Additive Background Noise Research Articles

Linear prediction coefficients correction method for digital speech processing systems with data compression based on the autoregressive model of a voice signal

The problem of distortion of the autoregressive model of the voice signal under the influence of additive background noise in digital speech processing systems with data compression based on linear prediction is considered. In the frequency domain, these distortions are observed in the weakening of the main formants responsible for the intelligibility of the speaker’s speech. To compensate for formant attenuation, it is proposed to modify the parameters of the autoregressive model (linear prediction coefficients) using the impulse response of a recursive shaping filter. Along with the amplitude amplification of the formants, their frequencies remain unchanged to make the speaker’s voice recognizable. The effectiveness of the method was studied experimentally using specially developed software. Based on the experimental results, conclusions were drawn about a significant increase in the relative level of formants in the power spectrum of the corrected voice signal.

Single-grain luminescence and combined U-series/ESR dating of the early Upper Palaeolithic Lagar Velho Rock Shelter, Leiria, Portugal

The existing radiocarbon (14C) dating framework for Lagar Velho is broadly consistent but provides limited constraint on several geoarchaeological complexes (GCs) and does not favour detailed chronological comparisons across all sectors of the site; including the stratigraphically disconnected child burial complex in the east area and the broader archaeological sequence in the central-west area. In this study, we undertake a complementary chronological assessment of Lagar Velho Rock Shelter using single-grain quartz OSL, single-grain quartz TT-OSL and combined U-series/ESR dating of fossil teeth, with the aim of establishing more comprehensive reconstructions of archaeological events and placing the human occupation sequence in a firmer regional climatic context. Expanding on the original chronological study, we also compare the published 14C datasets against widely used sample quality indicators (i.e., organic preservation and contamination proxies) to ensure reliable comparisons with our new luminescence and combined U-series/ESR ages. Eight 14C samples pass our combined chronological and stratigraphic hygiene criteria and are included in the final dating evaluations. Ten of the twelve optical dating samples produce homogeneous equivalent dose (De) datasets indicative of suitable daylight exposure. The replicate single-grain TT-OSL De datasets exhibit more pronounced scatter, consistent with enhanced potential for insufficient bleaching of TT-OSL residual doses in some karst settings. The fossil enamel samples dated using U-series/ESR span relatively low natural dose ranges, necessitating the inclusion of maximum dose (Dmax)-adjusted dose response curve fitting and additional background noise subtraction to avoid De biases of up to 13%. Stratigraphically consistent ages (n = 26) spanning the full archaeological infill sequence are obtained using the four different dating methods, providing a robust interpretive framework and underscoring the significant role of single-grain OSL and combined U-series/ESR dating for refining early Upper Palaeolithic histories of Iberia. Bayesian modelling of the combined chronological dataset reveals a site accumulation history spanning 35,750–23,440 cal. BP and provides improved age constraints on all GCs; particularly the al, bs, tc, gs and ls deposits that were previously undated or imprecisely constrained. The age of the child burial complex is refined to 29,410–28,280 cal. BP, potentially occurring within Greenland Interstadial 4. Comparison of the modelled GC ages from the east and west-central areas enables improved temporal correlations of depositional events across stratigraphically disconnected sectors of the site. This analysis suggests that the child burial complex and the Late Gravettian occupation are statistically indistinguishable based on current dating evidence. The new dating assessments reveal three periods of human occupation at Lagar Velho: the Late Gravettian (including the shorter duration child burial event) 31.6–26.0 ka from the end of Heinrich Stadial (HS) 3 until the emergence of HS2; the Terminal Gravettian 26.5–24.5 ka coinciding with HS2; and the Middle Solutrean 25.2–23.4 ka extending from HS2 to the end of Greenland Stadial 3. The timing of these occupations during late Marine Isotope Stage (MIS) 3 and early MIS 2 highlights the significance of central western Portugal for supporting regional human populations during periods when much of Europe was experiencing extreme cold, aridity and expanded glacial coverage.

Investigating conversational dynamics in triads: Effects of noise, hearing impairment, and hearing aids.

Communication is an important part of everyday life and requires a rapid and coordinated interplay between interlocutors to ensure a successful conversation. Here, we investigate whether increased communication difficulty caused by additional background noise, hearing impairment, and not providing adequate hearing-aid (HA) processing affected the dynamics of a group conversation between one hearing-impaired (HI) and two normal-hearing (NH) interlocutors. Free conversations were recorded from 25 triads communicating at low (50 dBC SPL) or high (75 dBC SPL) levels of canteen noise. In conversations at low noise levels, the HI interlocutor was either unaided or aided. In conversations at high noise levels, the HI interlocutor either experienced omnidirectional or directional sound processing. Results showed that HI interlocutors generally spoke more and initiated their turn faster, but with more variability, than the NH interlocutors. Increasing the noise level resulted in generally higher speech levels, but more so for the NH than for the HI interlocutors. Higher background noise also affected the HI interlocutors' ability to speak in longer turns. When the HI interlocutors were unaided at low noise levels, both HI and NH interlocutors spoke louder, while receiving directional sound processing at high levels of noise only reduced the speech level of the HI interlocutor. In conclusion, noise, hearing impairment, and hearing-aid processing mainly affected speech levels, while the remaining measures of conversational dynamics (FTO median, FTO IQR, turn duration, and speaking time) were unaffected. Hence, although experiencing large changes in communication difficulty, the conversational dynamics of the free triadic conversations remain relatively stable.

Numerical simulation of the equatorial plasma bubble: the effect of seeding by the vertical winds and random background noise perturbations

A wide variety of small-amplitude waves widely exist in the ionosphere and have significant effects on the evolution of equatorial plasma bubbles. In this paper, we simulated equatorial plasma bubbles (EPB) seeded by vertical neutral wind perturbations with wavelengths of 125 km and 250 km, and compared the morphology characteristics of plasma bubble structures with those under random noise perturbations in the background density. The numerical results showed that both vertical winds and random background noise perturbations can contribute to the growth of plasma bubbles, and the perturbations under additional random background noise can promote the growth of the plasma bubble structures faster. Additionally, several processes of the nonlinear behavior of bifurcated EPB structures, including bifurcation, pinching, and small-scale turbulent structures, were successfully obtained. Our simulation captured supersonic flows within the low-density plasma structures characterized by vertical velocities of about 1.5 km/s, which is consistent with experimental studies found in the literature.

Deep neural network analysis models for complex random telegraph signals

Time-fluctuating signals are ubiquitous and diverse in many physical, chemical, and biological systems, among which random telegraph signals (RTSs) refer to a series of instantaneous switching events between two discrete levels from single-particle movements. A reliable RTS analysis is a crucial prerequisite to identify underlying mechanisms related to device performance and sensitivity. When numerous levels are involved, complex patterns of multilevel RTSs occur and make their quantitative analysis exponentially difficult, hereby systematic approaches are often elusive. In this work, we present a three-step analysis protocol via progressive knowledge-transfer, where the outputs of the early step are passed onto a subsequent step. Especially, to quantify complex RTSs, we resort to three deep neural network architectures whose trained models can process raw temporal data directly. We furthermore demonstrate the model accuracy extensively with a large dataset of different RTS types in terms of additional background noise types and amplitude size. Our protocol offers structured schemes to extract the parameter values of complex RTSs as imperative information with which researchers can draw meaningful and relevant interpretations and inferences of given devices and systems.

Robust Localization for Near- and Far-Field Signals with an Unknown Number of Sources

Source location is a constant issue of importance of both theoretical study and practical engineering. Many pioneers have put out the corresponding solutions for near- or far-field signals, and preferred contributions are suggested. To our best knowledge, there are currently few focused approaches to the complicated situation where both near- and far-field signals exist with an unknown number of sources. Additionally, the robustness of the method must be taken into account when the additive background noise does not follow Gaussian or super-Gaussian distribution. To solve these problems, a novel method based on phased fractional lower-order moment (PFLOM) is proposed to simultaneously better preserve the signal and suppress the noise. Secondly, the whole procedure of the method containing direction of arrival (DOA) estimation, range estimation, separation of near-and far-field sources, and crucial parameter settings are studied in detail. Finally, comprehensive Monte Carlo experiments are carried out in the simulation to demonstrate the superiority of the proposed method compared to the existing competitive methods. Due to the novel method’s effectiveness with an unknown number of sources and robustness against various noises, it is believed that it could be fully utilized in more fields.

Speech Enhancement and Dereverberation With Diffusion-Based Generative Models

In this work, we build upon our previous publication and use diffusion-based generative models for speech enhancement. We present a detailed overview of the diffusion process that is based on a stochastic differential equation and delve into an extensive theoretical examination of its implications. Opposed to usual conditional generation tasks, we do not start the reverse process from pure Gaussian noise but from a mixture of noisy speech and Gaussian noise. This matches our forward process which moves from clean speech to noisy speech by including a drift term. We show that this procedure enables using only 30 diffusion steps to generate high-quality clean speech estimates. By adapting the network architecture, we are able to significantly improve the speech enhancement performance, indicating that the network, rather than the formalism, was the main limitation of our original approach. In an extensive cross-dataset evaluation, we show that the improved method can compete with recent discriminative models and achieves better generalization when evaluating on a different corpus than used for training. We complement the results with an instrumental evaluation using real-world noisy recordings and a listening experiment, in which our proposed method is rated best. Examining different sampler configurations for solving the reverse process allows us to balance the performance and computational speed of the proposed method. Moreover, we show that the proposed method is also suitable for dereverberation and thus not limited to additive background noise removal. Code and audio examples are available online <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1</sup> <uri xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">https://github.com/sp-uhh/sgmse</uri> .

Optimal Multi-Level Amplitude-Shift Keying Modulation for a Reliable Channel Magnitude Detection-Based Receive Diversity Powerline Communication System

The implementation of a coherent powerline communication (PLC) receiver increases its hardware complexity and energy consumption, which can be tackled by a non-coherent receiver. However, this reduces the reliability of the PLC systems. To meet this trade-off, a novel energy-efficient receiver structure is proposed in this study that exploits the knowledge of the channel magnitude of a receive diversity PLC system. The transmitter employs multi-level amplitude-shift keying (ASK) modulation for data transmission via the PLC channels subjected to Rayleigh multipath phenomenon and corrupted by additive Nakagami-m background noise, which would lead to a superior performance as compared to other modulation schemes owing to the channel magnitude-based proposed receiver. Utilizing the receiver structure, closed-form expression for symbol error probability (SEP) and its asymptotic expression at a high average signal-to-noise ratio (SNR) are derived using a characteristic function approach. Furthermore, to achieve the simultaneous design of an energy-efficient modulation scheme and a reliable PLC system, optimal ASK signaling levels are obtained with a constraint on the total available energy at the transmitter end. The analytical relation of the derived optimal multi-level ASK constellation with the traditional ASK constellation is obtained at high SNR values. The proposed channel magnitude-based receiver is found to achieve full diversity order and numerical results presented provide various insights towards the design of the energy-efficient PLC system.

Performance Analysis of Mixed PLC-FSO Dual-Hop Communication Systems

In this article, we analyze the performance of a mixed dual-hop power line communication/free-space optical communication (PLC-FSO) transmission system supporting both decode-and-forward (DF) and fixed-gain amplify-and-forward (AF) relaying protocols. Further, it is assumed that the PLC link experiences log-normal fading under the effects of additive background and impulsive noise, while the free-space optical (FSO) channel obeys an exponentiated Weibull distribution with pointing errors. On this basis, we derive the probability density function (PDF) and cumulative distribution function (CDF) of the end-to-end (e2e) signal-to-noise ratio (SNR). To evaluate the system performance, the closed-form expressions for the outage probability (OP) and average bit-error rate (BER) are derived. Additionally, to gain more insights, the asymptotic analysis of OP and average BER, and the upper bound of the average capacity are presented.

Performance Analysis of Dual-Hop Mixed PLC/RF Communication Systems

In this article, we study a dual-hop mixed power line communication and radio-frequency communication (PLC/RF) system, where the connection between the PLC link and the RF link is made by a decode-and-forward or amplify-and-forward relay. It is assumed that the PLC channel is affected by both additive background noise and Poisson&#x2013;Gaussian impulsive noise which suffer from log-normal fading, while the RF link undergoes generalized Rician fading. Based on this model, analytical expressions for the outage probability (OP), average bit error rate (BER), and the average channel capacity (ACC) are derived. Furthermore, asymptotic analyses for the OP and average BER as well as an upper bound expression for the ACC are presented. Illustrative numerical results are presented along with insightful discussions.

Counteracting Dark Web Text-Based CAPTCHA with Generative Adversarial Learning for Proactive Cyber Threat Intelligence

Automated monitoring of dark web (DW) platforms on a large scale is the first step toward developing proactive Cyber Threat Intelligence (CTI). While there are efficient methods for collecting data from the surface web, large-scale dark web data collection is often hindered by anti-crawling measures. In particular, text-based CAPTCHA serves as the most prevalent and prohibiting type of these measures in the dark web. Text-based CAPTCHA identifies and blocks automated crawlers by forcing the user to enter a combination of hard-to-recognize alphanumeric characters. In the dark web, CAPTCHA images are meticulously designed with additional background noise and variable character length to prevent automated CAPTCHA breaking. Existing automated CAPTCHA breaking methods have difficulties in overcoming these dark web challenges. As such, solving dark web text-based CAPTCHA has been relying heavily on human involvement, which is labor-intensive and time-consuming. In this study, we propose a novel framework for automated breaking of dark web CAPTCHA to facilitate dark web data collection. This framework encompasses a novel generative method to recognize dark web text-based CAPTCHA with noisy background and variable character length. To eliminate the need for human involvement, the proposed framework utilizes Generative Adversarial Network (GAN) to counteract dark web background noise and leverages an enhanced character segmentation algorithm to handle CAPTCHA images with variable character length. Our proposed framework, DW-GAN, was systematically evaluated on multiple dark web CAPTCHA testbeds. DW-GAN significantly outperformed the state-of-the-art benchmark methods on all datasets, achieving over 94.4% success rate on a carefully collected real-world dark web dataset. We further conducted a case study on an emergent Dark Net Marketplace (DNM) to demonstrate that DW-GAN eliminated human involvement by automatically solving CAPTCHA challenges with no more than three attempts. Our research enables the CTI community to develop advanced, large-scale dark web monitoring. We make DW-GAN code available to the community as an open-source tool in GitHub.

Microphone Array Beamforming With High Flexible Interference Attenuation and Noise Reduction

This paper studies the problem of microphone array beamforming to enhance a speech signal of interest in adverse acoustic environments, where interference and additive background noise coexist. The problem is formulated as one of convex optimization whose solution under a specified level of interference attenuation leads to an interference controlled maximum noise reduction (ICMR) beamformer, which can be expressed as a linear combination of two MVDR beamformers: one attempts to extract the desired source signal while the other attempts to extract the interference. The combination coefficients are functions of the array manifold vectors, noise coherence matrix, and the specified interference attenuation factor. By tuning the interference attenuation factor, the ICMR beamformer can be implemented to achieve aggressive interference attenuation or even eliminate interference completely; but this may lead to less additive noise suppression or even noise amplification. To control the maximum sacrifice in gain (SG) of the signal-to-noise ratio (SNR) that is acceptable for additive reduction, a variant of ICMR is derived, which is named as the ICMR-SG beamformer. Simulations are performed and the results show that the ICMR beamformer is able to control the amount of interference attenuation. In comparison, ICMR-SG controls the maximum SG of SNR while achieving the optimal possible level of interference attenuation.

On the dynamics and robustness of the chemostat with multiplicative noise

Abstract The stochastic dynamics of a two-state bioreactor model with random feed flow fluctuations and non-monotonic specific growth rate is analyzed. Using the Fokker-Planck equation approach for describing the probability density function (PDF) evolution the lack of stochastic robustness due to deterministic bifurcation phenomena for the open-loop reactor operating under optimal (maximum production) operation condition is established, and the associated stochastic stabilization problem is addressed. Inherent differences between the presence of multiplicative noise, due to the feed flow fluctuations, and additive background noise are analytically established. Numerical simulation results illustrate these inherent differences, the stochastic fragility of the open-loop operation yielding a stochastic extinction phenomenon, as well as the stochastic PDF stabilization with a proportional feedback control.

Hyperbolic tangent cyclic correlation and its application to the joint estimation of time delay and doppler shift

Gaussian distribution is a classic model for additive background noise in scientific research. By assuming and applying Gaussian noise, scholars in different fields can apply themselves to fundamental research and propose many innovative theories based upon Gaussianity. As the scope of engineering applications continues to expand, a single model of background noise cannot meet the more stringent conditions, and the errors caused by inappropriate noise models cannot be ignored. In this paper, alpha-stable distribution is utilized to model additive background noise because of the generality of this distribution. Correspondingly, a novel similarity measurement named hyperbolic tangent correlation (HTC) is proposed to replace the conventional correlation. Then, hyperbolic tangent cyclic correlation (HTCC) used in cyclostationary signal processing is proposed and thoroughly studied. HTC and HTCC emphasize preserving the signal of interest (SOI) rather than suppressing the noise, as traditionally practiced. In particular, they effectively process both the amplitude and phase information in the SOI. To demonstrate the superiority of HTCC in processing amplitude and phase information against background noise, a representative application in time-frequency analysis is presented, and the resulting performance is compared to that of the existing methods. Considering the beneficial mathematical properties of the hyperbolic tangent function, we hope that HTC and HTCC will have more applications in signal processing.

Autocorrelated Envelopes for early fault detection of rolling bearings

The rolling element bearings are extensively applied in rotating machines, and they are the most susceptible components in rotating machines. Early fault detection of bearings is to prevent machines from such typical failures and subsequent consequences. In this paper a detector based on Ensemble Average of Autocorrelated Envelopes (EAAE) is proposed to identify the early occurrence faults in rolling element bearings, of which the fault induced vibration signals are inevitably contaminated or masked by both additive background noise and random phase noise (or slippage between bearing components). To enhance the cyclostationary characteristics for fault detection, it utilizes the phase synchronization property of autocorrelation signals for aligning the cyclostationary signals in the lag domain to achieve an effective ensemble average which allows both types of random influences to be suppressed significantly. As a result, this detector shows very high performance of robustness in extracting the local fault signatures, which is verified by simulation signals and experimental investigations and benchmarked by the recent milestone method of Spectral Correlation (SC).

Bark scaled oversampled WPT based speech recognition enhancement in noisy environments

The performance of speech recognition system degrades significantly in real-world environment, is a case of the acoustic mismatch between the training and operating conditions. This paper presents a two-stage approach to make a speech recognition system immune to additive and uncorrelated background noise i.e. robust. In the first stage, an oversampled wavelet packet decomposes the entire input noisy speech into seventeen nonlinear frequency subbands like the Bark scale of the human hearing system and the adaptive noise estimation based spectral subtraction filters the noisy speech from each subband signal. The oversampled WPT is linear and advantageous as it causes to overcome the shift-invariance complexity by removing the decimation after the filtering at each decomposition level. In the second stage, a nonparametric approach is used for feature extraction from filtered speech, and the parameters from the feature extraction stage are compared with the parameters extracted from speech signals stored in a template to recognize the utterance. A series of experiments are carried out to evaluate the performance of the proposed two-stage system in a variety of real environments, with and without the use of the first stage. Recognition accuracy is evaluated at the word level in a wide range of SNRs for various types of noisy environments. The experimental results show significant improvement in recognition performance at low SNR using the proposed system.

Performance Analysis of a Mixed Cooperative PLC–VLC System for Indoor Communication Systems

In this paper we propose a dual-hop mixed cooperative power line communication (PLC)–visible light communication (VLC) system, in which the VLC system is linked to the PLC backhaul link with the assistance of a decode and forward relay. The performance of the PLC link is evaluated under the effect of log-normal fading, and the link is presumed to be affected by both additive background and impulsive noises. The VLC link's channel characteristics are taken as a function of the position of the end users. The proposed study provides novel closed-form expressions of various statistics of the equivalent end-to-end signal-to-noise ratio of the considered system. Taking these derived entities into account, we also evaluate closed-form expressions of various performance metrics of the proposed system.

Performance analysis of coordinate interleaved PLC system with Rayleigh channel gain under Nakagami‐ m additive noise

Power line communication (PLC) is an emerging field of communication that makes use of the existing power line infrastructure for the transmission of data and power. However, the presence of additive and multiplicative power line noises deteriorate the performance of a PLC system. Background and impulsive noises are the two main categories of additive noise, whereas channel gain accounts for the multiplicative noise. In this study, the authors propose coordinate interleaving (CI) technique and investigate its performance for a PLC system under the combined effect of Nakagami-m additive background noise and Rayleigh channel gain. They assume perfect channel state information at the transmitter and receiver of the PLC link. By interleaving the coordinates of the symbols sent over different channel realisations, CI results in significant improvement in the performance of a PLC system. Both the simulated and analytical results shown in this study reveal the effectiveness of the proposed technique in achieving better performance.

DF-Fit: A Robust Algorithm for Detection of Crystallographic Information in Atom Probe Tomography Data

We report on a new algorithm for the detection of crystallographic information in three-dimensional, as retained in atom probe tomography (APT), with improved robustness and signal detection performance. The algorithm is underpinned by one-dimensional distribution functions (DFs), as per existing algorithms, but eliminates an unnecessary parameter as compared to current methods.By examining traditional DFs in an automated fashion in real space, rather than using Fourier transform approaches, we utilize an error metric based upon the expected value for a spatially random distribution for detecting crystallography. We show cases where the metric is able to successfully obtain orientation information, and show that it can function with high levels of additive and displacive background noise. We additionally compare this metric to Fourier transform methods, showing fewer artifacts when examining simulated datasets. An extension of the approach is used to aid the automatic detection of high-quality data regions within an entire dataset, albeit with a large increase in computational cost.This extension is demonstrated on acquired aluminum and tungsten APT datasets, and shown to be able to discern regions of the data which have relatively improved spatial data quality. Finally, this program has been made available for use in other laboratories undertaking their own analyses.

Enhancing Target Speech Based on Nonlinear Soft Masking Using a Single Acoustic Vector Sensor

Enhancing speech captured by distant microphones is a challenging task. In this study, we investigate the multichannel signal properties of the single acoustic vector sensor (AVS) to obtain the inter-sensor data ratio (ISDR) model in the time-frequency (TF) domain. Then, the monotone functions describing the relationship between the ISDRs and the direction of arrival (DOA) of the target speaker are derived. For the target speech enhancement (SE) task, the DOA of the target speaker is given, and the ISDRs are calculated. Hence, the TF components dominated by the target speech are extracted with high probability using the established monotone functions, and then, a nonlinear soft mask of the target speech is generated. As a result, a masking-based speech enhancement method is developed, which is termed the AVS-SMASK method. Extensive experiments with simulated data and recorded data have been carried out to validate the effectiveness of our proposed AVS-SMASK method in terms of suppressing spatial speech interferences and reducing the adverse impact of the additive background noise while maintaining less speech distortion. Moreover, our AVS-SMASK method is computationally inexpensive, and the AVS is of a small physical size. These merits are favorable to many applications, such as robot auditory systems.