Speech enhancement is fundamental for various real time speech applications and it is a challenging task in the case of a single channel because practically only one data channel is available. We have proposed a supervised single channel speech enhancement algorithm in this paper based on a deep neural network (DNN) and less aggressive Wiener filtering as additional DNN layer. During the training stage the network learns and predicts the magnitude spectrums of the clean and noise signals from input noisy speech acoustic features. Relative spectral transform-perceptual linear prediction (RASTA-PLP) is used in the proposed method to extract the acoustic features at the frame level. Autoregressive moving average (ARMA) filter is applied to smooth the temporal curves of extracted features. The trained network predicts the coefficients to construct a ratio mask based on mean square error (MSE) objective cost function. The less aggressive Wiener filter is placed as an additional layer on the top of a DNN to produce an enhanced magnitude spectrum. Finally, the noisy speech phase is used to reconstruct the enhanced speech. The experimental results demonstrate that the proposed DNN framework with less aggressive Wiener filtering outperforms the competing speech enhancement methods in terms of the speech quality and intelligibility.
If we want to provide the efficient training intervention to increase the duration of using hearing protection devices (HPDs) by workers, we need a tool that can estimate the person’s hearing threshold taking into account noise exposure level, age, and work history, and compare them with audiometry to find out the percent reduction of workers hearing loss. First, the workers noise exposure level was determined according to ISO 9612, then 4000 Hz audiometry was done to find age and work history. On basis of ISO 1999 the hearing threshold was estimated and if the hearing protection device was not used continuously and correctly, the hearing protection device’s actual performance was reduced adjusted with person’s audiometry. After training intervention, the estimate was done again and was compared with the adjusted audiometry. According to ISO 1999 standard estimation results, the percent reduction of the workers hearing loss level was 6.48 dB in intervention group. This level remained unchanged in control group. The mean score of hearing threshold estimation (standard ISO 1999) was statistically more significant than mean score of hearing threshold (p-value ¡ 0.001). The results show not significant change in control group due to lack of changing of noise exposure level. In regards to the results of hearing threshold estimation based on ISO 1999 and comparing with workers audiometry, it can be seen that BASNEF training intervention increases the duration of using the HPDs and it could be effective in reducing hearing threshold related to noise.
The paper investigates the interdependence between the perceptual identification of the vocalic quality of six isolated Polish vowels traditionally defined by the spectral envelope and the fundamental frequency F0. The stimuli used in the listening experiments were natural female and male voices, which were modified by changing the F0 values in the ±1 octave range. The results were then compared with the outcome of the experiments on fully synthetic voices. Despite the differences in the generation of the investigated stimuli and their technical quality, consistent results were obtained. They confirmed the findings that in the perceptual identification of vowels of key importance is not only the position of the formants on the F1 × F2 plane but also their relationship to F0, the connection between the formants and the harmonics and other factors. The paper presents, in quantitative terms, all possible kinds of perceptual shifts of Polish vowels from one phonetic category to another in the function of voice pitch. An additional perceptual experiment was also conducted to check a broader range of F0 changes and their impact on the identification of vowels in CVC (consonant, vowel, consonant) structures. A mismatch between the formants and the glottal tone value can lead to a change in phonetic category.
For the purpose of making of a solid body of an electric guitar the acoustic- and mechanical properties of walnut- (Juglans regia L.) and ash wood (Fraxinus excelsior L.) were researched. The acoustic properties were determined in a flexural vibration response of laboratory conditioned wood elements of 430 × 186 × 42.8 mm used for making of a solid body of an electric guitar. The velocity of shearand compression ultrasonic waves was additionally determined in parallel small oriented samples of 80 × 40 × 40 mm. The research confirmed better mechanical properties of ash wood, that is, the larger modulus of elasticity and shear modules in all anatomical directions and planes. The acoustic quality of ash wood was better only in the basic vibration mode. Walnut was, on the other hand, lighter and more homogenous and had lower acoustic- and mechanical anisotropy. Additionally, reduced damping of walnut at higher vibration modes is assumed to have a positive impact on the vibration response of future modelled and built solid bodies of electric guitars. When choosing walnut wood, better energy transfer is expected at a similar string playing frequency and a structure resonance of the electric guitar.
The study makes an attempt to model a complete vibrating guitar including its non-linear features, specifically the tension-compression of truss rod and tension of strings. The purpose of such a model is to examine the influence of design parameters on tone. Most experimental studies are flawed by uncertainties introduced by materials and assembly of an instrument. Since numerical modelling of instruments allows for deterministic control over design parameters, a detailed numerical model of folk guitar was analysed and an experimental study was performed in order to simulate the excitation and measurement of guitar vibration. The virtual guitar was set up like a real guitar in a series of geometrically non-linear analyses. Balancing of strings and truss rod tension resulted in a realistic initial state of deformation, which affected the subsequent spectral analyses carried out after dynamic simulations. Design parameters of the guitar were freely manipulated without introducing unwanted uncertainties typical for experimental studies. The study highlights the importance of acoustic medium in numerical models.
The pump performance and occurrence of cavitation directly depends on different operating conditions. To cover a wide range of operation conditions for detecting cavitation in this work, investigations on the effect of various suction valve openings on cavitation in the pump were carried out. In order to analyse various levels of cavitation in different operation conditions, the effect of the decrease in the inlet suction pressure of the centrifugal pump by controlling the inlet suction valve opening was investigated using this experimental setup. Hence, the acoustic and pressure signals under different inlet valve openings and different flow rates, namely, 103, 200, 302 l/min were collected for this purpose. A detailed analysis of the results obtained from the acoustic signal was carried out to predict cavitation in the pump under different operating conditions. Also, the acoustic signal was investigated in time domain through the use of the same statistical features. The FFT technique was used to analyse the acoustic signal in the frequency domain. In addition, in this work an attempt was made to find a relationship between the cavitation and noise characteristics using the acoustic technique for identifying cavitation within a pump.
A hybrid artificial boundary condition (HABC) that combines the volume-based acoustic damping layer (ADL) and the local face-based characteristic boundary condition (CBC) is presented to enhance the absorption of acoustic waves near the computational boundaries. This method is applied to the prediction of aerodynamic noise from a circular cylinder immersed in uniform compressible viscous flow. Different ADLs are designed to assess their effectiveness whereby the effect of the mesh-stretch direction on wave absorption in the ADL is analysed. Large eddy simulation (LES) and FW-H acoustic analogy method are implemented to predict the far-field noise, and the sensitivities of each approach to the HABC are compared. In the LES computed propagation field of the fluctuation pressure and the frequency-domain results, the spurious reflections at edges are found to be significantly eliminated by the HABC through the effective dissipation of incident waves along the wave-front direction in the ADL. Thereby, the LES results are found to be in a good agreement with the acoustic pressure predicted using FW-H method, which is observed to be just affected slightly by reflected waves.
Combine harvesters are the source a large amount of noise in agriculture. Depending on different working conditions, the noise of such machines can have a significant effect on the hearing condition of drivers. Therefore, it is highly important to study the noise signals caused by these machines and find solutions for reducing the produced noise. The present study was carried out is order to obtain the fractal dimension (FD) of the noise signals in Sampo and John Deere combine harvesters in different operational conditions. The noise signals of the combines were recorded with different engine speeds, operational conditions, gear states, and locations. Four methods of direct estimations of the FD of the waveform in the time domain with three sliding windows with lengths of 50, 100, and 200 ms were employed. The results showed that the Fractal Dimension/Sound Pressure Level [dB] in John Deere and Sampo combines varied in the ranges of 1.44/96.8 to 1.57/103.2 and 1.23/92.3 to 1.51/104.1, respectively. The cabins of Sampo and John Deere combines reduced and enhanced these amounts, respectively. With an increase in the length of the sliding windows and the engine speed of the combines, the amount of FD increased. In other words, the size of the suitable window depends on the extraction method of calculating the FD. The results also showed that the type of the gearbox used in the combines could have a tangible effect on the trend of changes in the FD.
The increment in the number of automobiles and the densification of the city has increased noise pollution rates. In addition, the lack of regulation in Chile regarding the acoustic insulation of façades is a problem of a growing concern. The main objective of the present study was to obtain a model of the Sound Insulation of housing, façades, stratified in Santiago, Chile, based on constructive variables. It is expected to serve as a basis for one future regulation for acoustic façades of houses. In the present study, tests based on the international ISO 140-5 standard were carried out in situ. An estimation model of the Standardized Level Difference Dls,2m,nT,w + C, was obtained based on the opening/façade proportion, and the type of glass used for the windows.
In this paper, the applications of the multivariate data analysis and optimization on vibration signals from compressors have been tested on the assembly line to identify nonconforming products. The multivariate analysis has wide applicability in the optimization of weather forecasting, agricultural experiments, or, as in this case study, in quality control. The techniques of discriminant analysis and linear program were used to solve the problem. The acceleration and velocity signals used in this work were measured in twenty-five rotating compressors, of which eleven were classified as good baseline compressors and fourteen with manufacturing defects by the specialists in the final acoustic test of the production line. The results obtained with the discriminant analysis separated the conforming and nonconforming groups with a significance level of 0.01, which validated the proposed methodology.
A novel method of active noise control using adaptive radiation sound sources is investigated. A finite element model of a modal enclosed sound field is excited harmonically, representing a noise field in the low-frequency range. The control sources are comprised of elementary dipole sources for which the driving signals are adjusted by an optimization method. Two set-up cases of the proposed compound sources are investigated. The coupling of the control sources with the modal sound field is discussed. The simulated performance of the proposed method is compared with that of a system with distributed simple sources and the results show the effectiveness of the sources with adaptive radiation for active noise control in small enclosures.
The aim of this publication is to design a procedure for the synthesis of an IDT (interdigital transducer) with diluted electrodes. The paper deals with the surface acoustic waves (SAW) and the theory of synthesis of the asymmetrical delay line with the interdigital transducer with diluted electrodes. The authors developed a theory, design, and implementation of the proposed design. They also measured signals. The authors analysed acoustoelectronic components with SAW: PLF 13, PLR 40, delay line with PAV 44 PLO. The presented applications have a potential practical use.
Analytical relations, describing the electrical fields of cylindrical piezoceramic radiators with circular polarization as a member of the cylindrical systems with the baffle in the inner cavity, using the related fields method in multiply connected regions were obtained. Comparative analysis of the results of numerical experiments performed on the frequency characteristics of the electric field of the radiating systems for different modes of radiation allow to establish a number of subtle effects of the formation of the electric field of radiators.
In this paper, a new Multi-Layer Perceptron Neural Network (MLP NN) classifier is proposed for classifying sonar targets and non-targets from the acoustic backscattered signals. Besides the capabilities of MLP NNs, it uses Back Propagation (BP) and Gradient Descent (GD) for training; therefore, MLP NNs face with not only impertinent classification accuracy but also getting stuck in local minima as well as lowconvergence speed. To lift defections, this study uses Adaptive Best Mass Gravitational Search Algorithm (ABGSA) to train MLP NN. This algorithm develops marginal disadvantage of the GSA using the bestcollected masses within iterations and expediting exploitation phase. To test the proposed classifier, this algorithm along with the GSA, GD, GA, PSO and compound method (PSOGSA) via three datasets in various dimensions will be assessed. Assessed metrics include convergence speed, fail probability in local minimum and classification accuracy. Finally, as a practical application assumed network classifies sonar dataset. This dataset consists of the backscattered echoes from six different objects: four targets and two non-targets. Results indicate that the new classifier proposes better output in terms of aforementioned criteria than whole proposed benchmarks.
The article describes the method of controlling the recovered grade based on measuring the intensity of volume ultrasonic oscillations and Lamb waves covering a fixed distance through the test medium and on a metal plate contacting the test medium at various time points of deliberate motion of ground materials. The authors suggest a method of determining density of ground ore particles in the pulp periodically after isolating the pulp flow in the vertical part of the measuring vessel based on measuring attenuation change values in Lamb waves covering a fixed distance on a plate contacting the medium under study and high frequency volume ultrasonic oscillations that have come through it within a certain time period. There are given dependencies of amplitudes of measuring channels based on volume ultrasonic oscillations and surface Lamb waves, size distribution according to solid phase pulp particles for various types of ores under study, a set of curves for determining the recovered grade with regard to various types of ores under study.
The paper discusses acoustic problems in the contemporary Catholic church, and presents a study of the influence of the ceiling structure on acoustics in the interior for two types of ceiling structures, i.e. the truss type and the reinforced concrete one. The investigations involved six contemporary churches: three buildings with a truss type ceiling and three buildings with a reinforced concrete ceiling. The results reveal that in churches with a truss type ceiling, acoustic parameters reach values close to recommendations. In contrast, churches with a concrete ceiling create very unfavourable acoustic conditions. The investigations rendered it possible to calculate the sound absorption coefficient α for the truss type cover.
It has been shown in the present paper that exploitation of the experimental potential of a photoacoustic technique can provide information on a type of intermolecular interactions in aqueous mixtures containing organic liquids, when the basic parameters of these mixtures, such as density, ρ, specific heat, cp, or thermal conductivity, λ, are unknown. Earlier investigations of concentration dependence of effusivity in different aqueous solutions of organic liquids demonstrated that the photoacoustics method is a sensitive tool to identify hydrophobic properties of such liquids. In our experiment this suggestion was exploited for a solution of methanol which is known to display much weaker hydrophobicity than other alcohols. It was confirmed that the location of extreme deviations from linearity for the thermal effusivity, Δe, agrees well with that of characteristic points for the isentropic compressibility coefficient, κS, and the excess molar volume, V_m^E, as a function of the concentration.
Sub-bottom profiler (SBP) is an acoustic instrument commonly used to survey underwater shallow geological structure and embedded objects whose most important performance parameter is the actual vertical resolution. This paper presented a methodology to measure and evaluate the actual vertical resolution of SBP based on an experiment in an anechoic tank, which was divided into three components: building of artificial geological model, measurement of acoustic parameters, and determination of actual vertical resolution of the acoustic profiles. First, the wedge-shaped geological model, whose thickness could be accurately controlled, was designed and built in an anechoic tank to try to directly measure the vertical resolution of SBP. Then, the acoustic pulse width of SBP was measured to calculate the theoretical general vertical resolution and extreme vertical resolution. Finally, based on the acoustic profiles obtained in the experiment, the method which was used to evaluate the actual vertical resolution by measuring the duration of reflection event was put forward. Due to comparing measurement data of different parameter settings of the SBP, the study has revealed that the SBP had the lowest resolution in the 4 kHz–500 µs setting, which was 226.5 µs, or 36.2 cm, and the highest resolution in the 15 kHz–67 µs setting, which was 72.7 µs, or 11.6 cm. The vertical resolution decreased with the increase of the pulse width. The results also showed that the actual resolution was close to the theoretical general resolution and far from the extreme resolution.
Biography of Jozef John Zwislocki (March 22, 1922 – May 14, 2018) – Polish-born American neuroscientist. Granted fellowship into the Acoustical Society of America, as well as membership to the United States National Academy of Sciences, Polish Academy of Sciences, American Association for the Advancement of Science, and Association for Research in Otolaryngology (among others). He worked at Federal Institute of Technology in Zurich, Switzerland, taught at the University of Basel, was on a research fellowship at Harvard University, and was member of the Syracuse University faculty. Owner of twelve patents.