Speech Accent Profiles: Modeling and Synthesis

This article provides a brief introduction to speech accents and introduces a set of statistical signal processing methods for the modeling, analysis, synthesis, and morphing of English language accents. Accent morphing is the changing of the accent of a speech to a different accent, for example, from British to American.     

A new multi-level trust management framework (MLTM) for solving the invalidity and sparse problems of user feedback ratings in cloud environments

Choosing a trusted cloud service provider (CSP) is a major challenge for cloud users (CUs) in the cloud environment, as many CSPs offer cloud services (CSs) with the same functionality. Trust evaluation of CSPs is often based on information from quality of service (QoS) monitoring and CUs’ feedback ratings. Despite the volume of feedback ratings received in trust management systems, the quality of feedback storage is very low, as many CUs do not send their feedback ratings when using CSs. Additionally, a percentage of existing feedback ratings may not be valid, since some malicious CUs send unfair feedback ratings to change the trust evaluation results. As these lead to poor data quality, the accuracy of trust evaluation results might be affected. To overcome these limitations, this paper proposes a new multi-level trust management framework, which completes previous frameworks by defining new components to improve the data quality of feedback storage. In our framework, new components were defined to solve the invalidity and sparse problems of feedback storage. Certainly, the trust assessment of CSP would be more accurate based on high-quality feedback ratings. The performance of the MLTM was evaluated using two different datasets based on a real Quality of Web Services dataset (QWS) and an artificial data set (Cloud-Armor), whose quality was reduced for the purpose of this study. Analytical values revealed that our proposed approach significantly outperformed other approaches even with the poor data quality of feedback storage.

     

Interpolation of Lost Speech Segments Using LP-HNM Model With Codebook Post-Processing

This paper presents a method for interpolation of lost speech segments. The interpolation method can be used for packet loss concealment in voice communication over mobile phones, for voice over IP or for restoration of lost segments in speech recordings. The interpolation method employs a combination of a linear prediction (LP) model of the spectral envelope and a harmonic noise model (HNM) of the excitation of speech. The speech interpolation problem is transformed to the modeling and interpolation of the trajectories of LP parameters and the amplitude, phase and harmonicity of HNM tracks of speech excitation. In particular, the interpolation of harmonicity results in a smooth transition from voiced to unvoiced speech and vice versa. Crucially, the proposed interpolation method does not suffer from the consequences of zero-excitation of conventional autoregressive (AR) interpolation. Different combinations of linear and autoregressive interpolation methods are evaluated for the estimation of the time-varying parameters of LP-HNM tracks. Furthermore, a post-processing codebook mapping, employed to enhance the interpolation of the spectral envelope of speech, results in improved output quality for longer length speech gaps. For different packet loss rates and patterns of distributions of missing speech gaps, the proposed interpolation methods are evaluated and compared with popular AR-based interpolation methods and the speech packet recovery method specified in the ITU G.711 standard, as a reference. The evaluation results show that the proposed methods substantially improve the restoration of formants and harmonic tracks and consistently results in significant performance gain and improved perceptual quality of speech.     

Hidden Markov models with duration-dependent state transition probabilities (speech recognition)

A new method is proposed for incorporation of duration knowledge in the form of duration-dependent state transition probabilities in a left-right hidden Markov model. Duration-dependent transition probabilities are derived from integration of histograms of the state durations. The model re-estimation process becomes one of obtaining a new segmentation from which a new set of state and observation probabilities are derived.<>     

Constrained thickness optimization of rectangular orthotropic fiber-reinforced plate for fundamental frequency maximization

This paper addresses the thickness optimization of rectangular orthotropic fiber-reinforced composite plates by using mathematical programming theory of nonlinear constrained optimization. We optimize the thickness distribution of various composite plates for reaching a maximum in the plate natural frequency, subject to an equality constraint in a plate volume and a number of inequality constraints on the lower and upper bounds of allowable range of thicknesses. For the modeling part of our optimization algorithm, we formulate the kinematical hypothesis of the classical plate theory (CPT), reduced plane stress theory of linearized elasticity, and Hamilton’s principle. For the solution part, we implement the Rayleigh-Ritz semi-analytical solution technique to spatially discretize the weak form of the plate partial differential equation (PDE), and transforming it to a generalized algebraic eigen-problem. For the design optimization part, we use a conjugate Rosen’s gradient projection method. The design variables of the optimization design space are thicknesses of an arbitrary but rectangular partition of the plate into constant-thickness subdivisions. After simulating our algorithm, we noticed many interesting and impressive results, and tried to justify them. The most impressive finding is that, as the fiber orientation of the composite plate rotates, the plate thickness distribution rotates also in the same way, but with different manifestations for different boundary conditions. For parametric studies, we analyze the sensitivity of the algorithm against parameters, like, the degree of fineness in the number of plate partitions into constant-thickness sub-plates, the degree of plate anisotropy, the fiber orientation, the rectangular plate aspect ratio, the upper to lower bounds of allowable thickness ratio, and boundary conditions.     

Electronic behavior of native oxide films on Ti and TiN during 90-day immersion in electrolytes with different pH levels

This work presents the electronic behavior of Ti and TiN thin films when exposed to electrolytes with pH levels of 2,7 and 13 for 90 days.Staircase potentio-electrochemical impedance spectroscopy tests were performed on the 100-nm Ti and TiN monolithic films,and MottSchottky analysis of these tests was used to determine the films' semiconductive behavior and changes in the donor/acceptor density.In addition,the flat-band potential of each film's surface oxide was also characterized.No attempt was made to control oxide formation,and therefore,these tests reflected the native surfaces of these films.While the TiN films exhibited n-type semiconductivity in all electrolytes,the Ti films only showed n-type behavior in the acidic(pH=2) and neutral(pH=7) electrolytes.The semiconductivity of the Ti films transitioned to p-type during exposure to the basic electrolyte(pH=13) after reaching 60 days.Furthermore,there was a significant increase in the donor densities for both Ti and TiN films when immersed in the basic electrolyte relative to the acidic and neutral electrolytes.     

Engineered Cobalt Single-Atoms@BiFeO3 Heteronanostructures for Highly Efficient Solar Water Oxidation

Efficient charge–carrier separation and their utilization are the key factors in overcoming sluggish four-electron reaction kinetics involved in photocatalytic oxygen evolution. Here, a novel study demonstrates the significance of Na₂S₂O₈ as a sacrificial agent in comparison to AgNO₃. Resultantly, BiFeO₃ (BFO) and titanium doped-oxygen deficient BiFeO₃ (Ti-BFO-R) nanostructures achieve ≈64 and 44.5 times higher O₂ evolution in the presence of Na₂S₂O₈ compared to AgNO₃ as a sacrificial agent, respectively. Furthermore, the presence of Co single atoms (Co-SAs) deposited via immersion method on BFO and Ti-BFO-R nanostructures led to achieving outstanding O₂ evolution at a rate of 16.11 and 23.89 mmol g⁻¹ h⁻¹, respectively, which is 153 and 227.5 times higher compared to BFO (in the presence of AgNO₃), the highest O₂ evolution observed for BFO-based materials to date. The successful deposition of Co-SAs is confirmed by aberration-corrected high-angle annular dark-field scanning transmission electron microscopy (AC HAADF-STEM) and X-ray absorption near-edge structure (XANES). The charge transfer investigations confirm the significance of Co-SAs on BFO-based photocatalysts for improved charge–carrier separation, transport, and utilization. This novel study validates the excellent role of Na₂S₂O₈ as a sacrificial agent and Co-SAs as a cocatalyst for BFO-based nanostructures for efficient O₂ evolution.