Types Of Sounds Research Articles

A Dependent Nominal Physical Type System for Static Analysis of Memory in Low Level Code

We tackle the problem of checking non-proof-carrying code , i.e. automatically proving type-safety (implying in our type system spatial memory safety) of low-level C code or of machine code resulting from its compilation without modification. This requires a precise static analysis that we obtain by having a type system which (i) is expressive enough to encode common low-level idioms, like pointer arithmetic, discriminating variants by bit-stealing on aligned pointers, storing the size and the base address of a buffer in distinct parts of the memory, or records with flexible array members, among others; and (ii) can be embedded in an abstract interpreter. We propose a new type system that meets these criteria. The distinguishing feature of this type system is a nominal organization of contiguous memory regions, which (i) allows nesting, concatenation, union, and sharing parameters between regions; (ii) induces a lattice over sets of addresses from the type definitions; and (iii) permits updates to memory cells that change their type without requiring one to control aliasing. We provide a semantic model for our type system, which enables us to derive sound type checking rules by abstract interpretation, then to integrate these rules as an abstract domain in a standard flow-sensitive static analysis. Our experiments on various challenging benchmarks show that semantic type-checking using this expressive type system generally succeeds in proving type safety and spatial memory safety of C and machine code programs without modification, using only user-provided function prototypes.

Annoyance evaluation of impact sounds on lightweight, concrete and cross-laminated timber floors - A comparative study

For people living in multi-unit residential buildings, neighbor noise and especially impact noise from neighbors, such as dropping objects or footsteps, can affect the quality of living of those who hear the noise. In order to support the inclusion of limits for impact noise in a future edition of the National Building Code of Canada, a project at the National Research Council of Canada has focused on the perceived annoyance from different types of impact sounds for various timber floor-ceiling assemblies. The results from listening experiments showed poor correlation between subjective ratings and objective single number metrics for impact noise due to persons walking without shoes. To extend this evaluation to a wider variety of floor-ceiling assembly types and to validate previous results, another listening experiment was conducted at the National Research Council. Annoyance ratings for ball drops, hammer impacts and footsteps on different lightweight, concrete and cross-laminated timber floor-ceiling assemblies were compared.

Recommendations on bioacoustical metrics relevant for regulating exposure to anthropogenic underwater sounda).

Metrics to be used in noise impact assessment must integrate the physical acoustic characteristics of the sound field with relevant biology of animals. Several metrics have been established to determine and regulate underwater noise exposure to aquatic fauna. However, recent advances in understanding cause-effect relationships indicate that additional metrics are needed to fully describe and quantify the impact of sound fields on aquatic fauna. Existing regulations have primarily focused on marine mammals and are based on the dichotomy of sound types as being either impulsive or non-impulsive. This classification of sound types, however, is overly simplistic and insufficient for adequate impact assessments of sound on animals. It is recommended that the definition of impulsiveness be refined by incorporating kurtosis as an additional parameter and applying an appropriate conversion factor. Auditory frequency weighting functions, which scale the importance of particular sound frequencies to account for an animal's sensitivity to those frequencies, should be applied. Minimum phase filters are recommended for calculating weighted sound pressure. Temporal observation windows should be reported as signal duration influences its detectability by animals. Acknowledging that auditory integration time differs across species and is frequency dependent, standardized temporal integration windows are proposed for various signal types.

Recognition of Western Black-Crested Gibbon Call Signatures Based on SA_DenseNet-LSTM-Attention Network

As part of the ecosystem, the western black-crested gibbon (Nomascus concolor) is important for ecological sustainability. Calls are an important means of communication for gibbons, so accurately recognizing and categorizing gibbon calls is important for their population monitoring and conservation. Since a large amount of sound data will be generated in the process of acoustic monitoring, it will take a lot of time to recognize the gibbon calls manually, so this paper proposes a western black-crested gibbon call recognition network based on SA_DenseNet-LSTM-Attention. First, to address the lack of datasets, this paper explores 10 different data extension methods to process all the datasets, and then converts all the sound data into Mel spectrograms for model input. After the test, it is concluded that WaveGAN audio data augmentation method obtains the highest accuracy in improving the classification accuracy of all models in the paper. Then, the method of fusion of DenseNet-extracted features and LSTM-extracted temporal features using PCA principal component analysis is proposed to address the problem of the low accuracy of call recognition, and finally, the SA_DenseNet-LSTM-Attention western black-crested gibbon call recognition network proposed in this paper is used for recognition training. In order to verify the effectiveness of the feature fusion method proposed in this paper, we classified 13 different types of sounds and compared several different networks, and finally, the accuracy of the VGG16 model improved by 2.0%, the accuracy of the Xception model improved by 1.8%, the accuracy of the MobileNet model improved by 2.5%, and the accuracy of the DenseNet network model improved by 2.3%. Compared to other classical chirp recognition networks, our proposed network obtained the highest accuracy of 98.2%, and the convergence of our model is better than all the compared models. Our experiments have demonstrated that the deep learning-based call recognition method can provide better technical support for monitoring western black-crested gibbon populations.

Cross-referencing unidentified fish sound data sets to unravel sound sources: a case study from the Temperate Northern Atlantic

There is growing evidence that studying aquatic acoustic communities can provide ecologically relevant information. Understanding these communities may offer unique insights into species behaviour and ecology, while consolidating passive acoustic monitoring as a tool for mapping the presence of target species or estimating changes in aquatic biodiversity. Fish can be significant soundscape contributors, but most soniferous fish species are yet to be identified. Here, we crossed information of three key fish acoustic communities in the Lusitanian Province of the Temperate Northern Atlantic (the Madeira archipelago, the Azores archipelago and Arrábida in mainland Portugal) to unveil potential sources of unidentified fish sounds. We found that the three communities shared various sound types and we were able to narrow down the list of possible fish sound sources. Several sound types were suggested to be produced by species of the Pomacentridae, Scorpaenidae and Serranidae families. We also observed that the sound type /kwa/, associated with Scorpaena spp., exhibited more variations in the geographic area where more species of this genus are known to be present. This study showcases that, as databases of unidentified fish sounds continue to grow, future comparisons of multiple acoustic communities may provide insights into unknown fish sound sources and sound types.

Thermal–Acoustic Interaction Effects on Physiological and Psychological Measures in Urban Forests: A Laboratory Study

The environment in which people live is a complex system influenced by multiple factors interacting with each other, and therefore, it is crucial to deeply explore the influences of various factors on environmental perception. Among the numerous factors affecting the experience of urban forests visits, the thermal–acoustic environment stands out prominently. This study focuses on urban forests located in subtropical regions, with specific research conducted in the Xihu Park in Fuzhou, China. The study explores the thermal–acoustic interaction in urban forest environments. A total of 150 participants evaluated the perception of sound, thermal sensation, and overall perception through laboratory experiments, with 36 of them having their objective physiological indicators monitored. Different levels of sound and temperature were selected for the experiments, with three levels for each type of sound. Our results show that increasing temperature enhanced the perceived loudness of sound, especially when the environment was quiet. Sound type and loudness had a significant impact on thermal sensation, but no interaction was observed with temperature. Moreover, we found that certain sounds could improve overall comfort, and the effect was most evident at moderate loudness. Temperature had a significant influence on both comfort and annoyance, with increasing temperature leading to higher annoyance. These findings provide important insights into how the interplay between sound and heat affects human perception and emotional state, providing scientific guidance for the design of more human-centered environments.

Soundscape characteristics of RAS tanks holding Atlantic salmon (Salmo salar) during feeding and feed withdrawal

Behavioural monitoring can provide crucial information on welfare and feeding in aquaculture. Passive acoustic monitoring of behaviour can be particularly useful in recirculating aquaculture systems (RAS), as they often have turbid water that impairs visual monitoring. Currently, little is known about the sounds that make up the soundscapes in RAS tanks holding Atlantic salmon (Salmo salar). In this study, hydrophones were used to continuously record the soundscape in eight single tank RAS holding Atlantic salmon parr for 15 days, with the fish in four of the tanks being fasted by feed withdrawal for five days from the sixth to the tenth day. The results show that soundscapes in RAS tanks are affected by feeding. Two main sound sources were identified during feeding in RAS tanks, one related to pellets delivery and the other to fish behaviour. The sound of pellets hitting the water surface had energy concentrated at frequencies between 1.7 and 4.0 kHz, with peak frequency decreasing and amplitude increasing with increasing number of pellets hitting simultaneously. The feeding sounds of Atlantic salmon had energy concentrated at frequencies between 6.5 and 9.4 kHz.More complex soundscapes were recorded during feeding events. These were characterized by variations in amplitude and frequency that have been described by using acoustic indexes in RAS tanks for the very first time. The Acoustic Complexity Index (ACI), the Acoustic Entropy Index (H) and the Normalized Difference Soundscape Index (NDSI) showed distinct changes in the soundscape related to feeding events; ACI increased while H and NDSI decreased compared to the times in between scheduled feeding times. The sound types identified in this study and the outcomes of the acoustic indices indicate a possibility to monitor system performance as well as fish behaviour in the tank soundscapes in RAS. Soundscape monitoring can contribute to match feeding closer to fish appetite, improve water quality, and reduce risks that deviations in the system performance can have on fish welfare during production.

Broiler health monitoring technology based on sound features and random forest

The existing broiler health monitoring technology has problems such as low automation, unstable monitoring results, and low practical value, making it difficult to provide timely and reliable broiler health monitoring results. The broiler sound signal can provide feedback on their health. A widely validated and correct experience is to analyze the frequency of coughs in a segment of broiler sound signal to determine the health of the broiler group. Based on this, in this paper, the authors proposed a new broiler health monitoring technology based on sound detection. The broiler health monitoring problem is cleverly transformed into a multi-classification problem, which can be solved by identifying the sound types in broiler sound signals. Specifically, the audio signal collection system was designed to complete signal collection and preliminary signal filtering. Wiener filtering was used for deep signal filtering. The 60-dimensional sound features with good performance from three aspects, time-frequency domain, Mel-Frequency Cepstral Coefficients, and sparse representation were extracted, and a preliminary data set was created. Min-max normalization was used to align the numerical distribution of the data set, and a high-quality data set was created. Multi-classification models based on different classification algorithms and neural networks were trained, and the best-performing Random Forest was obtained, thus parameter optimization was carried out, and the optimal multi-classification model was obtained, achieving a classification accuracy of 91.14%. The visualization platform was built to process the classification results of the multi-classification model, completing majority voting processing and cough rate calculation, thereby achieving broiler health monitoring. In addition, the definitions of cough rate and prediction accuracy were newly proposed. A large number of experiments have verified the feasibility of the broiler health monitoring technology proposed in this paper, with an average prediction accuracy of 98.97% achieved.

Experimental study of sound wave propagation patterns

The present study compares the behavior of different sound types and their sources concerning distance. Experimental findings demonstrate a consistent reduction in noise levels with increasing distance from the sound origin, aligning with anticipated sound propagation patterns. Median noise level reductions are quantified, showing decreases from 72.7 dB at the source to 54.8 dB at a distance of 3 m. Pulsed sounds exhibit pronounced fluctuations and peaks at close distances, while steady and blended sounds maintain more uniform levels. An exponential model accurately characterizes the noise reduction phenomenon (R² = 0.8664), underscoring its applicability for construction noise control. These results offer valuable insights into sound propagation dynamics and provide a basis for developing effective noise control strategies.

Acoustic fish community in the Madeira Archipelago (North Atlantic Ocean): Characterization of sound diversity and daily patterns

Marine ecosystems are increasingly subjected to anthropogenic pressures, which demands urgent monitoring plans. Understanding soundscapes can offer unique insights into the ocean status providing important information and revealing different sounds and their sources. Fishes can be prominent soundscape contributors, making passive acoustic monitoring (PAM) a potential tool to detect the presence of vocal fish species and to monitor changes in biodiversity. The major goal of this research was to provide a first reference of the marine soundscapes of the Madeira Archipelago focusing on fish sounds, as a basis for a long-term PAM program. Based on the literature, 102 potentially vocal and 35 vocal fish species were identified. Additionally 43 putative fish sound types were detected in audio recordings from two marine protected areas (MPAs) in the Archipelago: the Garajau MPA and the Desertas MPA. The Garajau MPA exhibited higher fish vocal activity, a greater variety of putative fish sound types and higher fish sound diversity. Lower abundance of sounds was found at night at both MPAs. Acoustic activity revealed a clear distinction between diurnal and nocturnal fish groups and demonstrated daily patterns of fish sound activity, suggesting temporal and spectral partitioning of the acoustic space. Pomacentridae species were proposed as candidates for some of the dominant sound types detected during the day, while scorpionfishes (Scorpaena spp.) were proposed as sources for some of the dominant nocturnal fish sounds. This study provides an important baseline about this community acoustic behaviour and is a valuable steppingstone for future non-invasive and cost-effective monitoring programs in Madeira.

Observing System Simulation Experiments (OSSEs) in Support of Next-Generation NOAA Satellite Constellation

Abstract Between 2014 and 2018, the National Oceanic and Atmospheric Administration conducted the NOAA Satellite Observing System Architecture (NSOSA) study to plan for the next generation of operational environmental satellites. The study generated some important questions that could be addressed by observing system simulation experiments (OSSEs). This paper describes a series of OSSEs in which benefits to numerical weather prediction from existing observing systems are combined with enhancements from potential future capabilities. Assessments include the relative value of the quantity of different types of thermodynamic soundings for global numerical weather applications. We compare the relative impact of several sounding configuration scenarios for infrared (IR), microwave (MW), and radio occultation (RO) observing capabilities. The main results are 1) increasing the revisit rate for satellite radiance soundings produces the largest benefits but at a significant cost by requiring an increase in the number of polar-orbiting satellites from 2 to 12; 2) a large positive impact is found when the number of RO soundings per day is increased well beyond current values and other observations are held at current levels of performance; 3) RO can be used as a mitigation strategy for lower MW/IR sounding revisit rates, particularly in the tropics; and 4) smaller benefits result from increasing the horizontal resolution along the track of the satellites of MW/IR satellite radiances. Furthermore, disaggregating IR and MW instruments into six evenly distributed sun-synchronous orbits is slightly more beneficial than when the same instruments are combined and collocated on three separate orbits.

Perception of temporal structure in speech is influenced by body movement and individual beat perception ability.

The subjective experience of time flow in speech deviates from the sound acoustics in substantial ways. The present study focuses on the perceptual tendency to regularize time intervals found in speech but not in other types of sounds with a similar temporal structure. We investigate to what extent individual beat perception ability is responsible for perceptual regularization and if the effect can be eliminated through the involvement of body movement during listening. Participants performed a musical beat perception task and compared spoken sentences to their drumbeat-based versions either after passive listening or after listening and moving along with the beat of the sentences. The results show that the interval regularization prevails in listeners with a low beat perception ability performing a passive listening task and is eliminated in an active listening task involving body movement. Body movement also helped to promote a veridical percept of temporal structure in speech at the group level. We suggest that body movement engages an internal timekeeping mechanism, promoting the fidelity of auditory encoding even in sounds of high temporal complexity and irregularity such as natural speech.

Exploring the Multi-Sensory Coupling Relationship of Open Space on a Winter Campus

Exploring the combined effects of multisensory interactions in open spaces can help improve the comfort of campus environments. Nine typical spaces on a university campus in Fuzhou were selected for this study. Subjects perceived the environment and then completed an on-site subjective questionnaire. At the same time, meteorological data (global radiation, air temperature, globe temperature, wind speed, relative humidity, and illumination intensity) were measured to determine the interactions between visual and acoustic and thermal perceptions. Differences in the meteorological parameters between the measuring points were described using a one-way ANOVA and Tukey’s post hoc test, and a chi-square test of independence was used to determine significant associations between thermal, acoustic, and visual comfort, which in turn led to the study of interactions between visual, acoustic, and thermal comfort using a two-way ANOVA. The following conclusions were drawn: (1) the Thermal Comfort Vote (TCV) increased with the increasing Acoustic Comfort Vote (ACV) at all levels of thermal stress. (2) The highest and lowest Acoustic Sensation Vote (ASV) values for each sound type were derived from either “slightly cold” or “warm” conditions. Both the Thermal Comfort Vote (TCV) and the Acoustic Comfort Vote (ACV) were positively correlated. (3) When “neutral”, the Thermal Sensation Vote (TSV) increased with increasing illumination intensity (LUX). (4) The Sunlight Sensation Vote (SSV) increased with the increasing Universal Thermal Climate Index (UTCI) when illumination intensity (LUX) was moderate and bright. (5) The highest and lowest Acoustic Sensation Vote (ASV) values for each sound type came from either “slightly cold” or “warm” conditions.

Production of sounds by squirrelfish during symbiotic relationships with cleaner wrasses

Examples of symbiotic relationships often include cleaning mutualisms, typically involving interactions between cleaner fish and other fish, called the clients. While these cleaners can cooperate by removing ectoparasites from their clients, they can also deceive by feeding on client mucus, a behavior usually referred to as “cheating behavior” that often leads to a discernible jolt from the client fish. Despite extensive studies of these interactions, most research has focused on the visual aspects of the communication. In this study, we aimed to explore the role of acoustic communication in the mutualistic relationship between cleaner fishes and nine holocentrid client species across four regions of the Indo-Pacific Ocean: French Polynesia, Guam, Seychelles, and the Philippines. Video cameras coupled with hydrophones were positioned at various locations on reefs housing Holocentridae fish to observe their acoustic behaviors during interactions. Our results indicate that all nine species of holocentrids can use acoustic signals to communicate to cleaner fish their refusal of the symbiotic interaction or their desire to terminate the cooperation. These sounds were predominantly observed during agonistic behavior and seem to support visual cues from the client. This study provides a novel example of acoustic communication during a symbiotic relationship in teleosts. Interestingly, these vocalizations often lacked a distinct pattern or structure. This contrasts with numerous other interspecific communication systems where clear and distinguishable signals are essential. This absence of a clear acoustic pattern may be because they are used in interspecific interactions to support visual behavior with no selective pressure for developing specific calls required in conspecific recognition. The different sound types produced could also be correlated with the severity of the client response. There is a need for further research into the effects of acoustic behaviors on the quality and dynamics of these mutualistic interactions.

Vocal complexity in the long calls of Bornean orangutans.

Vocal complexity is central to many evolutionary hypotheses about animal communication. Yet, quantifying and comparing complexity remains a challenge, particularly when vocal types are highly graded. Male Bornean orangutans (Pongo pygmaeus wurmbii) produce complex and variable "long call" vocalizations comprising multiple sound types that vary within and among individuals. Previous studies described six distinct call (or pulse) types within these complex vocalizations, but none quantified their discreteness or the ability of human observers to reliably classify them. We studied the long calls of 13 individuals to: (1) evaluate and quantify the reliability of audio-visual classification by three well-trained observers, (2) distinguish among call types using supervised classification and unsupervised clustering, and (3) compare the performance of different feature sets. Using 46 acoustic features, we used machine learning (i.e., support vector machines, affinity propagation, and fuzzy c-means) to identify call types and assess their discreteness. We additionally used Uniform Manifold Approximation and Projection (UMAP) to visualize the separation of pulses using both extracted features and spectrogram representations. Supervised approaches showed low inter-observer reliability and poor classification accuracy, indicating that pulse types were not discrete. We propose an updated pulse classification approach that is highly reproducible across observers and exhibits strong classification accuracy using support vector machines. Although the low number of call types suggests long calls are fairly simple, the continuous gradation of sounds seems to greatly boost the complexity of this system. This work responds to calls for more quantitative research to define call types and quantify gradedness in animal vocal systems and highlights the need for a more comprehensive framework for studying vocal complexity vis-à-vis graded repertoires.

Resolving the Loss of Intermediate-Size Speech Aerosols in Funnel-Guided Particle Counting Measurements

Modeling of airborne virus transmission and protection against it requires knowledge of the amount of biofluid emitted into the atmosphere and its viral load. Whereas viral concentrations in biofluids are readily measured by quantitative PCR, the total volume of fluids aerosolized during speaking, as measured by different researchers using various technologies, differs by several orders of magnitude. We compared collection methods in which the aerosols first enter into a low-humidity chamber either by direct injection or via commonly used funnel and tubing arrangements, followed by standard optical particle sizer measurement. This “collect first, measure later” approach sacrifices the recording of the temporal correlation between aerosol generation and sound types such as plosives and vowels. However, the direct-injection mode prevents inertia deposition associated with the funnel arrangements and reveals far more intermediate-size (5–20 μm in diameter) particles that can dominate the total mass of ejected respiratory aerosol. The larger aerosol mass observed with our method partially reconciles the large discrepancy between the SARS-CoV-2 infectious dose estimated from superspreader event analyses and that from human challenge data. Our results also impact measures to combat airborne virus transmission because they indicate that aerosols that settle faster than good room ventilation rates can dominate this process.

Exploring an Intelligent Classification Model for the Recognition of Automobile Sounds Based on EEG Physiological Signals

The advancement of an intelligent automobile sound switching system has the potential to elevate the market standing of automotive products, with the pivotal prerequisite being the selection of automobile sounds based on the driver’s subjective perception. The subjective responses of diverse individuals to sounds can be objectively manifested through EEG signals. Therefore, EEG signals are employed herein to attain the recognition of automobile sounds. A subjective evaluation and EEG signal acquisition experiment are designed involving the stimulation of three distinct types of automobile sounds, namely comfort, power, and technology sounds, and a comprehensive database of EEG signals corresponding to these three sound qualities is established. Then, a specific transfer learning model based on a convolutional neural network (STL-CNN) is formulated, where the method of training the upper layer parameters with the fixed bottom weights is proposed to adaptively extract the EEG features related to automobile sounds. These improvements contribute to improving the generalization ability of the model and realizing the recognition of automobile sounds fused with EEG signals. The results of the comparison with traditional support vector machine (SVM) and convolutional neural network (CNN) models demonstrate that the accuracy of the test set of the STL-CNN model reaches 91.5%. Moreover, its comprehensive performance, coupled with the ability to adapt to individual differences, surpasses that of both SVM and CNN models. The demonstrated method in the recognition of automobile sounds based on EEG signals is of significance for the future implementation of switching driving sound modes fused with EEG signals.

Soundscape Characterization Using Autoencoders and Unsupervised Learning.

Passive acoustic monitoring (PAM) through acoustic recorder units (ARUs) shows promise in detecting early landscape changes linked to functional and structural patterns, including species richness, acoustic diversity, community interactions, and human-induced threats. However, current approaches primarily rely on supervised methods, which require prior knowledge of collected datasets. This reliance poses challenges due to the large volumes of ARU data. In this work, we propose a non-supervised framework using autoencoders to extract soundscape features. We applied this framework to a dataset from Colombian landscapes captured by 31 audiomoth recorders. Our method generates clusters based on autoencoder features and represents cluster information with prototype spectrograms using centroid features and the decoder part of the neural network. Our analysis provides valuable insights into the distribution and temporal patterns of various sound compositions within the study area. By utilizing autoencoders, we identify significant soundscape patterns characterized by recurring and intense sound types across multiple frequency ranges. This comprehensive understanding of the study area's soundscape allows us to pinpoint crucial sound sources and gain deeper insights into its acoustic environment. Our results encourage further exploration of unsupervised algorithms in soundscape analysis as a promising alternative path for understanding and monitoring environmental changes.

Formant indicators of «velarization». Part I

This article discusses the problem of determining velarization from acoustic data and its relationship to articulation. The aim of the study is to identify the range of formants by which «velarization» can be determined, as well as to determine the location of the highest point of localization of the part of the back of the tongue of the relative palatine arch with this articulation, modeling the sound quality of a specific consonantal sound type. In the course of the study, the methods used at the V.M. Nadelyaev LFEI of the sector of languages of the peoples of Siberia of the IFL SB RAS (Novosibirsk), as well as in the LEFI of the Department of Foreign Languages of the AmSU (Blagoveshchensk). The object of the study is dynamic speech. The subject of the study is «velarization» in languages of different families. The material for the analysis was static and dynamic MRI scans and audio recordings recorded at different times by employees (including the author of this work) of the V.M. Nadelyaev LEFI, made on time and after tomography from the appropriate media. Scientific novelty and theoretical significance of the study: the correlation between F2 , articulation and auditory perception of amplifying articulation of hardness – “velarization” is presented for the first time. The materials and results of the research can be used to develop textbooks on general phonetics, at seminars on dynamic speech analysis, and when writing scientific articles on experimental phonetics. The main conclusions presented in this paper are: 1) the acoustic effect of “velarization” is achieved not only as a result of lifting the back of the tongue to the soft palate not in guttural consonants (velarization), but also between the last third of the second half of the hard palate (prevelarization); 2) correlation analysis of somatic and acoustic material revealed the dependence of format indicators on a certain articulation; 3) a possible range of “velarization” in consonants has been identified – F2 = 1500 Hz – 550 Hz; 4) in the range from 1500 Hz to 1301 Hz, velarization sounds slightly “softer” by ear – a prevelarized consonant than from 1300 Hz to 550 Hz – a velarized consonant.

Speech Enhancement with Background Noise Suppression in Various Data Corpus Using Bi-LSTM Algorithm

Noise reduction is one of the crucial procedures in today’s teleconferencing scenarios. The signal-to-noise ratio (SNR) is a paramount factor considered for reducing the Bit error rate (BER). Minimizing the BER will result in the increase of SNR which improves the reliability and performance of the communication system. The microphone is the primary audio input device that captures the input signal, as the input signal is carried away it gets interfered with white noise and phase noise. Thus, the output signal is the combination of the input signal and reverberation noise. Our idea is to minimize the interfering noise thus improving the SNR. To achieve this, we develop a real-time speech-enhancing method that utilizes an enhanced recurrent neural network with Bidirectional Long Short Term Memory (Bi-LSTM). One LSTM in this sequence processing framework accepts the input in the forward direction, whereas the other LSTM takes it in the opposite direction, making up the Bi-LSTM. Considering Bi-LSTM, it takes fewer tensor operations which makes it quicker and more efficient. The Bi-LSTM is trained in real-time using various noise signals. The trained system is utilized to provide an unaltered signal by reducing the noise signal, thus making the proposed system comparable to other noise-suppressing systems. The STOI and PESQ metrics demonstrate a rise of approximately 0.5% to 14.8% and 1.77% to 29.8%, respectively, in contrast to the existing algorithms across various sound types and different input signal-to-noise ratio (SNR) levels.