

The diffracted field and its gradient near the edge of a thin screen.

The secondary edge source line integral formulation for diffraction modeling uses edge sources with a 1/r dependency, in addition to a directivity function. At first sight, this might seem to contradict the expected 1/r behavior for the gradient near the edge of a thin screen. An analysis is presented for the special case of perpendicular plane wave incidence onto a thin screen, which shows that the secondary edge source formulation does indeed lead to the expected behavior close to the edge of an ideal wedge.

Dynamic Systems Approach in Sensorimotor Synchronization: Adaptation to Tempo Step-Change.

This paper presents a dynamic systems model of a sensorimotor synchronization (SMS) task. An SMS task typically gives temporally discrete human responses to some temporally discrete stimuli. Here, a dynamic systems modeling approach is applied after converting the discrete events to regularly sampled time signals. To collect data for model parameter fitting, a previously published pilot study was expanded. Three human participants took part in an experiment: to tap a finger on a keyboard, following a metronome which changed tempo in steps. System identification was used to estimate the transfer function that represented the relationship between the stimulus and the step response signals, assuming a separate linear, time-invariant system for each tempo step. Different versions of model complexity were investigated. As a minimum, a second-order linear system with delay, two poles, and one zero was needed to model the most important features of the tempo step response by humans, while an additional third pole could give a somewhat better fit to the response data. The modeling results revealed the behavior of the system in two distinct regimes: tempo steps below and above the conscious awareness of tempo change, i.e., around 12% of the base tempo. For the tempo steps above this value, model parameters were derived as linear functions of step size for the group of three participants. The results were interpreted in the light of known facts from other fields like SMS, psychoacoustics and behavioral neuroscience.

Effects of Spatial Speech Presentation on Listener Response Strategy for Talker-Identification.

This study investigates effects of spatial auditory cues on human listeners' response strategy for identifying two alternately active talkers ("turn-taking" listening scenario). Previous research has demonstrated subjective benefits of audio spatialization with regard to speech intelligibility and talker-identification effort. So far, the deliberate activation of specific perceptual and cognitive processes by listeners to optimize their task performance remained largely unexamined. Spoken sentences selected as stimuli were either clean or degraded due to background noise or bandpass filtering. Stimuli were presented via three horizontally positioned loudspeakers: In a non-spatial mode, both talkers were presented through a central loudspeaker; in a spatial mode, each talker was presented through the central or a talker-specific lateral loudspeaker. Participants identified talkers via speeded keypresses and afterwards provided subjective ratings (speech quality, speech intelligibility, voice similarity, talker-identification effort). In the spatial mode, presentations at lateral loudspeaker locations entailed quicker behavioral responses, which were significantly slower in comparison to a talker-localization task. Under clean speech, response times globally increased in the spatial vs. non-spatial mode (across all locations); these "response time switch costs," presumably being caused by repeated switching of spatial auditory attention between different locations, diminished under degraded speech. No significant effects of spatialization on subjective ratings were found. The results suggested that when listeners could utilize task-relevant auditory cues about talker location, they continued to rely on voice recognition instead of localization of talker sound sources as primary response strategy. Besides, the presence of speech degradations may have led to increased cognitive control, which in turn compensated for incurring response time switch costs.

Efficient evaluation of edge diffraction integrals using the numerical method of steepest descent.

For the problem of edge diffraction from an edge of finite length a frequency-domain solution, obtained from an analytical time-domain solution, has been presented by Svensson et al. [Acta. Acust. Acust. 95, 568-572]. This formulation takes the form of a Fourier-type integral whose evaluation is expensive in the high frequency range. This paper demonstrates that by using tailored highly oscillatory quadrature methods based on asymptotic properties of the integral, accurate approximations in the high frequency case can be obtained with little computational effort.

Beamforming synthesis of binaural responses from computer simulations of acoustic spaces.

Auditorium designs can be evaluated prior to construction by numerical modeling of the design. High-accuracy numerical modeling produces the sound pressure on a rectangular grid, and subjective assessment of the design requires auralization of the sampled sound field at a desired listener position. This paper investigates the production of binaural outputs from the sound pressure at a selected number of grid points by using a least squares beam forming approach. Low-frequency axisymmetric emulations are derived by assuming a solid sphere model of the head, and a spherical array of 640 microphones is used to emulate ten measured head-related transfer function (HRTF) data sets from the CIPIC database for half the audio bandwidth. The spherical array can produce high-accuracy band-limited emulation of any human subject's measured HRTFs for a fixed listener position by using individual sets of beam forming impulse responses.