115 research outputs found
Study of Urban Atmosphere Harmful Substances Adsorption into Cement
No full textAbstractThe ecological safety issues of building materials are the most important area nowadays. The impact of different substances and materials on human organism is the major point for emission control of building industry enterprises. One of the main factors can be building dust that could be presented as cement. For that purpose the qualitative and quantitative composition of samples and its adsorption capacity should be known. The article presents a comparative analysis of dust and cement. Quantitative characteristics of adsorption activity of different branded cements (CEM-I 42.5N, CEM-II/A-SH 42.5N, PC 500-D0-N) are shown. The impact of atmosphere humidity on static adsorption capacity of selected samples at specified temperature was estimated
Improving the Computational Model for Approximation of Particle Functions over Diameter of Dust in the Work Area and at the Border of the Sanitary Protection Zone
No full textAbstractThis paper describes the basic ways of solving the problem of experimental data approximation. Disperse composition of crushed gypsum dust was analyzed; the analysis results are presented herein. Approximating function of particulate dust composition function is a piecewise function, defined as a three-tier spline “direct-parabola-hyperbole.” The approximation aims at finding seven function factors and two nodal points. The least square method was used to estimate the unknown parameters of regression models for the sample data. To apply this method to experimental data, a program complex calculation models for the approximation of the integral representation of the mass distribution function of particles in the dust diameter in CAS Maple was presented. It defines a function that describes the distribution of the particulate composition of the dust released from the open warehouse storage of crushed gypsum rock with the smallest error
Environmental and Working Area Dust Emission from the Gypsum Warehouse
No full textAbstractThe work analyses the emissions analysis for gypsum binder production. The major air pollutant is the suspended solids composed of a mixture of particles in air, which can be either solid or liquid and be a complex mixture of organic and inorganic substances. Studies have shown that excessive concentrations of inorganic dust (gypsum dust) generated by the open gypsum warehouse storage and thus the contribution of emission sources into total concentration of inorganic dust (dust plaster) average vary from 66,28% to 87,67%. The highest excessive concentrations of inorganic dust are registered at the border of regulated areas: up to 20% SiO2 (plaster dust) (from 5,9 to 16,0 shares of maximum allowable concentration (MAC))
Monitoring of Fine Particulate Air Pollution as a Factor in Urban Planning Decisions
No full textAbstractThe article substantiates the importance of assessing air pollution when making town-planning decisions. It focuses on the composition of fine dust. The article considers the issues of describing dust concentration using the theory of stationary random functions. Such approach allows not only to obtain the characteristics of the particulate composition of dust in the air, but also to determine a number of additional parameters, namely, mean residence time of fractional concentration above the predetermined level, the average number of times when fractional concentration per time unit exceeds the standard. The article offers the methods and calculations of probability of dust concentration exceeding hygienic standards by the example of three districts of Volgograd. The results of these calculations are presented. Based on this research, several industries in the city received recommendations to reduce their emissions into the atmosphere
AN EFFICIENT SPEECH GENERATIVE MODEL BASED ON DETERMINISTIC/STOCHASTIC SEPARATION OF SPECTRAL ENVELOPES
No full textThe paper presents a speech generative model that provides an efficient way of generating speech waveform from its amplitude spectral envelopes. The model is based on hybrid speech representation that includes deterministic (harmonic) and stochastic (noise) components. The main idea behind the approach originates from the fact that speech signal has a determined spectral structure that is statistically bound with deterministic/stochastic energy distribution in the spectrum. The performance of the model is evaluated using an experimental low-bitrate wide-band speech coder. The quality of reconstructed speech is evaluated using objective and subjective methods. Two objective quality characteristics were calculated: Modified Bark Spectral Distortion (MBSD) and Perceptual Evaluation of Speech Quality (PESQ). Narrow-band and wide-band versions of the proposed solution were compared with MELP (Mixed Excitation Linear Prediction) speech coder and AMR (Adaptive Multi-Rate) speech coder, respectively. The speech base of two female and two male speakers were used for testing. The performed tests show that overall performance of the proposed approach is speaker-dependent and it is better for male voices. Supposedly, this difference indicates the influence of pitch highness on separation accuracy. In that way, using the proposed approach in experimental speech compression system provides decent MBSD values and comparable PESQ values with AMR speech coder at 6,6 kbit/s. Additional subjective listening testsdemonstrate that the implemented coding system retains phonetic content and speaker’s identity. It proves consistency of the proposed approach.The paper presents a speech generative model that provides an efficient way of generating speech waveform from its amplitude spectral envelopes. The model is based on hybrid speech representation that includes deterministic (harmonic) and stochastic (noise) components. The main idea behind the approach originates from the fact that speech signal has a determined spectral structure that is statistically bound with deterministic/stochastic energy distribution in the spectrum. The performance of the model is evaluated using an experimental low-bitrate wide-band speech coder. The quality of reconstructed speech is evaluated using objective and subjective methods. Two objective quality characteristics were calculated: Modified Bark Spectral Distortion (MBSD) and Perceptual Evaluation of Speech Quality (PESQ). Narrow-band and wide-band versions of the proposed solution were compared with MELP (Mixed Excitation Linear Prediction) speech coder and AMR (Adaptive Multi-Rate) speech coder, respectively. The speech base of two female and two male speakers were used for testing. The performed tests show that overall performance of the proposed approach is speaker-dependent and it is better for male voices. Supposedly, this difference indicates the influence of pitch highness on separation accuracy. In that way, using the proposed approach in experimental speech compression system provides decent MBSD values and comparable PESQ values with AMR speech coder at 6,6 kbit/s. Additional subjective listening testsdemonstrate that the implemented coding system retains phonetic content and speaker’s identity. It proves consistency of the proposed approach
Convolutional neural network with semantically meaningful activations for speech analysis
No full textSemantic analysis of speech is more prospective
compared to analysis of text since speech contains more in-
formation that is important for understanding. The most im-
portant distinguishing feature of speech is intonation, which
is inaccessible in the text analysis. For successful semantic
analysis of speech it is necessary to transform the speech signal
into features with semantic interpretation. The mathematical
apparatus of convolutional neural networks (CNN) seems suitable
to implement this kind of transformation. However there is a
scalability problem that makes it hard to combine many CNN’s in
a single solution. To overcome this we propose to develop a CNN
model with semantically meaningful activations i.e. the model
that is capable of semantic interpretation of its internal states.
The ultimate goal of the transform is to extract all semantically
meaningful information, however the present work is confined to
voice activity detection (VAD) and intonation extraction. Unlike
other VADs based on artificial neural networks, the proposed
model does not require a lot of computing resources and has a
comparable or even better performance
Spectrum estimation of speech: coding and feature extraction
No full textSpeech analysis and spectrum estimation has been the fundamental problem of digital signal processing for recent decades. The problem still has a huge practical impact on modern speech processing applications
that involve coding and deep learning. The paper reviews the main speech spectral estimation techniques including
linear prediction and cepstrum
Voice activity detection in noisy conditions using tiny convolutional neural network
No full textThe paper investigates the problem of voice activity detection from a noisy sound signal. An extremely compact convolutional neural network is proposed. The model has only 385 trainable parameters. Proposed model doesn’t require a lot of computational resources that allows to use it as part of the “internet of things” concept for compact low power devices. At the same time the model provides state of the art results in voice activity detection in terms of detection accuracy. The properties of the model are achieved by using a special convolutional layer that considers the harmonic structure of vocal speech. This layer also eliminates redundancy of the model because it has invariance to changes of fundamental frequency. The model performance is evaluated in various noise conditions with different signal-to-noise ratios. The results show that the proposed model provides higher accuracy compared to voice activity detection model from the WebRTC framework by Google
Выделение речевой активности на фоне шумов при помощи компактной сверточной нейронной сети
No full textThe paper investigates the problem of voice activity detection from a noisy sound signal. An extremely compact convolutional neural network is proposed. The model has only 385 trainable parameters. Proposed model doesn’t require a lot of computational resources that allows to use it as part of the “internet of things” concept for compact low power devices. At the same time the model provides state of the art results in voice activity detection in terms of detection accuracy. The properties of the model are achieved by using a special convolutional layer that considers the harmonic structure of vocal speech. This layer also eliminates redundancy of the model because it has invariance to changes of fundamental frequency. The model performance is evaluated in various noise conditions with different signal-to-noise ratios. The results show that the proposed model provides higher accuracy compared to voice activity detection model from the WebRTC framework by Google.Исследуется задача выделения речевой активности из зашумленного звукового сигнала. Предлагается компактная модель сверточной нейронной сети, которая имеет всего 385 параметров. Модель нетребовательна к вычислительным ресурсам, что позволяет использовать ее в рамках концепции Интернета вещей для портативных устройств с низким энергопотреблением. В то же время эта модель обеспечивает высокую точность определения речевой активности на уровне лучших современных аналогов. Указанные полезные свойства достигаются путем применения специального сверточного слоя, учитывающего гармоническую структуру вокализованной речи и устраняющего избыточность модели за счет инвариантности к изменениям частоты основного тона. В рамках экспериментов производительность модели оценивалась в различных шумовых условиях для разных соотношений сигнала и шума. Результаты экспериментов показали, что предложенная модель обеспечивает более высокую точность определения речевой активности по сравнению с моделью, представленной компанией Google в фреймворке WebRTC
Amorphization of crystalline Si due to heavy and light ion irradiation
No full textThe formation of amorphous silicon in crystalline silicon by bombardment with light (Si) and heavy (Xe) ions has been investigated by transmission electron microscopy with in situ ion irradiation. Experiments have been carried out at room temperature and low temperature (50 K) and the results are compared to a simple numerical model for amorphization. The results indicate that the amorphization mechanisms for both irradiations are heterogeneous in nature and that numerous overlaps of the collision cascade are generally required to render the crystal amorphous. Following from this, the nature of the material within the confines of collision cascades will be discussed and it will be shown that the individual cascade volume is not necessarily amorphous as previously described in the scientific literature but contains varying degrees of damage depending on the energy deposited within the cascade
- …
