Acoustic Approach Research Articles

Short Utterance Speaker Recognition Based on Speech High Frequency Information Compensation and Dynamic Feature Enhancement Methods

This work aims to further compensate for the weaknesses of feature sparsity and insufficient discriminative acoustic features in existing short-duration speaker recognition. To address this issue, we propose the Bark-scaled Gauss and the linear filter bank superposition cepstral coefficients (BGLCC), and the multidimensional central difference (MDCD) acoustic feature extracted method. The Bark-scaled Gauss filter bank focuses on low-frequency information, while linear filtering is uniformly distributed, therefore, the filter superposition can obtain more discriminative and richer acoustic features of short-duration audio signals. In addition, the multi-dimensional central difference method captures better dynamics features of speakers for improving the performance of short utterance speaker verification. Extensive experiments are conducted on short-duration text-independent speaker verification datasets generated from the VoxCeleb, SITW, and NIST SRE corpora, respectively, which contain speech samples of diverse lengths, and different scenarios. The results demonstrate that the proposed method outperforms the existing acoustic feature extraction approach by at least 10% in the test set. The ablation experiments further illustrate that our proposed approaches can achieve substantial improvement over prior methods.

Investigation of navigation information correction techniques for master-slave AUV formations in unstable communication environments

Cooperative navigation within master–slave Autonomous Underwater Vehicles (AUVs) necessitates the periodic transmission of high-precision positional data from the master AUV, endowed with advanced navigational instrumentation, to the slave AUV characterized by limited positional accuracy. This data exchange is facilitated through an acoustic sensor mechanism, enabling the quantification of relative spatial separation between the two vehicles. Employing an Extended Kalman Filter (EKF), the positioning error of the slave AUV can be estimated and corrected, enhancing its positional accuracy. However, this collaborative navigation methodology encounters dual challenges. Firstly, the intricate underwater acoustic environment introduces the potential for data loss or outlier measurements in submarine acoustic communication, thereby jeopardizing the efficacy of collaborative navigation. Secondly, the current approach solely rectifies the slave AUV's positional error upon the receipt of data from the master AUV, neglecting the fundamental divergence in positional error that accrues between successive data receptions, stemming from the inherent imprecision in the dead reckoning process. To contend with these complexities, this study presents a novel approach to rectifying navigation data within master–slave AUV formations, particularly within the context of unstable communication conditions. The proposed method harnesses an outlier adaptive discriminant EKF framework, proficiently addressing the slave AUV's positional error while simultaneously estimating and compensating for azimuthal angular errors during underwater acoustic communication between the master and slave AUVs. This method establishes a pivotal groundwork for heightening the accuracy of independent autonomous navigation within the slave AUV during subsequent operational phases. The accuracy and applicability of the proposed method are rigorously validated through simulation-based verification and lake-water field experiments. The empirical results unequivocally highlight the efficacy of the underwater acoustic communication approach put forth in this study, with an outlier recognition rate exceeding 80%. Furthermore, the method adeptly rectifies azimuthal angular errors and positional inaccuracies within the AUV, thus effecting a comprehensive enhancement of overall positioning precision. This research bears significant implications for advancing autonomous underwater navigation, providing a robust framework to bolster the precision and reliability of master–slave AUV formations, particularly under challenging communication circumstances.

Computational optimization of trailing-edge designs to reduce airfoil self-noise

This paper presents an investigation and optimization of Trailing Edge (TE) design to reduce airfoil self-noise using Computational Fluid Dynamics (CFD). The case for study is a NACA0012 airfoil with a chord length (C) of 0.2286 m, a varying Angle of Attack (AoA) between 0° and 15°, and free stream velocity between 35 and 70 m/s. The flow domain consists of a c-type domain with a length and height of 18C and 9C, respectively. The parametric mesh maintains a structured mesh on the entire domain for different designs and TE shapes. Simulations employ a hybrid Stress-Blended Embedded Large-Eddy Simulations (SB-ELES) model to calculate the flow properties. Different turbulence models are tested to address their performance in determining pressure fluctuations. A correlation length also accounts for spanwise effects in the Ffowcs-Williams and Hawkings (FW-H) acoustic analogy approach to forecasting the far-field noise. Furthermore, multi-objective optimization is employed to determine the optimum airfoil TE configuration for different flow velocities and AoAs. The optimum designs generate the lowest Sound Pressure Level (SPL) without significantly sacrificing the aerodynamic performance of the airfoil within specified parameters.

Knowledge enhancement for speech emotion recognition via multi-level acoustic feature

Speech emotion recognition (SER) has become an increasingly attractive machine learning task for domain applications. It aims to improve the discriminative capacity of speech emotion utilising a certain type of features (e.g. MFCC, Spectrograms, Wav2vec2) or multi-type combination features. However, the potential of acoustic-related deep features is frequently overlooked in existing approaches that rely solely on a single type of feature or employ a basic combination of multiple feature types. To address this challenge, a multi-level acoustic feature cross-fusion approach is proposed, aiming to compensate for missing information between various features. It helps to enhance the SER performance by integrating different types of knowledge through the cross-fusion mechanism. Moreover, multi-task learning is utilised to share useful information through gender recognition, which can also obtain multiple common representations in a fine-grained space. Experimental results show that the fusion approach can capture the inner connections between multilevel acoustic features to refine the knowledge. The SOTA results were obtained under the same experimental conditions.

A Borehole Acoustic Calculation Approach with Gas Hydrate Saturation Inversion in Gas Hydrate-Bearing Sediments

The inversion of gas hydrate saturation is a critical procedure in the evaluation of hydrate reservoirs. In this paper, a theoretical model for a borehole acoustic wavefield excited by multipole sources is established for the first time. On this basis, the attenuation of the dipole flexural waves is obtained, and in combination with the results of sensitivity analysis, an approach for inverting natural gas hydrates using the attenuation characteristics of the dipole flexural wave is proposed. The results of the sensitivity analysis demonstrate that the attenuation of the dipole flexural wave is sensitive to gas hydrate saturation. Numerical results derived from synthetic logging data are provided to illustrate the viability of the inversion of gas hydrate saturation. Even when significant noise is introduced into the receiver signal arrays, the inversion method remains stable and accurately assesses gas hydrate saturation. The correctness and effectiveness of the proposed approach are substantiated through the processing of numerical simulation data. This work provides a potent processing approach for evaluating reservoir hydrate saturation utilizing acoustic well-logging information.

A novel non-destructive acoustic approach for investigating pool boiling phenomena

Analyzing boiling phenomena and performance within invisible boiling systems, where the interior cannot be observed, has traditionally been challenging. In this study, we propose a non-destructive approach using acoustic analysis method to figure out the boiling dynamics or performance, non-invasively. In pool boiling experiment, the acoustic emission (AE) sensor captured the emitted signals at the outside of the heater. Our analysis focused on two key aspects: single bubble characterization and multi-boiling phenomena. For single bubble, we observed that the AE signals originated from the pressure waves generated at the solid-vapor interface of the bubbles. Utilizing a finite difference method, the bubble radius was reverse calculated from the AE signals, enabling predictions up to a certain duration. Frequency analysis revealed a direct correlation between the signal frequency and the oscillation of the bubble radius. In multi-boiling phenomena, as heat flux increases, the number and intensity of AE signals rise, indicating more frequent and rapid bubble generation. A correlation was established between the AE signal counts in the 40 kHz range and the heat transfer coefficient (HTC). The AE count increased linearly with heat flux until it saturated near the critical heat flux. This correlation provided the non-destructive method to estimate the boiling HTC based on AE measurements. In conclusion, the findings offer insights into bubble dynamics, heat transfer characteristics, and the potential application for HTC estimation.

Control of Intrinsic ThermoAcoustic instabilities for methane and hydrogen air flames: DNS and network analysis

Replacing methane by hydrogen in premixed burners is a typical decarbonation scenario. This is usually done by ensuring a constant flame speed, i.e. by using ultra-lean H2 mixtures. This change can modify the thermoacoustic properties of burners. This study focuses on one class of thermoacoustic instabilities, Intrinsic ThermoAcoustic (ITA) modes, for flames stabilized on a diaphragm. An acoustic network approach is built to construct an ITA stability criterion (called ITAS criteria) for a three-duct configuration and a fully analytical expression is obtained in the limit of thin diaphragms. It shows that instability will grow when the Flame Transfer Function (FTF) gain n−π, located at the frequency at which u′-q′ phase ϕ=−π, is greater than a critical gain nc expressed by the ITAS criteria. Two-dimensional DNS, with anechoic terminations, of methane–air and hydrogen–air flames stabilized on a diaphragm are simulated with similar laminar flame speed sL0 and theoretical flame length, with an equivalence ratio of 0.73 and 0.4 respectively. The theory developed allows to analyze the differences with these two strong unstable ITA modes. For ultra-lean hydrogen flame, preferential diffusion near the diaphragm lips and tip opening reduces the flame time delay and shifts the ITA frequency towards higher values compared to methane. The assessment of two new control strategies is possible with the help of the analytical network model by increasing nc. A first strategy involves the use of preheated leaner mixtures: increasing Tu and reducing ϕ while keeping the flame speed sL0 constant. A second strategy consists of modifying the combustor geometry by changing the section ratio S2/S0, between plenum and combustion chamber. Both strategies successfully mitigate the hydrogen–air ITA modes and seems promising strategies for methane cases despite the fact that instabilities are not fully mitigated for this fuel at the conditions of interest in this study.

Active control of target sound fields using structural-acoustic brightness applied to a ship model's acoustic signature.

This manuscript presents a method for reproducing sound fields actively by using a vibrating submerged structure as the field reproduction source, with the target sound field to be reproduced defined in the frequency domain using the acoustic brightness approach. To balance the predetermination of a mono- or multi-zone target sound pressure field and the control effort required, singular value decomposition of the structural-acoustic system matrix is proposed. The dependency of the radiation efficiency into the target zone on the singular modes representing source and pressure modes is investigated using a wavenumber-domain approach. Furthermore, a feedforward control principle is adopted for adaptive sound-field reproduction with mode matching from the least squares perspective. Finally, an experiment is reported that involve synthesizing a tonal target underwater acoustic signature of a model of a fast attack craft (scale 1:8) at a measurement facility at Lake Plön in Germany. The results show that with the structural-acoustic brightness approach structural modes with radiation coupling into the target zone are excited and related pressure modes exhibit individual focus in the direction of hydrophones in use. Finally, a predetermined narrowband sound pressure field is actively reproduced at the hydrophone positions using inertial actuators and accelerometers on the ship model's hull.

Classification of dementia from spoken speech using feature selection and the bag of acoustic words model

<p>Memory disorders and dementia are a central factor in the decline of functioning and daily activities in older individuals. The workload related to standardized speech tests in clinical settings has led to a growing emphasis on developing automatic machine learning techniques for analyzing naturally spoken speech. This study presented a bag of acoustic words approach for distinguishing dementia patients from control individuals based on audio speech recordings. In this approach, each individual's speech was segmented into voiced periods, and these segments were characterized by acoustic features using the open-source openSMILE library. Word histogram representations were formed from the characterized speech segments of each speaker, which were used for classifying subjects. The formation of word histograms involved a clustering phase where feature vectors were quantized. It is well-known that partitional clustering involves instability in clustering results due to the selection of starting points, which can cause variability in classification outcomes. This study aimed to address instability by utilizing robust K-spatial-medians clustering, efficient K-means$ ++ $ clustering initialization, and selecting the smallest clustering error from repeated clusterings. Additionally, the study employed feature selection based on the Wilcoxon signed-rank test to achieve computational efficiency in the methods. The results showed that it is possible to achieve a consistent 75% classification accuracy using only twenty-five features, both with the external ADReSS 2020 test data and through leave-one-subject-out cross-validation of the entire dataset. The results rank at the top compared to international research, where the same dataset and only acoustic features have been used to diagnose patients.</p>

Acoustic Attack Mitigation Approach for MEMS Inertial Sensors Using Change Point Detection on MhIMU Framework

Acoustic Attack Mitigation Approach for MEMS Inertial Sensors Using Change Point Detection on MhIMU Framework

Measurement of jet gas–liquid interface fluctuations based on ultrasonic scattering

Aero-engines and rocket engines regularly experience unstable combustion phenomena. In order to reveal the complex mechanism, it is necessary to measure the jet gas–liquid interface disturbances. However, most techniques require optical access and the measurement fails when the optical access is limited. Ultrasonic method can be considered as an alternative under this condition. The present work proposes an acoustic approach for measuring the jet gas–liquid interface based on ultrasonic scattering. A thorough investigation of the scattered acoustic field by the jet is conducted by experimental methods. The high-speed imaging technology is used to verify the ultrasonic measurement results. The direct measurement results demonstrate a relationship between the perturbation of the jet gas–liquid interface and that of ultrasonic scattered acoustic pressure in certain measuring direction. The scattered acoustic field is analyzed theoretically and practically by using customized metal scatters of certain size. Based on these findings, a database is created to enable the ultrasonic measurements to be calibrated. The corrected result shows that the correlation of ultrasonic measurements and results from cameras has been greatly improved, and the maximum relative error of the ultrasonic measurements is 30.9%, the average relative measurement error is 2.1%. It is proved that the method of determining the gas–liquid interface of jet by ultrasonic scattering wave is feasible. The method may also be used for the measurement of the overall jet fluctuations and breakup.

The application of the coupled acoustic–structural approach (CASA) method on the free vibration of submerged structures

Traditionally, the consideration of the surrounding water during the modal analysis of submerged structures has been done by applying constant added mass factors. This method may lose its accuracy for complicated structures, especially those with bends and multi-plane topology. The present study utilizes the coupled acoustic structural approach (CASA) to obtain natural frequencies on submerged structures. The work compares the simulated CASA results with experimental results. Four experiments with objects with different geometry were chosen; a straight pipe, cantilever plates, stiffened cylindrical shell, and a disk. We found that the CASA method produced accurate results for the three first eigenmodes compared to the experimental results. It could simulate bending, twisting, ovality, and stretching mode shapes with an accuracy of less than 10 % difference. Additionally, we applied the CASA method to two submerged jumper structures for the first time. We found that the CASA method gives accurate modal analysis results and is recommended for use.

Deep learning and acoustic approach for mechanical failure detection in industrial machinery

Research into acoustic signal-based failure detection has developed into a subject that has attracted the attention of many researchers in recent years. Acoustic signal data collection can be performed without having to interrupt or stop the operation of the machine to be inspected. Therefore, it is very beneficial for the development of nondestructive testing and predictive maintenance. In this study, a collection of pump sound recordings that are part of the Malfunctioning Industrial Machine Investigation and Inspection dataset, known as the MIMII dataset, is used as test material. Several deep learning algorithms such as long short-term memory (LSTM), gate recurrent unit (GRU), autoencoder, and convolutional neural network (CNN) were involved and compared to determine their ability to detect failures. Based on the training results with 300 epochs and a learning rate of 10−6 it was found that CNN produced the classification with the highest accuracy compared to the other algorithms. In addition, the CNN algorithm is also capable of performing classification amidst the problem of imbalance in the amount of data.

Feature-Based Place Recognition Using Forward-Looking Sonar

Some structures in the harbour environment need to be inspected regularly. However, these scenarios present a major challenge for the accurate estimation of a vehicle’s position and subsequent recognition of similar images. In these scenarios, visibility can be poor, making place recognition a difficult task as the visual appearance of a local feature can be compromised. Under these operating conditions, imaging sonars are a promising solution. The quality of the captured images is affected by some factors but they do not suffer from haze, which is an advantage. Therefore, a purely acoustic approach for unsupervised recognition of similar images based on forward-looking sonar (FLS) data is proposed to solve the perception problems in harbour facilities. To simplify the variation of environment parameters and sensor configurations, and given the need for online data for these applications, a harbour scenario was recreated using the Stonefish simulator. Therefore, experiments were conducted with preconfigured user trajectories to simulate inspections in the vicinity of structures. The place recognition approach performs better than the results obtained from optical images. The proposed method provides a good compromise in terms of distinctiveness, achieving 87.5% recall considering appropriate constraints and assumptions for this task given its impact on navigation success. That is, it is based on a similarity threshold of 0.3 and 12 consistent features to consider only effective loops. The behaviour of FLS is the same regardless of the environment conditions and thus this work opens new horizons for the use of these sensors as a great aid for underwater perception, namely, to avoid degradation of navigation performance in muddy conditions.

Localization of mechanical and electrical defects in dry-type transformers using an optimized acoustic imaging approach.

This paper presents an acoustic imaging localization system designed to pinpoint common defects in dry-type transformers by analyzing the unique sounds they produce during operation. The system includes an optimized microphone array and an improved multiple signal classification algorithm. Sound signal characteristics of typical defects, such as foreign object intrusion, screw loosening, and partial discharge, are investigated. A 64-element, 8-arm spiral microphone array is designed using a particle swarm optimization algorithm. The multiple signal classification algorithm enhances acoustic imaging quality in field environments by transforming the input from time-domain to preprocessed frequency-domain signals. The power spectra of subarray and main array are combined, forming the optimization algorithm's output. Experimental results demonstrate the system's effectiveness and accuracy.

Stochastic Noise Sources for Computational Aeroacoustics of a Vehicle Side Mirror

<div>The broadband aeroacoustics of a side mirror is investigated with a stochastic noise source method and compared to scale-resolving simulations. The setup based on an already existing work includes two geometrical variants with a plain series side mirror and a modified mirror with a forward-facing step mounted on the inner side. The aeroacoustic near- and farfield is computed by a hydrodynamic–acoustic splitting approach by means of a perturbed convective wave equation. Aeroacoustic source terms are computed by the Fast Random Particle-Mesh method, a stochastic noise source method modeling velocity fluctuations in time domain based on time-averaged turbulence statistics. Three RANS models are used to provide input data for the Fast Random Particle-Mesh method with fundamental differences in local flow phenomena. Results of aeroacoustics simulations excited by the Fast Random Particle-Mesh method based on well-matching RANS data are in good agreement to the scale-resolving simulations in the integral acoustic Delta on the side window induced by the different side mirror geometries. For relative levels in between the variations, the robustness of the Fast Random Particle-Mesh method can be shown with secondary influences on the choice of the integral length scale. Absolute levels are only achieved with an adaptation of the length scale from literature. Two different RANS models with a missing separation bubble on the mirror or an overestimated wake flow show a good agreement with the plain series side mirror. However, they fail at computing the Delta to the step variant due to the missing amplification of the local turbulent kinetic energy interacting with the step and downstream mirror surfaces. Computational aeroacoustics simulations excited by the Fast Random Particle-Mesh method method based on RANS data only needs 14% of the computational effort compared to the conventional hybrid RANS-LES approach. This reveals its enormous potential for aeroacoustic broadband noise optimization purposes.</div>

Investigation on the behavior of flow and aerodynamic noise generated around the tandem seal-vibrissa-shaped cylinder

Through comparing with the tandem circular and elliptic cylinders with the same characteristic dimensions, the behavior of flow and flow-induced noise generated around the tandem seal-vibrissa-shaped cylinder is studied based on delayed detached-eddy simulation model and acoustic analogy approach. The co-shedding pattern of flow developed around the tandem cylindrical-like bars is investigated. The spatial modes, mode energy, and mode coefficients of turbulent flow around the geometries are analyzed through spectral proper orthogonal decomposition. Results show that the lift fluctuations of downstream bar are stronger than those of upstream bar, and more aerodynamic noise is radiated from the downstream bar than from the upstream bar. The alternative arrangement of nodal and saddle planes of seal-vibrissa-shaped cylinder introduces three-dimensional flow separations and suppresses the shear layer interactions, inhibiting the regular vortex shedding of Kármán vortex street occurring in the tandem cylinder wake. The reversed vortex shedding generated by two adjacent saddle surfaces in the wake of seal-vibrissa-shaped cylinder balances the lateral force and mitigates the lift fluctuations greatly, thereafter reduces the aerodynamic noise generated by wall pressure fluctuations introduced by unsteady fluctuating forces exerting on the surfaces of geometries. Compared to the tandem circular and elliptic cylinders, the good noise reduction effect with sound pressure level reduced at main frequency range has been achieved from the tandem seal-vibrissa-shaped cylinder. The calculated spectra and amplitude levels of aerodynamic noise agree well with the experimental measurements from the anechoic wind tunnel, verifying the accuracy of the numerical simulations.

Flow-induced noise from a seal-vibrissa-shaped cylinder

Inspired by the special structure of seal vibrissae and their ability to effectively suppress vortex-induced vibrations, this paper studies the characteristics of flow and aerodynamic noise generated around a seal-vibrissa-shaped cylinder at a Reynolds number (based on the hydrodynamic diameter and freestream velocity) of 6 × 104 using the delayed detached-eddy simulation model combined with the acoustic analogy approach. The far-field aerodynamic noise has also been measured at Reynolds numbers of 6 × 104∼1.2 × 105 in an anechoic wind tunnel to compare the seal-vibrissa-shaped cylinder with cylindrical and elliptical bars of the same characteristic dimensions. The calculated spectra of aerodynamic noise agree well with the wind tunnel measurements, verifying the accuracy of the numerical simulations. It is found that the three-dimensional flow separations introduced by the alternative saddle and nodal planes of the seal-vibrissa-shaped cylinder inhibit shear layer interactions and reduce the spanwise coherence, resulting in a wake with no dominant coherent structures and thus suppressing pressure fluctuations on the solid surfaces. The configuration of seal-vibrissa-shaped cylinder eliminates the regular Kármán vortex street which generally occurs in a cylinder wake, and consequently inhibits the tonal peak in the aerodynamic noise. The sound pressure level is also reduced in most frequencies. The consistency of the normalized spectra of the far-field aerodynamic noise and the vortex shedding frequency at different inflow speeds demonstrates the high accuracy of the anechoic wind tunnel measurements. Compared with cylinders with circular or elliptical cross sections, the seal-vibrissa-shaped cylinder effectively reduces the flow-induced noise by, respectively, 13 dB or 11 dB for a far-field receiver of 90° at Reynolds number of 6 × 104, showing great potential in reducing aerodynamic noise.

Host–Guest Interaction Studies of Isoproterenol Hydrochloride with Β-Cyclodextrin in an Aqueous Medium Probed by Volumetric, Acoustic, and Spectroscopic Approach

Host–Guest Interaction Studies of Isoproterenol Hydrochloride with Β-Cyclodextrin in an Aqueous Medium Probed by Volumetric, Acoustic, and Spectroscopic Approach

A Note on Performance Assessment of Signal Energy–Based Acoustic Source Localization in a Carbon Fiber–Reinforced Polymer Plate

Abstract The effectiveness of the signal energy–based acoustic source localization approach in practical applications has yet to be established. This is addressed herein by conducting an experimental study on a 500 mm × 500 mm carbon fiber–reinforced polymer plate and generating artificial acoustic events in the plate. Upon acquiring the propagating wave signals at several well-scattered sensors, the signal energy–based approach is applied, and the accuracy of the source localization results is noted. Seven experiments are performed with varying source locations, sensor-plate bonding, and excitation types. This approach has performed well for five experiments with source localization errors below 15 mm. The remaining two experiments where the acoustic sources are relatively close to the plate edges compared to the other experiments have, however, produced large localization errors, indicating a scope of improvement in the approach to encompass all situations.