Sound Signals Research Articles

Optimal design of spatial structures by a novel meta-heuristic algorithm: Sound energy optimizer

Sound signals are used by various species of living creaters to commonicate with each other. Scientific experiments have demonstrated that sound energy can travel through media and cause the receiver objects to vibrate or move. The present work simulates such a phenomenon in development of a new meta-heuristic algorithm called: Sound Energy Optimizer (SEO). Concerning physical rules, SEO leverages natural or ambient sounds for optimization. SEO utilizes a tunable virtual machine to control the decaying pattern of the sound energy in a medium representing the search space. Performance of the proposed algorithm is first evaluated on unconstrained optimization of some mathematical test functions. Then, a variety of pin-jointed structural problems are optimized including space tower, double-layer grid, barrel vault, helipad structure and power transmission towers. In this set, performance evaluation is done not only against the previous works but also by running SEO versus eight recent meta-heuristics under fair comparison conditions. New quantitative indices are introduced and traced as well as the common statistical measures and convergence curves. The results demonstrate competitive performance of SEO versus recent meta-heuristics in the discrete and continuous design of spatial structures.

Complex dynamic behavioral transitions in auditory neurons induced by chaotic activity

Chaotic sequences are widely used in secure communication due to their high randomness. Chaotic resonance (CR) refers to the resonant response of a system to weak signals induced by chaotic activity, but its practical application remains limited. This study designs a simplified FitzHugh-Nagumo (FHN) auditory neuron model by simulating the physiological activities of auditory neurons and considering the combined stimulation of chaotic activity and sound signals. It is found that the neuron dynamics depend on both external sound stimuli and chaotic current intensity. Chaotic currents induce spikes in the neuron output sequence through CR, and the spikes become more frequent with increasing current intensity, eventually leading to a chaotic state regardless of the initial state. However, the sensitivity of the initial value of this chaotic sequence shifts to the chaotic current excitation system. The injection of chaotic currents can reduce the system's average Hamiltonian energy under certain conditions. By measuring the complexity of the generated sequences, we find that the addition of chaotic currents can enhance the complexity of the original sequences, and the enhancement ability increases with the intensity. This provides a new approach to enhance the complexity of original chaotic sequences. Moreover, different chaotic currents can induce different chaotic sequences with varying abilities to enhance the complexity of the original sequences. We hope our work can contribute to secure communication.

A review on deep learning methods for heart sound signal analysis

IntroductionApplication of Deep Learning (DL) methods is being increasingly appreciated by researchers from the biomedical engineering domain in which heart sound analysis is an important topic of study. Diversity in methodology, results, and complexity causes uncertainties in obtaining a realistic picture of the methodological performance from the reported methods.MethodsThis survey paper provides the results of a broad retrospective study on the recent advances in heart sound analysis using DL methods. Representation of the results is performed according to both methodological and applicative taxonomies. The study method covers a wide span of related keywords using well-known search engines. Implementation of the observed methods along with the related results is pervasively represented and compared.Results and discussionIt is observed that convolutional neural networks and recurrent neural networks are the most commonly used ones for discriminating abnormal heart sounds and localization of heart sounds with 67.97% and 33.33% of the related papers, respectively. The convolutional neural network and the autoencoder network show a perfect accuracy of 100% in the case studies on the classification of abnormal from normal heart sounds. Nevertheless, this superiority against other methods with lower accuracy is not conclusive due to the inconsistency in evaluation.

Effect of human sound component on the sociability of urban public spaces—a case study in Sheffield, UK

AbstractSoundscape quality is important for fostering social interactions in urban public spaces. This study focuses on how and to what extent the compositions of the soundscape affect the sociability of urban public spaces where human sound is focused. Four 360° experimental soundscape samples were designed as foreground/background/no human sound type and signal sound type, based on the raw data collected in Peace Gardens, Sheffield, United Kingdom. Following Affordance Theory, sociability was analyzed by the experimental survey through two aspects—suitability and stimulation level—through a questionnaire consisting of Soundscape Quality evaluation (SQE) and Level of Willingness for social interactions (LoW), respectively. 120 participants were recruited in the same site representing the compositions of space users. Results show that only foreground human sound type significantly enhanced both suitability and stimulation levels, other samples were evaluated with either high suitability/low stimulation or low suitability/high stimulation. Results indicate people prefer a human sound dominated soundscape with high complexity and this preference was varied among different companion types and age groups. To achieve high sociability, it is suggested to create a complex and eventful soundscape within the range of pleasantness accompanied by a centripetal spatial form, and concerns users' diverse demographic backgrounds.

Neural Networks for Audio Classification: Multi-scale CNN-LSTM Approach to Animal Sound Recognition

Abstract. This research explores the application of neural networks, specifically CNN-LSTM models, for classifying sound signals from dogs, frogs, and cats, selected from the ESC-50 dataset. The sound data was preprocessed using Mel-frequency cepstral coefficients (MFCCs) and augmented through time stretching, pitch shifting, and noise addition to enhance model generalization in varied acoustic environments. We compared two deep learning models: a traditional CNN-LSTM and an improved version with multi-scale feature extraction, allowing for capturing both short-term and long-term sound patterns. Our findings show that the multi-scale CNN-LSTM architecture outperforms the traditional CNN-LSTM, achieving a test accuracy of 86.11% compared to 80.56%. These results highlight the effectiveness of multi-scale feature extraction for handling complex audio signals. This research offers valuable insights into bioacoustics and has broader applications in areas such as environmental sound monitoring, wildlife preservation, and animal behavior analysis.

Robust Biometric Verification Using Phonocardiogram Fingerprinting and a Multilayer-Perceptron-Based Classifier

Recently, a new set of biometric traits, called medical biometrics, have been explored for human identity verification. This study introduces a novel framework for recognizing human identity through heart sound signals, commonly referred to as phonocardiograms (PCGs). The framework is built on extracting and suitably processing Mel-Frequency Cepstral Coefficients (MFCCs) from PCGs and on a classifier based on a Multilayer Perceptron (MLP) network. A large dataset containing heart sounds acquired from 206 people has been used to perform the experiments. The classifier was tuned to obtain the same false positive and false negative misclassification rates (equal error rate: EER = FPR = FNR) on chunks of audio lasting 2 s. This target has been reached, splitting the dataset into 70% and 30% training and testing non-overlapped subsets, respectively. A recurrence filter has been applied to also improve the performance of the system in the presence of noisy recordings. After the application of the filter on chunks of audio signal lasting from 2 to 22 s, the performance of the system has been evaluated in terms of recall, specificity, precision, negative predictive value, accuracy, and F1-score. All the performance metrics are higher than 97.86% with the recurrence filter applied on a window lasting 22 s and in different noise conditions.

On the accuracy of distance estimates by propagation time of sound signals on the arctic shelf

As part of numerical modeling, estimates are made of the accuracy of determining the distance between underwater sources and sound receivers located at a distance of several kilometers from each other in the Kara Sea in the autumn. It is assumed that the main source of possible errors in determining the distance is the lack of accurate data on the vertical profile of the sound speed along the propagation path of acoustic signals. Data from September and November were analyzed, in the interval between which significant changes in the profile take place, when the vertical gradient of sound speed changes from negative to positive values. Characteristic values of sound speed variations were obtained by statistical processing of hydrological data taken from the World Ocean Database. The results obtained are important for analyzing the capabilities of underwater acoustic navigation.

Visually Guided Sound Source Separation With Audio-Visual Predictive Coding.

The framework of visually guided sound source separation generally consists of three parts: visual feature extraction, multimodal feature fusion, and sound signal processing. An ongoing trend in this field has been to tailor involved visual feature extractor for informative visual guidance and separately devise module for feature fusion, while utilizing U-Net by default for sound analysis. However, such a divide-and-conquer paradigm is parameter-inefficient and, meanwhile, may obtain suboptimal performance as jointly optimizing and harmonizing various model components is challengeable. By contrast, this article presents a novel approach, dubbed audio-visual predictive coding (AVPC), to tackle this task in a parameter-efficient and more effective manner. The network of AVPC features a simple ResNet-based video analysis network for deriving semantic visual features, and a predictive coding (PC)-based sound separation network that can extract audio features, fuse multimodal information, and predict sound separation masks in the same architecture. By iteratively minimizing the prediction error between features, AVPC integrates audio and visual information recursively, leading to progressively improved performance. In addition, we develop a valid self-supervised learning strategy for AVPC via copredicting two audio-visual representations of the same sound source. Extensive evaluations demonstrate that AVPC outperforms several baselines in separating musical instrument sounds, while reducing the model size significantly. Code is available at: https://github.com/zjsong/Audio-Visual-Predictive-Coding.

The Critical Thing about the Ear's Sensory Hair Cells.

The capabilities of the human ear are remarkable. We can normally detect acoustic stimuli down to a threshold sound-pressure level of 0 dB (decibels) at the entrance to the external ear, which elicits eardrum vibrations in the picometer range. From this threshold up to the onset of pain, 120 dB, our ears can encompass sounds that differ in power by a trillionfold. The comprehension of speech and enjoyment of music result from our ability to distinguish between tones that differ in frequency by only 0.2%. All these capabilities vanish upon damage to the ear's receptors, the mechanoreceptive sensory hair cells. Each cochlea, the auditory organ of the inner ear, contains some 16,000 such cells that are frequency-tuned between ∼20 Hz (cycles per second) and 20,000 Hz. Remarkably enough, hair cells do not simply capture sound energy: they can also exhibit an active process whereby sound signals are amplified, tuned, and scaled. This article describes the active process in detail and offers evidence that its striking features emerge from the operation of hair cells on the brink of an oscillatory instability-one example of the critical phenomena that are widespread in physics.

Heart sound classification algorithm based on bispectral feature extraction and convolutional neural networks

Cardiovascular disease (CVD) is one of the leading causes of death worldwide. Heart sound classification plays a key role in the early detection of CVD. The difference between normal and abnormal heart sounds is not obvious. In this paper, in order to improve the accuracy of the heart sound classification model, we propose a heart sound feature extraction method based on bispectral analysis and combine it with convolutional neural network (CNN) to classify heart sounds. The model can effectively suppress Gaussian noise by using bispectral analysis and can effectively extract the features of heart sound signals without relying on the accurate segmentation of heart sound signals. At the same time, the model combines with the strong classification performance of convolutional neural network and finally achieves the accurate classification of heart sound. According to the experimental results, the proposed algorithm achieves 0.910, 0.884 and 0.940 in terms of accuracy, sensitivity and specificity under the same data and experimental conditions, respectively. Compared with other heart sound classification algorithms, the proposed algorithm shows a significant improvement and strong robustness and generalization ability, so it is expected to be applied to the auxiliary detection of congenital heart disease.

Теорія фантомізації радіозображень РСА: базові відомості та обґрунтування розмірів ділянки фантомізації

The article's subject of study is the creation of the foundations of the theory of the formation of false elements in the images of aerospace radio vision systems by injecting informational signs of objects absent on the underlying surface. The goal is to substantiate the dimensions of the subsurface areas on which it is possible to simulate objects with different methods of viewing the exploring area from the aerospace radar carrier. Task: to develop the geometry of the problem, which will reflect the interaction between the radar with the synthesis of the Synthetic-aperture radar (SAR) and the radar of phantomization (FR); to develop a general structural diagram of the FR; determine the main parameters, taking into account which will allow planning an "interaction" session between the SAR and the FR; to determine the sizes of the underlying surface areas, where it is possible to provide phantomization of radio images formed by SAR with different methods of examination. The methods used are mathematical and functional analysis, navigation, and remote sensing. The following results were obtained. The geometry of the problem of the interaction between the SAR and the FR was defined; it allowed us to calculate the interaction time between the systems and to establish the law of change of the distance between them at any placement of the SAR in orbit. The influence of the interposition of the considered systems on the expected shape of the SAR sounding signal and the size of the area on which the problem of radio image phantomization arises has been investigated. Examples of calculations of the specified parameters for various types of exploring of the earth's surface by radar from an aerospace carrier are given. Conclusions. The article summarizes and develops the theory of phantom radio images constructed using SAR. The obtained results establishing the dependence between the phantomization geometry and such parameters as the potential and actual time of interaction between the systems, the envelope of the bundle of received radio pulses, and the number of separation bands by range are important for further research in this direction, which at the moment it is advisable to focus on solving the optimization problem phantomization of separate blocks of the radar and determining the conditions for planning the processes of injecting objects into the transmitter signal and into the radio image formed by the SAR.

A Framework for Detecting Pulmonary Diseases from Lung Sound Signals Using a Hybrid Multi-Task Autoencoder-SVM Model

Research focuses on the efficacy of Multi-Task Autoencoder (MTAE) models in signal classification due to their ability to handle many tasks while improving feature extraction. However, researchers have not thoroughly investigated the study of lung sounds (LSs) for pulmonary disease detection. This paper introduces a new framework that utilizes an MTAE model to detect lung diseases based on LS signals. The model integrates an autoencoder and a supervised classifier, simultaneously optimizing both classification accuracy and signal reconstruction. Furthermore, we propose a hybrid approach that combines an MTAE and a Support Vector Machine (MTAE-SVM) to enhance performance. We evaluated our model using LS signals from a publicly available database from King Abdullah University Hospital. The model attained an accuracy of 89.47% for four classes (normal, pneumonia, asthma, and chronic obstructive pulmonary disease) and 90.22% for three classes (normal, pneumonia, and asthma cases). Using the MTAE-SVM, the accuracy was further improved to 91.49% for four classes and 93.08% for three classes, respectively. The results indicate that the MTAE and MTAE-SVM have a considerable potential for detecting pulmonary diseases from lung sound signals. This could aid in the creation of more user-friendly and effective diagnostic tools.

A novel monitoring method based on the fusion of sound and image signals for laser welding penetration status

Accurately identifying penetration status is crucial for ensuring high-quality welding. However, relying on a single sensor to monitor the welding status falls short in accuracy. Thus, a novel monitoring method that fuses sound and image is proposed for assessing the laser welding penetration status. First, a microphone and a high-speed camera are employed to collect sound and image signals during welding. Second, to overcome noise challenges in sound signal, a denoising method is proposed that utilizes Grey Wolf Optimization (GWO) to find the optimal parameters for Variational Mode Decomposition (VMD). The time and frequency domain features from the denoised sound and Local Binary Pattern (LBP) features from the image are extracted. These features constitute a high-dimensional dataset, which is reduced using Principal Component Analysis (PCA). Finally, a back propagation neural network (BPNN) is constructed for fusion monitoring. The results demonstrate that fusion monitoring achieves higher accuracy (97.9%) compared to using sound (79.9%) or image (92.3%) features independently. In comparison to other methods like K-Nearest Neighbor (KNN), Naive Bayesian (NB), Support Vector Machine (SVM), and Decision Tree, BPNN fusion monitoring achieves higher accuracies, with increases of 18.09%, 12.79%, 6.53%, and 25.84% respectively. The proposed method helps to promote the development of welding intelligence.

The Relationship Between the Burpee Movement Program and Strength and Endurance Performance Measures in Active Young Adults: A Cross-Sectional Analysis.

This paper aimed to assess the motor performance in the Burpee Movement Program through the acceleration recorded by the Phyphox mobile app and define its relationship to strength and endurance parameters. Altogether, 15 students in physical education teaching completed the 3 × 3 min Burpee Movement Program, consisting of the repeated execution of a single burpee with maximum effort at regular intervals triggered by a sound signal. During the load phase, the intensity of the burpee and the fatigue index expressed in percentages were evaluated by means of the acceleration recorded through a mobile phone. In the second part of testing, we evaluated the performance parameters during a bench press and squat where the intensity was measured using a linear displacement transducer (Tendo Power Analyzer) and aerobic endurance was assessed with a 20 m shuttle run test (20 mSR). The average intensity of the burpee ranged from 3.12 to 11.12 ms-2. The fatigue index ranged from -21.95% (which represented an increase in performance) to 33.63% (which represented a decrease in performance). The performances in the bench presses ranged from 58 to 480 W and from 175 to 696 W during the squats. The distance in the 20 m shuttle run test (20 mSR) ranged from 540 to 2000 m. The intensity of the burpee showed a significant correlation to the performances achieved in the bench presses and squats r = 0.82 and 0.79. The fatigue index showed a significant correlation to the 20 m shuttle run test (20 mSR) r = -0.67. These findings indicate that in, our case, the results from the Burpee Movement Program are significantly associated with the participants' strength and endurance abilities. We recommend using BMP for the development of strength-endurance abilities, but further exploration is needed regarding the potential use of BMP as a diagnostic test.

Deep Learning-Based Gender Recognition in Cherry Valley Ducks Through Sound Analysis.

Gender recognition is an important part of the duck industry. Currently, the gender identification of ducks mainly relies on manual labor, which is highly labor-intensive. This study aims to propose a novel method for distinguishing between males and females based on the characteristic sound parameters for day-old ducks. The effective data from the sounds of day-old ducks were recorded and extracted using the endpoint detection method. The 12-dimensional Mel-frequency cepstral coefficients (MFCCs) with first-order and second-order difference coefficients in the effective sound signals of the ducks were calculated, and a total of 36-dimensional feature vectors were obtained. These data were used as input information to train three classification models, include a backpropagation neural network (BPNN), a deep neural network (DNN), and a convolutional neural network (CNN). The training results show that the accuracies of the BPNN, DNN, and CNN were 83.87%, 83.94%, and 84.15%, respectively, and that the three classification models could identify the sounds of male and female ducks. The prediction results showed that the prediction accuracies of the BPNN, DNN, and CNN were 93.33%, 91.67%, and 95.0%, respectively, which shows that the scheme for distinguishing between male and female ducks via sound had high accuracy. Moreover, the CNN demonstrated the best recognition effect. The method proposed in this study can provide some support for developing an efficient technique for gender identification in duck production.

Molecular specializations underlying phenotypic differences in inner ear hair cells of zebrafish and mice.

Hair cells (HCs) are the sensory receptors of the auditory and vestibular systems in the inner ears of vertebrates that selectively transduce mechanical stimuli into electrical activity. Although all HCs have the hallmark stereocilia bundle for mechanotransduction, HCs in non-mammals and mammals differ in their molecular specialization in the apical, basolateral, and synaptic membranes. HCs of non-mammals, such as zebrafish (zHCs), are electrically tuned to specific frequencies and possess an active process in the stereocilia bundle to amplify sound signals. Mammalian HCs, in contrast, are not electrically tuned and achieve amplification by somatic motility of outer HCs (OHCs). To understand the genetic mechanisms underlying differences between adult zebrafish and mammalian HCs, we compared their RNA-seq-characterized transcriptomes, focusing on protein-coding orthologous genes related to HC specialization. There was considerable shared expression of gene orthologs among the HCs, including those genes associated with mechanotransduction, ion transport/channels, and synaptic signaling. However, there were some notable differences in expression among zHCs, OHCs, and inner HCs (IHCs), which likely underlie the distinctive physiological properties of each cell type. For example, OHCs highly express Slc26a5 which encodes the motor protein prestin that contributes to OHC electromotility. However, zHCs have only weak expression of slc26a5, and subsequently showed no voltage-dependent electromotility when measured. Notably, the zHCs expressed more paralogous genes including those associated with HC-specific functions and transcriptional activity, though it is unknown whether they have functions similar to their mammalian counterparts. There was overlap in the expressed genes associated with a known hearing phenotype. Our analyses unveil substantial differences in gene expression patterns that may explain phenotypic specialization of zebrafish and mouse HCs. This dataset also includes several protein-coding genes to further the functional characterization of HCs and study of HC evolution from non-mammals to mammals.

Morten Qvenild – The HyPer(sonal) Piano Project

Towards a (per)sonal topography of grand piano and electronics How can I develop a grand piano with live electronics through iterated development loops in the cognitive technological environment of instrument, music, performance and my poetics? The instrument I am developing, a grand piano with electronic augmentations, is adapted to cater my poetics. This adaptation of the instrument will change the way I compose. The change of composition will change the music. The change of music will change my performances. The change in performative needs will change the instrument, because it needs to do different things. This change in the instrument will show me other poetics and change my ideas. The change of ideas demands another music and another instrument, because the instrument should cater to my poetics. And so it goes… These are the development loops I am talking about. I have made an augmented grand piano using various music technologies. I call the instrument the HyPer(sonal) Piano, a name derived from the suspected interagency between the extended instrument (HyPer), the personal (my poetics) and the sonal result (music and sound). I use old analogue guitar pedals and my own computer programming side by side, processing the original piano sound. I also take out control signals from the piano keys to drive different sound processes. The sound output of the instrument is deciding colors, patterns and density on a 1x3 meter LED light carpet attached to the grand piano. I sing, yet the sound of my voice is heavily processed, a processing decided by what I am playing on the keys. All sound sources and control signal sources are interconnected, allowing for complex and sometimes incomprehensible situations in the instrument´s mechanisms. Credits: First supervisor: Henrik Hellstenius Second Supervisors: Øyvind Brandtsegg and Eivind Buene Cover photo by Jørn Stenersen, www.anamorphiclofi.com All other photo, audio and video recording/editing by Morten Qvenild, unless stated.

Sound and vibration control crocodile hatching

Abstract The birth of a baby or the hatching of an egg are moments when the risk of predation is particularly high. In oviparous species where clutches contain many eggs, one strategy for limiting predation is based on mechanisms that synchronise hatching, which reduces the individual risk of being targeted and may allow hatchlings to emerge under parental protection. These mechanisms involve communication signals between eggs and between eggs and parents. While previous studies in various species have shown the importance of sound and vibration signals in these exchanges, how embryos integrate this information remains poorly understood. Here we show that crocodile embryos respond differently to calls emitted by siblings (hatching calls) and to vibrations mimicking parental intervention (rubbing and scratching the shell) or the presence of a predator, suggesting that these two stimuli carry different information. Through playback experiments, we confirm that hatching calls elicit vocal responses and synchronous hatching in mature embryos, while embryos hatch earlier when they perceive vibrations. Our study underlines that the control of hatching by external factors can rely on differential apprehension of multiple sensory inputs by embryos.

Computer vision based early fire-detection and firefighting mobile robots oriented for onsite construction

Fires are one of the most dangerous hazards and the leading cause of death in construction sites. This paper proposes a video-based firefighting mobile robot (FFMR), which is designed to patrol the desired territory and will constantly observe for fire-related events to make sure the camera without any occlusions. Once a fire is detected, the early warning system will send sound and light signals instantly and the FFMR moves to the right place to fight the fire source using the extinguisher. To improve the accuracy and speed of fire detection, an improved YOLOv3-Tiny (namely as YOLOv3-Tiny-S) model is proposed by optimizing its network structure, introducing a Spatial Pyramid Pooling (SPP) module, and refining the multi-scale anchor mechanism. The experiments show the proposed YOLOv3-Tiny-S model based FFMR can detect a small fire target with relatively higher accuracy and faster speed under the occlusions by outdoor environment. The proposed FFMR can be helpful to disaster management systems, avoiding huge ecological and economic losses, as well as saving a lot of human lives.

Simulation and experimental research on the propagation mechanisms of acoustic energy inside the pipeline to detect the leakage

Leakage from pipelines has negative impacts on property and environment. The diameters of internal acoustic detectors in pipelines are small and not easily blocked, making it suitable for pipelines under various working conditions. Studying the characteristics of pipeline leakage sound signal and the propagation mechanisms within pipelines is crucial for timely detection of small leaks. In this study, a 3D simulation model of pipeline leakage was established, and a CFD/CAA hybrid calculation method was used to obtain effective leakage flow field information via LES. The equivalent sound sources in the flow field were extracted based on the Lighthill acoustic analogy theory. Determine the characteristic frequency range of the leakage sound signal was 1–2 kHz. This study founded that an increase in the leakage pressure and aperture size both lead to an enhancement in the leakage sound signal, and the propagation of leakage sound signal in pipelines is affected by background flow. An experimental pipeline leakage system was designed, and the accuracy of the simulation results was verified experimentally.