Acoustics Research Institute

The recent theory of compressive sensing leverages upon the structure of signals to acquire them with much fewer measurements than was previously thought necessary, and certainly well below the traditional Nyquist-Shannon sampling rate. However, most implementations developed to take advantage of this framework revolve around controlling the measurements with carefully engineered material or acquisition sequences. Instead, we use the natural randomness of wave propagation through multiply scattering media as an optimal and instantaneous compressive imaging mechanism. Waves reflected from an object are detected after propagation through a well-characterized complex medium. Each local measurement thus contains global information about the object, yielding a purely analog compressive sensing method. We experimentally demonstrate the effectiveness of the proposed approach for optical imaging by using a 300-micrometer thick layer of white paint as the compressive imaging device. Scattering media are thus promising candidates for designing efficient and compact compressive imagers.

We consider the problem of designing spectral graph filters for the construction of dictionaries of atoms that can be used to efficiently represent signals residing on weighted graphs. While the filters used in previous spectral graph wavelet constructions are only adapted to the length of the spectrum, the filters proposed in this paper are adapted to the distribution of graph Laplacian eigenvalues, and therefore lead to atoms with better discriminatory power. Our approach is to first characterize a family of systems of uniformly translated kernels in the graph spectral domain that give rise to tight frames of atoms generated via generalized translation on the graph. We then warp the uniform translates with a function that approximates the cumulative spectral density function of the graph Laplacian eigenvalues. We use this approach to construct computationally efficient, spectrum-adapted, tight vertex-frequency and graph wavelet frames. We give numerous examples of the resulting spectrum-adapted graph filters, and also present an illustrative example of vertex-frequency analysis using the proposed construction.

In this paper, we present a new algorithm to estimate a signal from its short-time Fourier transform modulus (STFTM). This algorithm is computationally simple and is obtained by an acceleration of the well-known Griffin-Lim algorithm (GLA). Before deriving the algorithm, we will give a new interpretation of the GLA and formulate the phase recovery problem in an optimization form. We then present some experimental results where the new algorithm is tested on various signals. It shows not only significant improvement in speed of convergence but it does as well recover the signals with a smaller error than the traditional GLA.

Signal analysis with classical Gabor frames leads to a fixed time-frequency resolution over the whole time-frequency plane. To overcome the limitations imposed by this rigidity, we propose an extension of Gabor theory that leads to the construction of frames with time-frequency resolution changing over time or frequency. We describe the construction of the resulting nonstationary Gabor frames and give the explicit formula for the canonical dual frame for a particular case, the painless case. We show that wavelet transforms, constant-Q transforms and more general filter banks may be modeled in the framework of nonstationary Gabor frames. Further, we present the results in the finite-dimensional case, which provides a method for implementing the above-mentioned transforms with perfect reconstruction. Finally, we elaborate on two applications of nonstationary Gabor frames in audio signal processing, namely a method for automatic adaptation to transients and an algorithm for an invertible constant-Q transform.

The Linear Time Frequency Analysis Toolbox is a MATLAB/Octave toolbox for computational time-frequency analysis. It is intended both as an educational and computational tool. The toolbox provides the basic Gabor, Wilson and MDCT transform along with routines for constructing windows (filter prototypes) and routines for manipulating coefficients. It also provides a bunch of demo scripts devoted either to demonstrating the main functions of the toolbox, or to exemplify their use in specific signal processing applications. In this paper we describe the used algorithms, their mathematical background as well as some signal processing applications.

Monaural spectral features are important for human sound-source localization in sagittal planes, including front-back discrimination and elevation perception. These directional features result from the acoustic filtering of incoming sounds by the listener's morphology and are described by listener-specific head-related transfer functions (HRTFs). This article proposes a probabilistic, functional model of sagittal-plane localization that is based on human listeners' HRTFs. The model approximates spectral auditory processing, accounts for acoustic and non-acoustic listener specificity, allows for predictions beyond the median plane, and directly predicts psychoacoustic measures of localization performance. The predictive power of the listener-specific modeling approach was verified under various experimental conditions: The model predicted effects on localization performance of band limitation, spectral warping, non-individualized HRTFs, spectral resolution, spectral ripples, and high-frequency attenuation in speech. The functionalities of vital model components were evaluated and discussed in detail. Positive spectral gradient extraction, sensorimotor mapping, and binaural weighting of monaural spatial information were addressed in particular. Potential applications of the model include predictions of psychophysical effects, for instance, in the context of virtual acoustics or hearing assistive devices.

In Brief Objectives: To assess the sensitivity of two bilateral cochlear implant users to interaural level and time differences (ILDs and ITDs) for various signals presented through the auxiliary inputs of clinical sound processors that discard fine timing information and only preserve envelope cues. Design: In a lateralization discrimination experiment, the just noticeable difference (JND) for ILDs and envelope ITDs was measured by means of an adaptive 2-AFC method. Different stimuli were used, including click trains at varying repetition rates, a speech fragment, and noise bursts. For one cochlear implant listener and one stimulus, the sensitivity to envelope ITDs was also determined with the method of constant stimuli. The dependency of ILD-JNDs on the interaural place difference was studied with stimulation at single electrode pairs by using sinusoidal input signals in combination with appropriate single-channel processor fittings. In a lateralization position experiment, subjects were required to use a visual pointer on a computer screen to indicate in-the-head positions for blocks of stimuli containing either ILD or ITD cues. All stimuli were loudness balanced (before applying ILD) and fed directly into the auxiliary inputs of the BTE processors (TEMPO+, Med-El Corp.). The automatic gain control and the processors’ microphones were deactivated. Results: Both cochlear implant listeners were highly sensitive to ILDs in all broadband stimuli used; JNDs approached those of normal-hearing listeners. Pitch-matched single electrode pairs showed significantly lower ILD-JNDs than pitch-mismatched electrode pairs. Envelope ITD-JNDs of cochlear implant listeners obtained with the adaptive method were substantially higher and showed a higher test-retest variability than waveform ITD-JNDs of normal-hearing control listeners and envelope ITD-JNDs of normal-hearing listeners reported in the literature for comparable signals. The envelope ITD-JNDs for the click trains were significantly lower than for the speech token or the noise bursts. The best envelope ITD-JND measured was ca. 250 μs for the click train at 100 cycles per sec. Direct measurement of the psychometric function for envelope ITD by the method of constant stimuli showed discrimination above chance level down to 150 μs. The lateralization position experiment showed that both ILDs and envelope ITDs can lead to monotonic changes in lateral percept. Conclusions: The two cochlear implant users tested showed strong effects of ILDs in various broadband stimuli with respect to JNDs as well as lateralization position. The high dependency of ILD-JNDs on the interaural pitch difference suggests the potential importance of pitch-matched assignment of electrodes in the two ears by the speech processors. Envelope ITDs appear to be more ambiguous cues than ILDs, as reflected by the higher and more variable JNDs compared with normal-hearing listeners. The envelope ITD-JNDs of cochlear implant listeners depended on the stimulus. Interaural level and time differences (ILDs and ITDs) are important cues for the localization of sound sources and for the understanding of speech in noise. This study investigated the ability of bilateral cochlear implant (CI) listeners to lateralize on ILD and envelope ITD using various stimuli presented through the auxiliary inputs of clinical sound processors. The processing done by these sound processors preserves envelope cues but removes fine timing information. The results of a lateralization discrimination experiment using broadband stimuli revealed generally high sensitivity of CI listeners to ILDs but low sensitivity and high variability with respect to envelope ITDs compared with normal hearing listeners. The effect of interaural place difference on ILD sensitivity was studied with stimulation at single electrode pairs. Sensitivity was found to be significantly higher for pitch-matched electrode pairs than for pitch-mismatched pairs, emphasizing the potential importance of pitch-matched assignment of electrodes in the two ears by the speech processors. A lateralization position experiment showed that both ILDs and envelope ITDs can lead to monotonic changes in lateral percept.

Fusing probabilistic information is a fundamental task in signal and data processing with relevance to many fields of technology and science. In this work, we investigate the fusion of multiple probability density functions (pdfs) of a continuous random variable or vector. Although the case of continuous random variables and the problem of pdf fusion frequently arise in multisensor signal processing, statistical inference, and machine learning, a universally accepted method for pdf fusion does not exist. The diversity of approaches, perspectives, and solutions related to pdf fusion motivates a unified presentation of the theory and methodology of the field. We discuss three different approaches to fusing pdfs. In the axiomatic approach, the fusion rule is defined indirectly by a set of properties (axioms). In the optimization approach, it is the result of minimizing an objective function that involves an information-theoretic divergence or a distance measure. In the supra-Bayesian approach, the fusion center interprets the pdfs to be fused as random observations. Our work is partly a survey, reviewing in a structured and coherent fashion many of the concepts and methods that have been developed in the literature. In addition, we present new results for each of the three approaches. Our original contributions include new fusion rules, axioms, and axiomatic and optimization-based characterizations; a new formulation of supra-Bayesian fusion in terms of finite-dimensional parametrizations; and a study of supra-Bayesian fusion of posterior pdfs for linear Gaussian models.

Weighted and controlled frames have been introduced recently to improve the numerical efficiency of iterative algorithms for inverting the frame operator. In this paper, we develop systematically these notions, including their mutual relationship. We will show that controlled frames are equivalent to standard frames and so this concept gives a generalized way to check the frame condition, while offering a numerical advantage in the sense of preconditioning. Next, we investigate weighted frames, in particular their relation to controlled frames. We consider the special case of semi-normalized weights, where the concepts of weighted frames and standard frames are interchangeable. We also make the connection with frame multipliers. Finally, we analyze weighted frames numerically. First, we investigate three possibilities for finding weights in order to tighten a given frame, i.e. decrease the frame bound ratio. Then, we examine Gabor frames and how well the canonical dual of a weighted frame is approximated by the inversely weighted dual frame.

This paper presents a new family of zero-voltage switching phase-shift-modulated dc-dc converters for a solid oxide fuel cell application. The proposed converters use a new technique of adaptive energy storage to minimize auxiliary circulating currents for all line and load conditions, thereby maximizing efficiency. A thorough analysis of the proposed topology, including the effects of rectifier junction capacitance, is conducted; and a design procedure presented. Experimental results from a 300 W per bridge, 20 V/300 V prototype are presented to validate the analysis. When compared to a widely used reference topology, the proposed topologies exhibit an efficiency improvement of 7% at full-load.

The recently described anti-A2/RA33 autoantibodies occur in 20-40% of patients with RA, SLE, and mixed connective tissue disease (MCTD). They are directed to the A2 protein of the heterogeneous nuclear ribonucleoprotein complex (hnRNP-A2), an abundant nuclear protein associated with the spliceosome. The NH2-terminal half of the antigen contains two conserved RNA binding domains whereas its COOH-terminal part is extremely glycine-rich. The aim of this study was to characterize the autoepitopes of hnRNP-A2 and to investigate the effects of anti-A2/RA33 autoantibodies on possible functions of the antigen. Using bacterially expressed fragments, two major discontinuous epitopes were identified. One containing the complete second RNA binding domain was recognized by the majority of patients with RA and SLE but not by patients with MCTD. The second epitope contained sequences of both RNA binding domains and was preferentially targeted by patients with MCTD. When the RNA binding properties of the antigen were investigated, oligoribonucleotides containing the sequence motif r(UUAG) were found to bind to a site closely adjacent or overlapping with the epitope targeted by autoantibodies from patients with RA and SLE. Moreover, anti-A2/RA33 autoantibodies from patients with RA or SLE, but not from patients with MCTD, inhibited binding of RNA. Thus, anti-A2/RA33 autoantibodies recognize conformation-dependent epitopes located in a functionally important region of the antigen. Furthermore, the specific recognition of an epitope by MCTD patients may be used as another argument in favor of considering MCTD a distinct connective tissue disease.

In this paper we study a distributed weighted least-squares estimation problem for a large-scale system consisting of a network of interconnected sub-systems. Each sub-system is concerned with a subset of the unknown parameters and has a measurement linear in the unknown parameters with additive noise. The distributed estimation task is for each sub-system to compute the globally optimal estimate of its own parameters using its own measurement and information shared with the network through neighborhood communication. We first provide a fully distributed iterative algorithm to asymptotically compute the global optimal estimate. The convergence rate of the algorithm will be maximized using a scaling parameter and a preconditioning method. This algorithm works for a general network. For a network without loops, we also provide a different iterative algorithm to compute the global optimal estimate which converges in a finite number of steps. We include numerical experiments to illustrate the performances of the proposed methods.

Audio signal processing frequently requires time-frequency representations and in many applications, a non-linear spacing of frequency bands is preferable. This paper introduces a framework for efficient implementation of invertible signal transforms allowing for non-uniform frequency resolution. Non-uniformity in frequency is realized by applying nonstationary Gabor frames with adaptivity in the frequency domain. The realization of a perfectly invertible constant-Q transform is described in detail. To achieve real-time processing, independent of signal length, slice-wise processing of the full input signal is proposed and referred to as sliCQ transform. By applying frame theory and FFT-based processing, the presented approach overcomes computational inefficiency and lack of invertibility of classical constant-Q transform implementations. Numerical simulations evaluate the efficiency of the proposed algorithm and the method's applicability is illustrated by experiments on real-life audio signals .

The Auditory Modeling Toolbox (AMT) is a MATLAB/Octave toolbox for the development and application of computational auditory models with a particular focus on binaural hearing. The AMT aims for a consistent implementation of auditory models, well-structured in-code documentation, and inclusion of auditory data required to run the models. The motivation is to provide a toolbox able to reproduce the model predictions and allowing students and researchers to work with and to advance existing models. In the AMT, model implementations can be evaluated in two stages: by running so-called demonstrations, which are quick presentations of a model, and by starting so-called experiments aimed at reproducing results from the corresponding publications. Here, we describe the tools and mechanisms available within the framework of all AMT 1.x versions. The recently released AMT 1.1 includes over 60 models and is freely available as an open-source package from https://www.amtoolbox.org .

Sound externalization, or the perception that a sound source is outside of the head, is an intriguing phenomenon that has long interested psychoacousticians. While previous reviews are available, the past few decades have produced a substantial amount of new data.In this review, we aim to synthesize those data and to summarize advances in our understanding of the phenomenon. We also discuss issues related to the definition and measurement of sound externalization and describe quantitative approaches that have been taken to predict the outcomes of externalization experiments. Last, sound externalization is of practical importance for many kinds of hearing technologies. Here, we touch on two examples, discussing the role of sound externalization in augmented/virtual reality systems and bringing attention to the somewhat overlooked issue of sound externalization in wearers of hearing aids.

Bilateral cochlear implant (CI) listeners currently use stimulation strategies which encode interaural time differences (ITD) in the temporal envelope but which do not transmit ITD in the fine structure, due to the constant phase in the electric pulse train. To determine the utility of encoding ITD in the fine structure, ITD-based lateralization was investigated with four CI listeners and four normal hearing (NH) subjects listening to a simulation of electric stimulation. Lateralization discrimination was tested at different pulse rates for various combinations of independently controlled fine structure ITD and envelope ITD. Results for electric hearing show that the fine structure ITD had the strongest impact on lateralization at lower pulse rates, with significant effects for pulse rates up to 800 pulses per second. At higher pulse rates, lateralization discrimination depended solely on the envelope ITD. The data suggest that bilateral CI listeners benefit from transmitting fine structure ITD at lower pulse rates. However, there were strong interindividual differences: the better performing CI listeners performed comparably to the NH listeners.

The combination of ultrasound and microreactor is an emerging and promising area, but the report of designing high-power ultrasonic microreactor (USMR) is still limited. This work presents a robust, high-power and highly efficient USMR by directly coupling a microreactor plate with a Langevin-type transducer. The USMR is designed as a longitudinal half wavelength resonator, for which the antinode plane of the highest sound intensity is located at the microreactor. According to one dimension design theory, numerical simulation and impedance analysis, a USMR with a maximum power of 100 W and a resonance frequency of 20 kHz was built. The strong and uniform sound field in the USMR was then applied to intensify gas-liquid mass transfer of slug flow in a microfluidic channel. Non-inertial cavitation with multiple surface wave oscillation was excited on the slug bubbles, enhancing the overall mass transfer coefficient by 3.3-5.7 times.

Empirical studies in export performance research are characterized by vastly empiricist and eclectic research designs. Partial models—i.e., models that investigate selected determinants of export performance—are symptomatic of this stream of research. Structural modeling techniques allow for the analysis of complex relations, which integrate a vast number of success factors into an extensive model in order to avoid over-emphasis of single determinants and to conduct an integrative investigation of different bundles of variables. In this article, the cross-validation and further development of a structural model for the prediction of export performance that was recently conceived in Austria is outlined.

This study examined the sensitivity of four cochlear implant (CI) listeners to interaural time difference (ITD) in different portions of four-pulse sequences in lateralization discrimination. ITD was present either in all the pulses (referred to as condition Wave), the two middle pulses (Ongoing), the first pulse (Onset), the last pulse (Offset), or both the first and last pulse (Gating). All ITD conditions were tested at different pulse rates (100, 200, 400, and 800 pulses/s pps). Also, five normal hearing (NH) subjects were tested, listening to an acoustic simulation of CI stimulation. All CI and NH listeners were sensitive in condition Gating at all pulse rates for which they showed sensitivity in condition Wave. The sensitivity in condition Onset increased with the pulse rate for three CI listeners as well as for all NH listeners. The performance in condition Ongoing varied over the subjects. One CI listener showed sensitivity up to 800 pps, two up to 400 pps, and one at 100 pps only. The group of NH listeners showed sensitivity up to 200 pps. The result that CI listeners detect ITD from the middle pulses of short trains indicates the relevance of fine timing of stimulation pulses in lateralization and therefore in CI stimulation strategies.

We present an algorithm for removing time-frequency components, found by a standard Gabor transform, of a ldquoreal-worldrdquo sound while causing no audible difference to the original sound after resynthesis. Thus, this representation is made sparser. The selection of removable components is based on a simple model of simultaneous masking in the auditory system. Important goals were the applicability to any real-world music and speech sound, integrating mutual masking effects between time-frequency components, coping with the time-frequency spread of such an operation, and computational efficiency. The proposed algorithm first determines an estimation of the masked threshold within an analysis window. The masked threshold function is then shifted in level by an amount determined experimentally, and all components falling below this function (the irrelevance threshold) are removed. This shift gives a conservative way to deal with uncertainty effects resulting from removing time-frequency components and with inaccuracies in the masking model. The removal of components is described as an adaptive Gabor multiplier. Thirty-six normal hearing subjects participated in an experiment to determine the maximum shift value for which they could not discriminate the irrelevance filtered signal from the original signal. On average across the test stimuli, 32 percent of the time-frequency components fell below the irrelevance threshold.