Source Separation Research Articles

Application of Lock-in Amplifier Technique in Signal Blind Source Separation

Abstract In the realm of signal processing, frequency error stands as a critical challenge that cannot be overlooked, particularly in scenarios characterized by noise disturbances and constantly changing environmental variables, as it significantly undermines the precision of waveform separation. Faced with a spectrum of frequency error phenomena stemming from hardware instability, such as clock drift, transmission distortion, and environmental changes, this paper proposes a waveform separation scheme centered around a core closed-loop control architecture aimed at addressing such challenges. Leveraging the high-performance STM32F407 microcontroller platform, the system harnesses the powerful algorithmic advantages of Fast Fourier Transform (FFT) to precisely locate and delineate the spectral characteristics of signals. Furthermore, we creatively integrate lock-in amplifier technology into the meticulously designed closed-loop control system, coupled with state-of-the-art zero-crossing detection algorithms for in-depth optimization. The research findings demonstrate outstanding signal-to-noise ratio performance within predefined target frequency bands for the separation device based on lock-in amplifier technology, showcasing not only remarkable resilience against interference but also the ability to accurately separate and reconstruct target signal components with high fidelity from complex and dynamic signal environments. This significantly enhances the robustness and accuracy of the entire system.

Method of field measurement of electromagnetic environment of high speed maglev train

Abstract In order to reduce the complexity of the magnetic field test of maglev trains and improve the accuracy of measurement data, this paper proposes a field measurement method of maglev train radiation emission based on high-speed parallel adaptive filter denoising technology. This measurement method is not limited by special measurement sites, and has the characteristics of high-speed measurement, real-time data processing, and large-capacity data storage. It breaks through the technical bottleneck of the traditional measurement system. The electromagnetic radiation emission of high-speed maglev trains is measured in real-time through multi-antenna synchronous measurement technology, and multi-channel high-speed RF signal measurement and acquisition of external signals are further carried out by blind source separation technology, environmental pre-scanning is carried out, environmental interference characteristics are obtained through feature statistics, and the radiation signal to be measured is estimated by blind source separation algorithm. It helps to achieve the purpose of accurately measuring the electromagnetic radiation emission of high-speed maglev trains. Through field testing, the denoising performance of the method reached 20 dB. Through field testing, the denoising performance of the method reached 20 dB.

Decomposing the AIA 304 Å Channel into Its Cool and Hot Components

The AIA 304 Å channel on board the Solar Dynamics Observatory (SDO) offers a unique view of ≈105 K\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$\\approx 10^{5}\ ext{ K}$\\end{document} plasma emitting in the He ii 304 Å line. However, when observing off-limb, the emission of the (small) cool structures in the solar atmosphere (such as spicules, coronal rain and prominence material) can be of the same order as the surrounding hot coronal emission from other spectral lines included in the 304 Å passband, particularly over active regions. In this paper, we investigate three methods based on temperature and morphology that are able to distinguish the cool and hot emission within the 304 Å passband. The methods are based on the Differential Emission Measure (DEM), a linear decomposition of the AIA response functions (RFit) and the Blind Source Separation (BSS) technique. All three methods are found to produce satisfactory results in both quiescent and flaring conditions, largely removing the diffuse corona and leading to images with cool material off-limb in sharp contrast with the background. We compare our results with co-aligned data from the Interface Region Imaging Spectrograph (IRIS) in the SJI 1400 Å and 2796 Å channels, and find the RFit method to best match the quantity and evolution of the cool material detected with IRIS. Some differences can appear due to plasma emitting in the logT=5.1–5.5\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$\\log T=5.1\\,\ ext{--}\\,5.5$\\end{document} temperature range, particularly during the catastrophic cooling stage prior to rain appearance during flares. These methods are, in principle, applicable to any passband from any instrument suffering from similar cool and hot emission ambiguity, as long as there is good coverage of the high-temperature range.

The Validity and Reliability of the Turkish Form of the Epilepsy Disclosure Scale—Youth and Parent Versions

AbstractAlthough explaining epilepsy is a separate source of stress for children with epilepsy and their parents, studies evaluating the disclosure of epilepsy by patients and their parents are insufficient. The aim of this study was to test the validity and reliability of the Turkish form of the “Epilepsy Disclosure Scale (EDS)—Youth and Parent Versions,” which measures the concealment/disclosure of epilepsy by youth patients with epilepsy and their parents. The population of the study consisted of 126 children who were diagnosed with epilepsy and who were between the ages of 8 to 18 and their parents (63 children and 63 parents) who applied to two hospitals pediatric neurology. Both scales consist of six items. When the scale was adapted, language, content, structural, and reliability analyses were conducted. The factor loads varied between 0.78 and 0.88 and contributed 71.99% to the total variance in the Youth Version. In the Parent Version, they varied between 0.79 and 0.88 and contributed 67.09% to the total variance. The Cronbach's α coefficients of the youth and parent versions of the scale were calculated as 0.92 in the youth version and 0.90 in the parent version. The Composite Reliability Index of the youth version was 0.94, and that of the parent version was 0.92. It was concluded that all statistical studies in the study were compatible with the original scale and that it could be applied to children with epilepsy between the ages of 8 to 18 and their parents in Turkish society.

Enhancement the wastewater treatment performance of multistage living machine by underwater lamp

Source separation and decentralized domestic wastewater treatment represent effective strategies to enhance sewage treatment performance and facilitate water reuse economically. The Living Machine (LM) system has gained widespread adoption for decentralized sewage treatment. While underwater light source has been demonstrated to enhance the treatment performance of open aerobic reactors in LM systems, its influence on the treatment efficiency of a fully multistage LM system remains underreported. In this study, an underwater lamp-added LM system (ULLM) with eight reactors was constructed and investigated. The introduction of underwater light source obviously improved the removal capacity of chemical oxygen demand (COD) and NH4+-N, which was 96.1% and 61.6%, respectively. The diversity of algae, zooplankton, and aquatic animals was notably higher in the light-treated reactors than in the control group (CK) without underwater light source, and substantial alteration in the microbial community of the light-treated reactors was observed compared with CK reactors. At the phylum level, Proteobacteria and Nitrospirae enriched in the underwater light-treated reactors, while Bacteroidetes and Actinobacteria exhibited a decrease after light exposure. At the genus level, Nitrospira and Rhodanobacter were enriched in the ULLM system. Importantly, the prevalence of these two dominant genera was sustained until the final operational stage, indicating their potential key roles in enhancing wastewater treatment performance. The addition of underwater light source proves to be an effective strategy for augmenting the treatment efficiency of the multistage living machine systems, resulting in substantial improvements in pollutant removal. These findings contribute valuable insights into optimizing LM systems for decentralized wastewater treatment.

Comparing the three-dimensional morphological asymmetries in the ejecta of Kepler and Tycho in X-rays

Aims. Recent simulations have shown that asymmetries in the ejecta distribution of supernova remnants (SNRs) may be a reflection of asymmetries left over from the initial supernova explosion. Thus, SNR studies provide a vital means for testing and constraining model predictions in relation to the distribution of heavy elements, which are key to improving our understanding of the explosion mechanisms in Type Ia supernovae. Methods. The use of a novel blind source separation method applied to the megasecond X-ray observations of the historic Kepler and Tycho supernova remnants has revealed maps of the ejecta distribution. These maps are endowed with an unprecedented level of detail and clear separations from the continuum emission. Our method also provides a three-dimensional (3D) view of the ejecta by individually disentangling red- and blueshifted spectral components associated with images of the Si, S, Ar, Ca, and Fe emission. This approach provides insights into the morphology of the ejecta distribution in those two remnants. Results. Those mappings have allowed us to thoroughly investigate the asymmetries in the intermediate-mass elements and Fe distribution in two Type Ia supernova remnants. We also compared the results with the core-collapse Cassiopeia A remnant, which we had studied previously. The images obtained confirm, as expected for Type Ia SNRs, that the Fe distribution is mostly closer to the core than that of intermediate-mass elements. They also highlight peculiar features in the ejecta distribution, such as the Fe-rich southeastern knot in Tycho.

Blind Sound Source Separation by Combining the Convolutional Neural Network and Degree Separator

Blind Sound Source Separation by Combining the Convolutional Neural Network and Degree Separator

Pengembangan Potensi Nelayan Melalui Pelatihan Manajemen Usaha Pengolahan Ikan Asap Di Gks Hanggaroru, Desa Kaliuda

Fisheries, especially capture fisheries, have quite large potential in the coastal area of ​​East Sumba. One of the potential catch fisheries in East Sumba is Kaliuda Village. The GKS Hanggaroru congregation is a community in Kaliuda village whose entire community depends on fishery products, both capture and aquaculture. Fisheries problems that often occur in Kaliuda village are the large number of catches that are damaged due to the way they are handled after being caught, due to the unavailability of ice, and the length of the marketing process. This results in losses at the community level so it is necessary to find a solution regarding partnerships and marketing the catch. This mentoring activity was carried out for 1 month through socialization and mentoring. Implementation of this activity includes socialization, distribution of leaflets on procedures for making smoked fish, training in making smoked fish using a cabinet with a separate smoke source using a drum, as well as designing packaging and packaging using a vacuum sealer. Partners are happy with the training activities carried out by the PKM team.

Unveiling Hidden Sources of Dynamic Functional Connectome through a Novel Regularized Blind Source Separation Approach

Abstract The investigation of the brain’s functional connectome and its dynamic changes can provide valuable insights into brain organization and its reconfiguration. However, the analysis of dynamic functional connectivity (dFC) using functional magnetic resonance imaging (fMRI) faces major challenges, including the high dimensionality of brain networks, unknown latent sources underlying observed dFC, and the large number of brain connections that increase the risk of spurious findings. In this paper, we propose a new regularized blind source separation (BSS) method called dyna-LOCUS to address these challenges. dyna-LOCUS decomposes observed dFC measures to reveal latent source connectivity traits and their dynamic temporal expression profiles. By utilizing low-rank factorization and novel regularizations, dyna-LOCUS achieves efficient and reliable mapping of connectivity traits underlying the dynamic brain functional connectome, characterizes temporal changes of the connectivity traits that contribute to the reconfiguration in the observed dFC, and generates parsimonious and interpretable results in identifying whole-brain dFC states. We introduce a highly efficient iterative Node-Rotation algorithm that solves the non-convex optimization problem for learning dyna-LOCUS. Simulation studies demonstrate the advantages of our proposed method. Application of dyna-LOCUS to the Philadelphia Neurodevelopmental Cohort (PNC) study unveils latent connectivity traits and key brain connections and regions driving each of these neural circuits, reveals temporal expressions and interactions of these connectivity traits, and generates new findings regarding gender differences in the neurodevelopment of an executive function-related connectivity trait.

Comprehensive agricultural ecological effects of aeration on regenerated liquid fertilizer of mini flush toilet

The high concentration of organic waste liquid obtained from the mini flush pipeline discharge technology based on source separation has the potential for fertilizer utilization. However, there are concerns about the risk of secondary pollution. This study proposes the idea of aeration treatment for regenerated liquid fertilizers to induce beneficial changes in their material composition and properties. Initially, this study compares the characteristic changes in nitrogen transformation of liquid fertilizer through aeration treatment. Subsequently, it examines the effects of different types of liquid fertilizers on soil properties, plant physiology, and soil microbial communities. Finally, we elucidate the flow and distribution of nitrogen in soil, plants, and nitrogen-containing gas emissions in agricultural ecosystems through material flow accounting. The study found that aeration treatment can reduce the ammonia nitrogen ratio while increasing the proportions of nitrite nitrogen and nitrate nitrogen. The regenerated liquid fertilizer through aeration treatment not only significantly increased the chlorophyll, protein, and polysaccharide content of vegetable leaves (P < 0.05) but also reduced nitrate accumulation. Moreover, it can reduce the risk of soil nitrate nitrogen leaching and increase the diversity of soil bacterial communities, enhancing the ecological functions of bacteria involved in carbon and nitrogen cycling. Material flow accounting indicated that aeration treatment for liquid fertilizer could reduce gaseous nitrogen loss by 50.0 %, improve the nitrogen utilization efficiency of vegetables by 95.5 %, and enhance soil nitrogen retention by 11.4 %. Overall, the results show that aeration treatment can improve the agricultural utilization of liquid fertilizer and reduce the risk of secondary pollution, providing preliminary decision-making support for optimizing resource treatment strategies for mini-flush toilet fecal waste to realize the agricultural cycle.

Quantifying NOx point sources with Landsat and Sentinel-2 satellite observations of NO2 plumes

We show that the Landsat and Sentinel-2 satellites can detect NO2 plumes from large point sources at 10 to 60 m pixel resolution in their blue and ultrablue bands. We use the resulting NO2 plume imagery to quantify nitrogen oxides (NOx) emission rates for several power plants in Saudi Arabia and the United States, including a 13-y analysis of 132 Landsat plumes from Riyadh power plant 9 from 2009 through 2021. NO2 in the plumes initially increases with distance from the source, likely reflecting recovery from ozone titration. The fine pixel resolutions of Landsat and Sentinel-2 enable separation of individual point sources and stacks, including in urban background, and the long records enable examination of multidecadal emission trends. Our inferred NOx emission rates are consistent with previous estimates to within a precision of about 30%. Sources down to ~500 kg h-1 can be detected over bright, quasi-homogeneous surfaces. The 2009 to 2021 data for Riyadh power plant 9 show a strong summer peak in emissions, consistent with increased power demand for air conditioning, and a marginal slow decrease following the introduction of Saudi Arabia's Ambient Air Standard 2012.

A geometric solution to microphone array position calibration and its confidence interval analysis

Wireless acoustic sensor networks (WASNs) comprised of spatially distributed acoustic nodes have been increasingly popular in many areas such as speech enhancement, blind source separation, and audio surveillance, where node position information is generally required to implement relevant processing algorithms. In this article, we focus on the node position calibration only using direction-of-arrival (DOA) information with each node consisting of a microphone array. State-of-the-art techniques translate the position calibration into optimization of DOA-based cost functions, and have shown notable success. However, a compromise between position calibration accuracy and computational load still exists. Furthermore, none of the existing studies can offer a quantitative evaluation of calibration performance before the optimization procedure is finished. To address these problems, a geometric solution to array node position calibration is proposed using DOA measurements. For a given array node, DOAs of three sources are first utilized to construct cylinder, plane, and cone equations. Next, the two-dimensional projection of intersections among these equations is calculated using Sylvesterresultantmatrix. Finally, a closed-form solution of node position in three-dimensional space is established. Moreover, explicit mathematical derivation of the confidence interval analysis on the calibration results under given node angle estimation error condition is presented, indicating that the proposed method can provide predictable performance with corresponding probabilities. The Cramér-Rao bound (CRB) is further investigated. Simulation results reveal that as compared with existing methods our approach achieves comparable calibration accuracy with much less computational burden. Real-world experiments also verify its effectiveness.

Tracking rivalry with neural rhythms: multivariate SSVEPs reveal perception during binocular rivalry.

The contents of awareness can substantially change without any modification to the external world. Such effects are exemplified in binocular rivalry, where a different stimulus is presented to each eye causing instability in perception. This phenomenon has made binocular rivalry a quintessential method for studying consciousness and the necessary neural correlates for awareness. However, to conduct research on binocular rivalry usually requires self-reports of changes in percept, which can produce confounds and exclude states and contexts where self-reports are undesirable or unreliable. Here, we use a novel multivariate spatial filter dubbed 'Rhythmic Entrainment Source Separation' to extract steady state visual evoked potentials from electroencephalography data. We show that this method can be used to quantify the perceptual switch-rate of participants during binocular rivalry and therefore may be valuable in experimental contexts where self-reports are methodologically problematic or impossible, particularly as an adjunct. Our analyses also reveal that 'no-report' conditions may affect the deployment of attention and thereby neural correlates, another important consideration for consciousness research.

Elicitation of fetal ECG from abdominal recordings using Blind Source Separation techniques and Robust Set Membership Affine Projection algorithm for signal quality enhancement

Background:The utilization of non-invasive techniques for fetal cardiac health surveillance is pivotal in evaluating fetal well-being throughout the gestational period. This process requires clean and interpretable fetal Electrocardiogram (fECG) signals. Method:The proposed work is the novel framework for the elicitation of fECG signals from abdominal ECG (aECG) recordings of the pregnant mother. The comprehensive approach encompasses pre-processing of the raw ECG signal, Blind Source Separation techniques (BSS), Decomposition techniques like Empirical Mode Decomposition (EMD), and its variants like Ensemble Empirical Mode Decomposition (EEMD), and Complete Ensemble Empirical Mode Decomposition with Additive Noise (CEEMDAN). The Robust Set Membership Affine Projection (RSMAP) Algorithm is deployed for the enhancement of the obtained fECG signal. Result:The results show significant improvements in the elicited fECG signal with a maximum Signal Noise Ratio (SNR) of 31.72 dB and correlation coefficient = 0.899, Maximum Heart Rate(MHR) obtained in the range of 108–142 bpm for all the records of abdominal ECG signals. The statistical test gave a p-value of 0.21 accepting the null hypothesis. The Abdominal and Direct Fetal Electrocardiogram Database (ABDFECGDB) from PhysioNet has been used for this analysis. Conclusion:The proposed framework demonstrates a robust and effective method for the elicitation and enhancement of fECG signals from the abdominal recordings.

Joint dereverberation and blind source separation using a hybrid autoregressive and convolutive transfer function-based model

Most frequency-domain blind source separation (BSS) methods are based on the multiplicative narrowband assumption, which is not valid in long reverberation environments. In contrast, convolutive transfer function (CTF)-based BSS methods do not rely on the narrowband assumption, and the separation performance is significantly improved compared to the traditional algorithms in long reverberation environments. However, the CTF-based BSS methods and their variants, e.g., autoregressive (AR) BSS methods, introduce modeling errors to some extent, due to the truncation or approximation during the optimization process. To address this problem, we propose a frequency-domain BSS method employing a hybrid AR and CTF model, which can provide more precise representations of the early reflections and late reverberations. Furthermore, we utilize the Gaussian noise model to deal with the BSS problem in noisy reverberant environments. We formulate the objective function using the maximum log-likelihood criterion, and derive an efficient iterative algorithm for parameter estimation with the block coordinate descent (BCD) method. Experimental results show that the proposed method has a better separation performance than the existing methods in long reverberation environments.

Enhancing Concrete Durability: Sustainable Ternary Blends Incorporating Agricultural and Industrial Wastes

&lt;p&gt;The demand for sustainable concrete in civil infrastructure operations grows by the day, resulting in the use of agricultural and industrial byproducts and controlling high-cost nanomaterials. Ternary blended concrete is an excellent alternative to conventional concrete for long-term development by lowering carbon footprint. Ternary blended concrete is a green and sustainable concrete made from three separate source components to make a binder. The fundamental benefit of ternary mixed concrete is its densely packed particles of various shapes and sizes, resulting in superior characteristics. This study discusses experimental studies that were conducted to assess the durability properties of regular ordinary Portland cement (OPC) concrete and ternary blended concrete containing 1% nano-SiO2 and industrial or agricultural byproducts. Concrete durability is defined as its capacity to sustain a safe degree of serviceability and various environmental exposure circumstances without requiring considerable repair or rehabilitation during its service life. Conventional concrete is prone to cracking because of its low tensile and durability properties. Six potential mixes were investigated in this work: three blends based on industrial byproducts (fly ash, ground granulated blast furnace slag, and metakaolin) and three mixes based on agricultural byproducts (rice husk ash, corncob ash, and sugarcane bagasse ash). This study investigated the durability characteristics such as water permeability, sorptivity, resistance to sulphate, acid attack, and alkaline attack. The experimental test findings were determined to be well within limitations when compared to the requirements specified in the standard/literature. The study also shows that ternary blended concrete with regulated nano-SiO2 enhances durability metrics significantly. Furthermore, energy dispersive spectrophotometry (EDS) measurements show dense phases with high proportions of Si atoms in the substrate's penetration layer. As a result, ternary mixed concrete is strongly advised in structural repair works.&lt;/p&gt;

MIRACLE—a microphone array impulse response dataset for acoustic learning

This work introduces a large dataset comprising impulse responses of spatially distributed sources within a plane parallel to a planar microphone array. The dataset, named MIRACLE, encompasses 856,128 single-channel impulse responses and includes four different measurement scenarios. Three measurement scenarios were conducted under anechoic conditions. The fourth scenario includes an additional specular reflection from a reflective panel. The source positions were obtained by uniformly discretizing a rectangular source plane parallel to the microphone for each scenario. The dataset contains three scenarios with a spatial resolution of 23mm\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$23\\,\ extrm{mm}$$\\end{document} at two different source-plane-to-array distances, as well as a scenario with a resolution of 5mm\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$5\\,\ extrm{mm}$$\\end{document} for the shorter distance. In contrast to existing room impulse response datasets, the accuracy of the provided source location labels is assessed and additional metadata, such as the directivity of the loudspeaker used for excitation, is provided. The MIRACLE dataset can be used as a benchmark for data-driven modelling and interpolation methods as well as for various acoustic machine learning tasks, such as source separation, localization, and characterization. Two timely applications of the dataset are presented in this work: the generation of microphone array data for data-driven source localization and characterization tasks and data-driven model order reduction.

A modified T-type multilevel inverter for renewable energy applications

The primary challenge in integrating renewable resources into grids using multilevel inverters (MLI) is the need for many separate DC sources and switching device counts. Transformer-based multilevel inverters (TMIs) have emerged to address this issue, aiming to minimize system components and boost source voltage with a single DC source. This research article introduces a novel TMI topology that utilizes only a single DC source and incorporates ten switches to produce good-quality load voltage with high magnitude. The proposed TMI offers several structural advantages, including self-galvanic isolation, reduced switching devices and uniform voltage levels across all turn ratios. Additionally, the TMI operates a switching method called pulse width modulation, which provides the gating pulses to all the power semiconductor devices in the proposed TMI. An experimental model has been created in a laboratory environment, and simulations are performed using the MATLAB/Simulink platform to assess the effectiveness of the suggested TMI. Furthermore, a comparison between the suggested TMI circuit and other recent TMI designs with similar characteristics is performed. This comparison is carried out to assess and validate the superior features of the proposed TMI over the alternative designs.

Sustainable Municipal Solid Waste Management Strategies and Practices: A review

Purpose: This paper introduces a multi-period, multi-objective model for creating an integrated, sustainable municipal solid waste management supply chain. The model includes source separation and a reward-penalty mechanism, aiming to minimize total system costs, greenhouse gas emissions, and environmental impacts on residential areas. Many countries strive to develop efficient solid waste management systems (SWMS) that handle and dispose of daily waste cost-effectively while maintaining sustainability. These systems include waste sources, collection stations, landfills, incinerators, recycling plants, and a transportation network. Methodology: A systematic literature review was conducted on existing literature from various studies. Recent studies were used for this systematic review of sustainable municipal solid waste management strategies and practices. Findings: Decision-makers must design or reconfigure a sustainable MSWM to determine the optimal supply chain network for treating and disposing of daily waste effectively. This paper suggests sustainable waste management strategies and practices from various countries, aiming to determine the best number and locations for facilities and optimal waste flow within the system to minimize daily costs. The primary contribution of this review is the theoretical development of sustainable MSWM practices applicable worldwide. Unique contribution to theory, policy and practice: Cooperation between the private and public sectors involved in solid waste management and the circular economy is imperative to ensure the technical, financial, and social sustainability of waste management systems. Large-scale awareness programs on solid waste management and the circular economy should be implemented, focusing on fostering environmentally responsible behaviors and attitudes, promoting sorting, selective collection, and the recovery of materials and energy from waste.

Application of Unet-SE-Bisru Algorithm for Music Signal Processing in Music Source Separation

At present, the use of time-domain deep learning for end-to-end neural network models has the problem of long training time and poor performance in music source separation. To address this issue, a U-network squeezing excitation bidirectional simple recursive unit model was proposed based on the deep extractor model. Replace Unet SE Bisru with Unet SE Bisru in the following text. This model improves the bidirectional long short-term memory network into a bidirectional simple recurrent unit, and then introduces attention mechanisms in the generalized encoding and decoding layers. The squeezing excitation block is used to selectively extract features based on the type of audio to be separated. Finally, group normalization is added after onedimensional convolution, And its effectiveness was verified. The experimental results show that the signal noise distortion ratio in the improved model is 5.68 decibels compared to the bidirectional simple recursive unit value, which is higher than the 5.55 decibels of bidirectional long short-term memory. After adding the squeezing excitation module, the overall increase is about 0.1-0.5 decibels. In addition, in the model comparison, the three indicators of the improved model with the same number of channels were 5.68 decibels, 5.91 decibels, and 11.28 decibels, respectively, higher than the benchmark model. Compared with other music source classification models, the improved model has better comprehensive separation performance. Although some indicators are lower than the comparison model, the signal noise distortion ratio of drum and bass is 6.11 decibels and 6.36 decibels, which is better than the comparison model. Overall, the improved model has high performance in music source separation for music signal processing and can be effectively applied in practical music source separation.