Individual Spectral Components Research Articles

Stacked U-Net with Time–Frequency Attention and Deep Connection Net for Single Channel Speech Enhancement

Deep neural networks have significantly promoted the progress of speech enhancement technology. However, a great number of speech enhancement approaches are unable to fully utilize context information from various scales, hindering performance enhancement. To tackle this issue, we introduce a method called TFADCSU-Net (Stacked U-Net with Time-Frequency Attention (TFA) and Deep Connection Layer (DCL)) for enhancing noisy speech in the time–frequency domain. TFADCSU-Net adopts an encoder-decoder structure with skip links. Within TFADCSU-Net, a multiscale feature extraction layer (MSFEL) is proposed to effectively capture contextual data from various scales. This allows us to leverage both global and local speech features to enhance the reconstruction of speech signals. Moreover, we incorporate deep connection layer and TFA mechanisms into the network to further improve feature extraction and aggregate utterance level context. The deep connection layer effectively captures rich and precise features by establishing direct connections starting from the initial layer to all subsequent layers, rather than relying on connections from earlier layers to subsequent layers. This approach not only enhances the information flow within the network but also avoids a significant rise in computational complexity as the number of network layers increases. The TFA module consists of two attention branches operating concurrently: one directed towards the temporal dimension and the other towards the frequency dimension. These branches generate distinct forms of attention — one for identifying relevant time frames and another for selecting frequency wise channels. These attention mechanisms assist the models in discerning “where” and “what” to prioritize. Subsequently, the TA and FA branches are combined to produce a comprehensive attention map in two dimensions. This map assigns specific attention weights to individual spectral components in the time–frequency representation, enabling the networks to proficiently capture the speech characteristics in the T-F representation. The results confirm that the proposed method outperforms other models in terms of objective speech quality as well as intelligibility.

Temperature measurements based on a composite of nanosized phosphors [Ru(dipy)<sub>3</sub>]<sup>2+</sup>@SiO<sub>2</sub> and NaYF<sub>4</sub>:Eu,Gd

The results of temperature testing of the two-phosphor composite ([Ru(dipy)3]2+@SiO2 and NaYF4:Eu,Gd) are presented in order to assess the prospects for its further application as a luminescent thermal sensor in the range of 200–310 K. By calibrating the temperature dependences of individual spectral components of the emission produced under irradiation by cw laser at 405 nm, the possibilities of temperature sensing by means of ratiometric method are analyzed in four different spectral ranges of the composite’s visible luminescence.

Detection of Water on Road Surface with Acoustic Vector Sensor.

This paper presents a new approach to detecting the presence of water on a road surface, employing an acoustic vector sensor. The proposed method is based on sound intensity analysis in the frequency domain. Acoustic events, representing road vehicles, are detected in the sound intensity signals. The direction of the incoming sound is calculated for the individual spectral components of the intensity signal, and the components not originating from the observed road section are discarded. Next, an estimate of the road surface state is calculated from the sound intensity spectrum, and the wet surface detection is performed by comparing the estimate with a threshold. The proposed method was evaluated using sound recordings made in a real-world scenario, and the algorithm results were compared with data from a reference device. The proposed algorithm achieved 89% precision, recall and F1 score, and it outperforms the traditional approach based on sound pressure analysis. The test results confirm that the proposed method may be used for the detection of water on the road surface with acoustic sensors as an element of a smart city monitoring system.

An experimental approach to unravel 2D ground resonances: application to an alluvial-sedimentary basin

The study of ground resonances is important to assess seismic site amplification and to infer information on the geometrical and mechanical properties of the resonating structures. 1D- and 2D-type resonances imply different dynamic behavior that can be distinguished by inspecting the individual spectral components of single-station microtremor measurements. Typically, 2D resonance modes develop along cross-sections of deep sediment-filled valleys and consist of longitudinal, transverse and vertical modes that can be identified as spectral peaks when ground motion is recorded parallel to the axes of the valley. In the case of more complex geometries, such as sedimentary basins, resonance modes are more difficult to predict and depend on the unknown complexity of the buried bedrock geometry. We show how a simple signal rotation procedure applied to single-station microtremor recordings reveals the underlying 2D resonance pattern. The method allows assessing the axes of motion of buried geological structures and identifying 2D resonance modes along these axes. Their directionality, frequency and amplitude features are then analyzed to extract information on the bedrock geometry. We test our method in the Bolzano alluvial-sedimentary basin and we observe that apparently complicated resonance patterns may be simplified by locally referring to the simplest description of the phenomenon as 2D resonance of a valley slice. The bedrock morphology can be decomposed into 2D-like geometries, i.e., excavated channels, and the observed resonances develop within cross-sections of these channels.Graphical

Evaluating the jet/accretion coupling of Aql X-1: probing the contribution of accretion flow spectral components

ABSTRACTThe coupling between radio and X-ray luminosity is an important diagnostic tool to study the connection between the accretion inflow and jet outflow for low-mass X-ray binaries (LMXBs). The radio/X-ray correlation for individual neutron star (NS) LMXBs is scattered, whereas for individual black hole (BH) LMXBs a more consistent correlation is generally found. Furthermore, jet quenching is observed for both types of LMXBs, but it is unclear whether jets in NS-LMXBs quench as strongly as those in BH-LMXBs. While additional soft X-ray spectral components can be detected in NS-LMXB spectra due to the presence of the NS surface, disentangling the individual X-ray spectral components has thus far not been considered when studying the radio/X-ray coupling. Here we present eleven epochs of Swift/XRT observations matched with quasi-simultaneous archival radio observations of the 2009 November outburst of Aql X-1. We decompose the thermal and Comptonized spectral components in the Swift/XRT spectra, with the aim of studying whether the presence of additional thermal emission affects the coupling of the radio/X-ray luminosity. We find that there is no evidence of a significant thermal contribution in Swift/XRT spectra that could cause scatter in the radio/X-ray coupling. To explore the role of potential spectral degeneracies in the X-ray models and consider the improvements from including hard X-rays, we perform joint fits with quasi-simultaneous RXTE/PCA spectra. Follow-up research using more sensitive, broad-band X-ray observations and densely sampled near-simultaneous radio observations is required to study this in more detail.

Метод функций Грина для расчета нестационарных спектров люминесценции неравновесных молекулярных систем

Interest in ultrafast photochemical processes is due to their role in wildlife and the possibility of using them in solar energy conversion devices. Significant part of experimental research in this area is carried out using different spectroscopic techniques, for example, based on recording the dynamic response of the system to optical excitation by a short laser pulse. This method is basic for photochemistry and is used for studying both small inorganic molecules in liquids or mixtures, and complex biological objects, such as photosynthetic reaction centers of bacteria and plants. Time-resolved luminescence/absorption spectra contain information about the processes of population of the electronic and vibrational states of the system, and therefore make it possible to study the mechanisms of photoreactions. At the same time, the question of the correct interpretation of experimental data in the case of ultrafast reactions at picosecond timescales is still relevant. The traditional approach is based on the decomposition of transient spectra into the dynamic components, associated with different electronic states. This state-associated analysis turns out to be much less accurate in ultrafast nonequilibrium reactions. The spectral dynamics of the system in this case depends not only on chemical transformations (that is, changes in state populations), but also on the evolution of the spectral response for each of the states. Particularly, relaxation of high-frequency intramolecular vibrations in ultrafast reactions was shown to cause not only a shift in the luminescence spectrum, but also significant changes in spectral profiles themselves. An assumption about the invariable shape of individual spectral components cannot be justified here. Relevant analysis of spectral dynamics in such systems thus requires taking into account the nonequilibrium state of intramolecular degrees of freedom and the environment. Development of theoretical models of nonequilibrium processes, as well as software tools for numerical simulation of spectral dynamics and the fitting techniques for experimental results data, allows us to improve the relevance of the analysis. This study is devoted to the development of a mathematical model of spectral dynamics of macromolecular systems, in which photoexcitation triggers a series of ultrafast electron transfer reactions involving several redox centers. The main attention is paid to accounting for nonequilibrium states of the medium and intramolecular degrees of freedom formed both at the stage of photoexcitation and in the course of a multistage reaction. This allows us to expand the range of phenomena under study and apply the model for interpretation of experimental data on multistage charge transfer in biological objects. The model is based on the semiclassical theory of multistage electron transfer in a multicomponent non-Debye solvent, as well as the Green’s function method for the classical (polarization) and quantum (intramolecular) coordinates of the system.

Directional diagrams of acoustic radiation from unmanned aerial vehicles

When solving the urgent task of detecting small unmanned aerial vehicles (UAVs) by their acoustic radiation (AR), there is a need to study the AR characteristics of UAVs. Therefore, considerable attention is paid in the literature to theoretical and experimental studies of the structure and parameters of the sound field created by UAVs.
 This article is devoted to the experimental study of the directional diagrams of the acoustic radiation from the DJI Phantom 3 unmanned aerial vehicle. The UAV AR recording experiment was conducted in a "silenced" chamber, the walls of which are covered with sound-absorbing panels with a surface of a special geometric shape. The experimental setup includes the UAV mounted on a boom, a microphone for sound recording, and a boom for the microphone.
 Studies of the structure and parameters of the sound field of the quadcopter have shown that the spectra of the emitted signal have pronounced harmonic components with frequencies multiples of the propeller rotation frequency. The spectral components have the greatest power in the frequency range up to 500 Hz, where the first harmonic has the largest amplitude, and then the spectrum components decrease to the ambient noise level.
 Obtained two-dimensional and three-dimensional directional diagrams of the UAV AR with and without propellers when only the aircraft engine is operating. It is shown that in the absence of screws, the acoustic radiation is much weaker in level. The experimental data was also presented in the form of three-dimensional AR diagrams for the four harmonics of the acoustic signal, and it was analyzed what changes in the spatial orientation of the UAV AR are observed based on changes in the three-dimensional figure for each of the radiation harmonics.
 It is shown that the spatial distribution of both the total energy (over the entire frequency range) of the acoustic signal and the energy of its individual spectral (harmonic) components is significantly anisotropic. It follows from this conclusion that the range of UAV detection and observation in real conditions is a statistical value that depends on the angle of its observation.

On the wake response of a square cylinder subjected to an inline nonharmonic forcing

In this work, we study the effect of two spectrally simple nonharmonic inline forcing waveforms on the wake response of a square cylinder. The characteristics of wake response under nonharmonic forcing were studied and compared to wake response under harmonic forcing for a constant Reynolds number 100. By varying the relative strength of the spectral components and the secondary frequency of the nonharmonic forcing of the inflow, an asymmetric modulated wake and an asymmetric 2P + 2S wake were observed for a two-tone nonharmonic forcing. These modes combine the features of the wake response of individual spectral components. In some regions of the parameter space, the wake response did not correlate well with the spectral components of harmonic forcing. An amplitude-modulated forcing was also studied. For amplitude-modulated forcing, the additional frequency component did not affect the wake structure. This observation indicates that the wake may act as a low pass filter for inline flow frequency components above the natural vortex shedding frequency.

Development and characterisation of an irradiation device for biomedical studies covering the solar spectrum with individual regulated spectral bands

To understand the importance of terrestrial solar exposure on human skin, not only individual spectral components need to be considered in biomedical studies, but also the relevance of the combined action profile of the complete solar spectrum (cSS) must be established. We therefore developed a novel irradiation device that combines the emission of four individual lamps (UVB, UVA, VIS and nIR) to achieve exposure from 280 to 1400 nm with individual controllable lamps. The integrated irradiance of each spectral band is similar to the solar spectrum. The lamps can be utilised individually or in any desired combination. Here we present the design, realisation, and validation of this irradiation device as well as biological results on cellular metabolism (MTT assay), cell cycle alterations, and clonogenic growth in HaCaT cells after exposures to the individual spectral bands as well as their simultaneous combinations. Thereby, we demonstrate that UVB combined with UVA is the main determinant for the metabolic activity within cSS. Also, UVB-dependent effects dominate cell cycle regulation in cSS, whilst UVA and nIR have little influence. Lastly, also clonogenic growth is dominated by the UVB action profile in cSS, despite nIR showing modulatory activity when applied in combination with UVB. Together, this highlights the regulatory influence of the different spectral bands on the three biological endpoints and demonstrates their modulation when being part of the complete solar spectrum.Graphical abstract

Multispectral Dip for Improved Geometric Attribute Computations

Accurate estimates of volumetric dip serve as input for curvature, aberrancy, reflector convergence attributes, guide the application of structure-oriented filters and the computation of coherence, amplitude gradient, as well as GLCM texture attributes. Different frequency components can exhibit slightly different dips, due to the differences in their resolution and sensitivity to noise. Research during the last few years has shown that multispectral coherence computed by summing the covariance matrices of individual spectral voice components can provide significant improvement over coherence computed from the covariance matrix computed from the original broadband data. A similar workflow can be constructed by computing multispectral dip by summing the gradient structure tensors computed from the individual spectral voice components. We apply this workflow to two different seismic datasets from the Taranaki and Canterbury Basins of New Zealand and find that the resulting multispectral dip estimates exhibit sharper images than those computed from the original broadband data. Images from the Taranaki dataset show improvements on the definition of a fault system, whereas those from the Canterbury dataset exhibit enhanced definition of faults and associated fractures as seen on the dip estimates, as well as both structural and amplitude curvature displays

Exploring the potential of PCA-based quantitation of NMR signals in T1 relaxometry

Principal component analysis (PCA) has proved to be a powerful technique for processing NMR data. It is particularly useful in signal quantitation where it often provides better results compared to a direct integration of individual signals. In the present work, we recapitulate the principles and theoretical framework underlying PCA-based quantitation with a special focus on T1 relaxometry. We show that under commonly encountered conditions, this approach can provide up to ~4-fold improvement in scatter of points in magnetization build-up curves compared to direct integration. Best practices to optimize the PCA performance in measuring the total magnetization are discussed, including minimization of the number of signal-related principal components and a proper selection of FT parameters and data quantitation intervals. For signals consisting of distinct relaxation components, formulas are provided for resolving the components relaxation and illustrated on a real-data example. In addition to the problem of quantitation, the use of PCA in denoising of partially relaxed spectra is discussed in connection with such applications as line shape analysis and monitoring relaxation of individual spectral components.

Study of the Perception of Written Speech Using Functional Tomography Based On Electroencephalography Data

The spectral and spatial characteristics of the electroencephalograms recorded during the perception of written speech were studied. For the experimental study, four groups were formed, each containing 100 words: words with a positive emotional rating, words with a negative emotional rating, words with concrete meanings, and words with abstract meanings. A separate experiment was conducted for each group with the subjects. Words were represented by white text on a black background, each word was presented for 1000 ms, after the presentation of the stimulus there was a pause of 500 ms. Brain activity was recorded using an electroencephalograph with 19 leads, arranged according to the 10–20 scheme. For detailed quantitative analysis of this activity, method of functional tomography of the brain, based on electroencephalography data, was used. This method is based on the Fourier transform of multichannel encephalographic data and the localization of individual spectral components. The method makes it possible to single out and stably localize in space various spectral features of the brain activity studied in experiments on speech research. The frequency band from 8 to 30 Hz was analyzed; for all spectral components in this band, the inverse problem was solved in the approximation of an equivalent current dipole in a single-layer spherical conductor, without any restrictions on the position of the source. As a result, three-dimensional maps of activity were built - the functional structures of the brain. The presentation of these functional structures on magnetic resonance imaging allows one to study the frequency and spatial characteristics of responses to various speech stimuli.

Multiplexed lock-in thermography.

Many modern measurement methods for heat transfer work in the frequency domain. A certain average temperature rise in the sample is unavoidable if the sample can only be heated, e.g., by an intensity modulated light source. This average temperature rise influences the measured thermal properties because they are, in general, temperature dependent. Here, we explore square waves and multiplexed sine waves with the goal of reducing the average temperature rise. We apply these signals to lock-in thermography and show the feasibility of a simultaneous measurement at multiple frequencies. In addition, we propose the use of the Goertzel algorithm to efficiently extract individual spectral components from the temperature signal.

Adopting multispectral dip components for coherence and curvature attribute computations

Although volumetric coherence is the most widely used geometric attribute, accurate estimates of volumetric dip are in some ways more important. Coherence, amplitude gradients, and gray-level co-occurrence matrix textures should be computed along structural dip. Curvature and aberrancy are computed from volumetric estimates of structural dip. Because of both differences in resolution and sensitivity to coherent noise, different frequency components may exhibit different dip. In recent years, improvements in coherence have been noticed where covariance matrices of individual spectral components are summed rather than summing the original broadband data. We extend the same concepts to compute multispectral dip estimates by using a gradient structure tensor algorithm. The results are sharper, less smeared images on the dip components. The higher-resolution dip estimates result in higher-resolution curvature and aberrancy estimates. Availability of sharper estimates of dip to guide coherence attribute results in more continuous, less noisy discontinuities.

Some features of the day-to-day MLT wind variability in winter 2017–2018 as seen with a European/Siberian meteor radar network

We present results of wind measurements near the mesopause carried out with meteor radars (MRs) at Collm (51°N, 13°E), Obninsk (55°N, 37°E), Kazan (56°N, 49°E), Angarsk (52°N, 104°E) and Anadyr (65°N, 178°E) from October 1, 2017 till March 31, 2018. The Collm and Kazan MRs are SKiYMET radars with vertical transmission and radio echo height finding, while the other radars operate with horizontal transmission and without height finding. We paid particular attention to the meridional wind variability with periods of 4–6 days and 9–11 days. The waves with these periods are seen as spots of the wave activity in the wavelet spectra and include oscillations with different periods and different discrete zonal wavenumbers. These wave packets successively propagate as a group of waves from one site to another one in such a way that they are observed at one site and almost disappear at the previous one. The 4–6 wave group includes planetary-scale oscillations (individual spectral components) which have eastward phase velocities and mostly zonal wavenumbers 2 and 3, and the vertical wavelength exceeds 70 km at middle latitudes. The source of the oscillations is the polar jet instability. The wave group itself propagates westward, and the amplitudes of wind oscillations are approximately 5–6 m/s as obtained from the wind data averaged over the meteor zone. The 9–11 day wave set propagates westward as a group and mainly consists of spectral components which have westward phase velocity and zonal wavenumber 1. Amplitudes of these wind perturbations strongly vary from station to station and can reach, approximately, 8 m/s. The vertical wavenumber is 0.014 km−1 as taken from the Kazan and 0.05 km−1 according to the Collm data. We obtained a global view on the waves by using the AURA MLS geopotential data. We found a good correspondence between wave features obtained from the MR wind measurements and the MLS data. To our knowledge, such a wave propagation of planetary wave in the mesosphere/lower thermosphere (MLT) region has so far not obtained much attention.

X-ray analysis of the accreting supermassive black hole in the radio galaxy PKS 2251+11

Context. We have investigated the dichotomy between jetted and non-jetted active galactic nuclei (AGNs), focusing on the fundamental differences of these two classes in the accretion physics onto the central supermassive black hole (SMBH). We tested the validity of the unification model of AGNs through the characterization of the mutual interaction between accreting and outflowing matter in radio galaxies. Aims. Our aim is to study and constrain the structure, kinematics and physical state of the nuclear environment in the broad line radio galaxy (BLRG) PKS 2251+11. The high X-ray luminosity and the relative proximity make such AGN an ideal candidate for a detailed analysis of the accretion regions in radio galaxies. The investigation will help to shed light on the analogies and differences between the BLRGs and the larger class of radio-quiet Seyfert galaxies and hence on the processes that trigger the launch of a relativistic jet. Methods. We performed a spectral and timing analysis of a ∼64 ks observation of PKS 2251+11 in the X-ray band with XMM-Newton. We modeled the spectrum considering an absorbed power law superimposed to a reflection component. We performed a time-resolved spectral analysis to search for variability of the X-ray flux and of the individual spectral components. Results. We find that the power law has a photon index Γ = 1.8 ± 0.1, absorbed by an ionized partial covering medium with a column density NH = (10.1 ± 0.8) × 1023 cm−2, a ionization parameter log ξ = 1.3 ± 0.1 erg s−1 cm and a covering factor f ≃ 90%. Considering a density of the absorber typical of the broad line region (BLR), its distance from the central SMBH is of the order of r ∼ 0.1 pc. An Fe Kα emission line is found at 6.4 keV, whose intensity shows variability on timescales of hours. We derive that the reflecting material is located at a distance r ≳ 600rs, where rs is the Schwarzschild radius. Conclusions. Concerning the X-ray properties, we found that PKS 2251+11 does not differ significantly from the non-jetted AGNs, confirming the validity of the unified model in describing the inner regions around the central SMBH, but the lack of information regarding the state of the very innermost disk and SMBH spin still leaves unconstrained the origin of the jet.

Energy‐dependent timing studies of the low‐hard state of black hole X‐ray binaries with XMM‐Newton

Almost all low‐mass black hole X‐ray binaries are transient sources. Most of these sources show a hysteresis: they evolve from the low‐hard state through intermediate state(s) into the high‐soft state and then return to the hard state at lower luminosity, again passing through the intermediate states. However, there are outbursts that remain in the hard state (so called “failed” outbursts). While mean X‐ray spectra allow us to study the time‐averaged spectral shape, we need to make use of “variability spectra,” like, for example, covariance spectra, to investigate the variability of individual spectral components on different time scales. Comprehensive studies of covariance spectra of a sample of black hole X‐ray binaries observed in the low‐hard state of “normal” outbursts revealed an increase of the covariance ratios toward lower energies that has been interpreted as the sign of additional disk variability on long time scales. In the case of GX 339‐4, we found that the covariance ratio increases and steeps during outburst decay, consistent with increased disk instabilities. There are two sources (H 1743‐322 and GS 1354‐64) that do not show an increase toward lower energies in their covariance ratio. We show that there is no correlation between the shape of the covariance ratio and the amount of fractional rms variability. Both these sources have been observed during “failed” outbursts and their photon indices are at the hard end of the photon index range usually observed in black hole X‐ray binaries.

Spatio-temporal and spatiospectral metrology of terahertz broadband uniformly topologically charged vortex beams.

A comprehensive characterization of the diffraction properties of terahertz (THz) pulsed broadband vortex beams consisting of several electromagnetic field oscillations requires state-of-the-art techniques for studying the evolution of a wavefront as it propagates. For this purpose, we have applied the capabilities offered by THz pulse time domain holography. Accurate metrological study of pulsed single-period THz field propagation allowed us to reveal the spatio-temporal and spatiospectral couplings in broadband uniformly topologically charged vortex beams. Here, we reveal dynamics of such beam propagation in a free space as well as in the experiment with edge diffraction with 50% blocking of the beam focal waist. In this study, we compare the dynamics of freely propagating and edge-diffracted THz vortex. Despite the fact that in the amplitude representation one can observe the emergence of strong asymmetry, analysis of the spectral trajectory of the singular point at some distance from the obstacle and the visualization of phase distribution for individual spectral components testify to the conservation of transverse energy circulation. Similar to the edge diffraction of monochromatic optical vortices, it can be interpreted as self-reconstruction of vortex properties. The given term has not previously been used for the case of pulsed broadband THz beams, to the best of our knowledge.

Heteroassociation of Caffeine and a Phenanthridine Dye in Aqueous Solution

Association of ethidium bromide and caffeine in aqueous solution was studied using spectrophotometry. The results were analyzed using a molecular dimer heteroassociation model. The molecular association equilibrium constants were determined. The experimental absorption spectrum of a mixture of heterocyclic molecules was deconvoluted into individual components using a regularization method. Quantitative information on the structure of the heterocomplex was obtained using data for the individual spectral components. The distance and angle between point transition dipole moments of the heterogeneous molecules were 0.3 nm and 16o. A calculation of the relative contents of different types of associates in a mixed solution showed that heteroassociation of the molecules decreased the effective concentration of ethidium bromide and, accordingly, the mutagenic activity of the dye.

Coherence attribute applications on seismic data in various guises — Part 2

We have previously discussed some alternative means of modifying the frequency spectrum of the input seismic data to modify the resulting coherence image. The simplest method was to increase the high-frequency content by computing the first and second derivatives of the original seismic amplitudes. We also evaluated more sophisticated techniques, including the application of structure-oriented filtering to different spectral components before spectral balancing, thin-bed reflectivity inversion, bandwidth extension, and the amplitude volume technique. We further examine the value of coherence computed from individual spectral voice components, and alternative means of combining three or more such coherence images, providing a single volume for interpretation.