8,798 research outputs found
Dataset for A Compact Two-Loudspeaker Virtual Sound Reproduction System for Clinical Testing of Spatial Hearing With Hearing-Assistive Devices
No full textDataset for A Compact Two-Loudspeaker Virtual Sound Reproduction System for Clinical Testing of Spatial Hearing With Hearing-Assistive Devices
Dataset DOI: https://doi.org/10.5258/SOTON/D2081
Authors: Eric Hamdan & Mark Fletcher, Auditory Implant Service, University of Southampton
This dataset supports the publication:
AUTHORS: Eric Hamdan and Mark Fletcher
TITLE: A Compact Two-Loudspeaker Virtual Sound Reproduction System for Clinical Testing of Spatial Hearing With Hearing-Assistive Devices
JOURNAL: Frontiers in Neuroscience
This dataset contains CSV files for each figure in the manuscript that has graphed data, i.e., Figures 4 - 15. The CSV files contain the x,y data for each graph (some graphs have multiple dependent variables, e.g., y1, y2, etc.). There is a folder for each figure which contains the corresponding CVS file, e.g.,
Figure5/Figure_5.csv
For figures with subfigures, i.e., (A), (B), etc., there is a CSV file for each subfigure, e.g.,
Figure7/Figure_7A.csv
Figure7/Figure_7B.csv
Figures 14A,B and 15A,B feature 2D graphs, i.e., color maps. The two independent variables are head rotation angle (degrees) and frequency (Hz). For these subfigures, there is a CSV file for each head rotation angle, e.g.,
Figure15/Figure_15A_Angle-10.csv -> corresponds to a head rotation angle of -10 degrees
.
.
.
Figure15/Figure_15A_Angle00.csv -> corresponds to a head rotation angle of 0 degrees
.
.
.
Figure15/Figure_15A_Angle10.csv -> corresponds to a head rotation angle of 10 degrees
Additionally, the folders LAC_AE/ and Booth_AE/ contain the CSV files for the smoothed absolute error (AE) associated with Figures 14 and 15 (corresponding to the large anechoic chamber (LAC) and the audiological booth). There is a CSV file for the AE calculated for each head rotation angle (as done with Figures 14A,B and 15A,B), and with and without filter compensation for the head rotation. CSV files for data with head rotation compensation are denoted with "LACC" or "BC". CSV files for data with no head rotation compensation are denoted with "LACNC" and "BNC", e.g.
LAC_AE/LACC_AEhat_Angle00.csv -> corresponds to a head rotation angle of 0 degrees, in the LAC, with head rotation compensation
LAC_AE/LACNC_AEhat_Angle00.csv -> corresponds to a head rotation angle of 0 degrees, in the LAC, no head rotation compensation
Booth_AE/BC_AEhat_Angle00.csv -> corresponds to a head rotation angle of 0 degrees, in the booth, with head rotation compensation
Booth_AE/BNC_AEhat_Angle00.csv -> corresponds to a head rotation angle of 0 degrees, in the booth, no head rotation compensation
Finally, there are stereo HAD recordings of a target speech recording played from each of the six loudspeakers used in the experiment (L3, L2, L1, R1, R2, R3) and recordings of the VA system (L1 and R1) attempting to reproduce the target signal, as detailed in section Results->Stationary Measurements of the manuscript. Recordings were made in the LAC and audiological booth. The LAC and booth recordings are in the folders HAD_Recordings/LAC and HAD_Recordings/Booth, respectively. The target recordings are denoted with the letter 'd' and the reproduced recordings are denoted with the letter 'p', e.g.,
HAD_Recordings/LAC/LAC_d_L1.wav -> recording of the target speech played from loudspeaker L1 in the LAC
HAD_Recordings/LAC/LAC_p_L1.wav -> recording of the VA system reproduced speech played from loudspeaker L1 in the LAC
HAD_Recordings/Booth/Booth_d_L1.wav -> recording of the target speech played from loudspeaker L1 in the audiological booth
HAD_Recordings/Booth/Booth_p_L1.wav -> recording of the VA system reproduced speech played from loudspeaker L1 in the audiological booth
Date of data collection: 09/09/2020 - 29/01/2021
All data was collected at the University of Southampton, U.K.</span
A compact two-loudspeaker virtual sound reproduction system for clinical testing of spatial hearing with hearing-assistive devices
No full textExciting developments in hearing aid and cochlear implant technology for linking signal processing across the ears have improved spatial hearing outcomes. This has resulted in an increased emphasis on clinical assessment of the spatial hearing abilities of hearing-assistive device users. Effective assessment of spatial hearing currently requires a large and costly loudspeaker array system, housed in a heavily acoustically treated testing room. This imposes economic and logistical constraints that limit proliferation of array systems, particularly in developing nations. Despite their size and cost, the ability of current clinical array systems to reproduce realistic spatial sound fields is limited, which substantially reduces the range of realistic acoustic scenes that can be used for diagnostic testing. We propose an alternative low-cost, compact virtual acoustics system with just two loudspeakers. This system uses crosstalk cancelation to reproduce pressure signals at the device microphones that match those for real-world sound sources. Furthermore, in contrast to clinical array systems, the system can adapt to different room acoustics, removing the requirement for a heavily acoustically treated testing environment. We conducted a proof-of-concept study in two stages: in the first, we evaluated the physical performance of the system for a stationary listener in anechoic conditions and in a small audiological testing booth with moderate acoustic treatment. To do this, a head and torso simulator was fitted with specially adapted hearing-assistive devices that allowed direct access to the microphone signals. These microphone signals were compared for real and virtual sound sources at numerous source locations. In the second stage, we quantified the system’s robustness to head rotations with and without the system adapting for head position. In the stationary case, the system was found to be highly effective at reproducing signals, such as speech, at all tested source locations. When head rotation was added, it performed well for rotations of up to 2°, even without adapting. However, performance improved markedly for larger rotations when the system adapted. These findings suggest that a compact, low-cost virtual acoustics system can give wider access to advanced and ecologically valid audiological testing, which could substantially improve clinical assessment of hearing-assistive device users.</p
A modal analysis of multichannel crosstalk cancellation systems and their relationship to amplitude panning
No full textThe single-listener multichannel crosstalk cancellation system is analysed using the singular value decomposition. Analytic expressions for the singular system are derived considering a sound-field of point sources placed in the far-field and in free-field conditions. The derived singular system yields theoretical insight into the benefit of multichannel loudspeaker geometries over the conventional two-channel system. From the singular system the source strength solution is derived for a target far-field virtual source. The low frequency approximation of the derived source strengths is an amplitude panning function. This panning function is a generalisation of the sine law solution that allows for head rotation and asymmetric loudspeaker geometry. When the geometry is symmetric about the median plane of the listener the low frequency sources strengths are the minimum ℓ2 norm solution of the multichannel sine law. The low frequency approximation is also shown to be representative when the free-field assumption is relaxed
Theoretical Advances in Multichannel Crosstalk Cancellation Systems
No full textThis thesis is concerned with the development of physical models of multichannel crosstalk cancellation (CTC) systems and the analysis of those models. The physical principles of multichannel CTC systems are established through a course of analytical studies that rely on the singular value decomposition (SVD), eigendecomposition, and a novel analysis framework that was developed in the course of these investigations. In the first of these studies, a simple acoustic model is introduced and the singular system for asymmetric and symmetric geometries is derived. An analysis of the low frequency source strength solution leads to an amplitude panning function that is shown to be a solution to the multichannel sine law. The eigenvalues of the Gram matrix reveal that the system is generally dependent on a phenomenon identified as focusing crosstalk. In the second of these works, it is shown that the multichannel CTC system robustness and stability, as quantified by the system condition number and amplification factor of the pseudoinverse, respectively, can be described directly in terms of the focusing crosstalk at a single point. This realisation leads to a physical study of the symmetric CTC system and the focusing behaviour of specific multichannel geometries. In the absence of reflections, the focusing crosstalk is largely due to sidelobe radiation, and thus an intuitive link is formed between the physical concepts of beamforming and the stability of the inverse solution. In the third of these works, the analysis culminates in an examination of the general underdetermined linear discrete inverse problem in the context of sound-field control, and is centred on the focusing behaviour inherent to the inverse solution. The physical requirements for optimally-conditioned systems are rigorously established and shown to occur only when the focusing crosstalk is minimised at all control points in the field. Finally, an alternate method to achieve CTC is developed based on the derived source strength modes of an optimally-conditioned system. The practical implications of these physical studies are discussed and supported with simulations using measured data
Ideal focusing and optimally-conditioned systems in sound field control with loudspeaker arrays
No full textThe focusing operation inherent to sound field control based on inverse filters is a fundamental aspect of the overall sound field reproduction that completely determines the system robustness and stability. The behaviour of the system is fundamentally tied to the amount of acoustic crosstalk at each control point that manifests as a result of the focus- ing operation. The maximisation of the crosstalk at just one point in space leads to linear dependence in the plant matrix and system instability. On the other hand, its minimisation leads to the ideal focusing state, wherein the sources can selectively focus at each point while a null is created at all other points. Two theoretical case studies are presented that demonstrate ideal and super ideal focusing, wherein the latter the plant matrix is optimally-conditioned. First, the application of binaural audio reproduction using an array of loudspeakers is examined and cases of ideal focusing are presented. In the process, the Optimal Source Distribution is re-derived and shown to be a specific case of super ideal focusing. Secondly, the application of recreating multiple sound zones using a linear array is examined and the conditions to achieve super ideal focusing at control points angled arbitrarily in the far-field are derived. In all cases, the ability to maintain ideal focusing as a function of frequency requires proportional changes in the source or control point geometry
Weighted orthogonal vector rejection method for loudspeaker-based binaural audio reproduction
No full textA method for reproducing binaural audio over an array of loudspeakers is presented. The proposed method is based on spatially-matched filters and is shown to be, in general, more stable than the conventional inverse filter approach. The non-weighted formulation of the method introduces reductions in the reproduced pressure magnitude response at each control point and is therefore inherently lossy. On the other hand, far less error is introduced forfrequencies where the Hermitian angle between the row vectors of the plant matrix is close to ninety degrees. For sufficiently dense arrays, little error is introduced at mid-low and high frequencies. The trade-off between reductions in the reproduced target signals and crosstalk cancellation is controlled by introducing frequency dependent weights to the left and right channel filters. The weights that minimise the error are derived. It is shown that the optimalweight solution emphasises achieved crosstalk cancellation overloss less focusing
Low frequency crosstalk cancellation and its relationship to amplitude panning
No full textThe low frequency behaviour of the source strength solution of the two loudspeaker crosstalk cancellation system is analysed. Using the rigid sphere head model, an accurate 1st order approximation of the source strength solution is derived, assuming a centred listener position. It is shown that the 1st order approximation of the source strength solution corresponds to the well-known stereo sine law. The validity of this approximation for real systems is supported by a comparison of simulated low frequency complex source strengths and the real stereo panning functions. To obtain a realistic analysis, the simulated source strengths are created using KEMAR head-related transfer functions. It is shown that the simulated source strengths converge to the stereo sine law in the low frequency limit under the expected conditions
An optimization approach to control sound source spread with multichannel amplitude panning
No full textThe perceived spread (or width) of a sound source is an important attribute of an audio object that should be controlled by a multi-channel audio reproduction system. It is desirable to either synthesize a plausible spread of a sound object or to maintain a constant spread when the virtual source moves. Existing sound spreading techniques such as Multiple Direction Amplitude Panning (MDAP) or time-frequency decomposition typically do not account for the specific loudspeaker arrangement and the inherent source spread generated by this layout. In this paper we propose an optimization-based sound field control approach, termed the `1/`2 method, to adjust the spread of a sound source. To this end we use the velocity vector magnitude to quantify the desired source spread. Based on the equivalence between amplitude panning and the maximization of this vector, the control of the source spread is formulated as a convex optimization problem. We show how the velocity objective function and additional constraints affect the resulting loudspeaker gain distribution. The proposed approach can be integrated into amplitude panning systems, and allows for either a position-independent spread of moving sources or for a smooth and continuous control of the spread parameter
Stage compression in transaural audio
No full textThe reproduction of binaural audio with loudspeakers, also referred to as transaural audio, is affected by a number of artefacts. This work focuses on the effect of reproduction error on low frequency Interaural Time Difference (ITD). Transaural systems do not provide perfect cross-talk cancellation between the left and right ear signals, especially at low frequencies. It is shown that increase in cross-talk leads to a perceived source azimuth angle that is smaller than intended. The authors show that in ideal theoretical conditions the angular error calculated from the interaural phase difference indicates stage compression for frequencies for which high cross-talk occurs. This trend is shown in the resultant ITD calculated from Interaural Cross Correlation (IACC), examined in one-third octave bands.</p
- …
