精细结构常数及其物理意义

精细结构常数是基本物理常数之一,阐述了精细结构常数的由来,揭示了在不同情况下该常数表现出来的不同物理本质.     

听觉计算模型在鲁棒性语音识别中的应用

利用听觉感知机理,建立一个基于听觉感知机理的语音信号特征提取模型。本文由两部分组成,一部分是在传统听觉计算模型基础上提出听觉倒谱特征AFCC(AnditoryFrequencyCepstralCoefficient)的提取方法,这样既压缩了特征维数,减小计算量,又使各个特征维之间相互独立,满足HMM模型的要求。并且根据听觉神经中枢的长时整合特性,文中提出了用低通滤波模型来模拟这种功能。结合该低通模型,提取的语音信号的听觉倒谱特征在HMM框架下取得较好的鲁律性。另一部分在研究听觉侧抑制机理的基础上,提出一个简单有效的听觉侧抑制处理模型。美尔倒谱特征MFCC谱特征经过该侧抑制模型处理,得到侧抑制美倒谱特征MFCCI,实验表明,该新特征MFCCI鲁棒性能比MFCC有大大提高。听觉倒谱特征AFCC经过该侧抑制处理得到侧抑制听觉倒谱特征AFCCI,实验表明,该新特征AFCCI鲁律性能比AFCC有大大提高。     

汉语语音的非线性动力学特征及其降噪应用

分析了汉语语音的相关维、最小嵌入维数以及重构相图。分析结果表明汉语语音具有混沌特征。根据这些非线性特征可以有效区分汉语中的浊音、清音和随机噪声,从而可以用于语音降噪。介绍了本地投影法混沌语音降噪的原理与算法,并利用该算法对一些典型的元音和辅音进行降噪,获得了较好的降噪效果。     

Fine structure of dark matter halos and its effect on terrestrial detection experiments

Terrestrial dark matter detection experiments probe the velocity-space distribution of dark matter particles in the vicinity of the Earth. We present a novel method, to be used in conjunction with standard cosmological simulations of hierarchical clustering, that allows one to extract a truly local velocity-space distribution in exquisite detail. Preliminary results suggest a new picture for this distribution which is decidedly non-Maxwellian but instead is characterized by randomly positioned peaks in velocity space. We discuss the implications of these results for experiments to detect axions and weakly interacting massive particles.     

Fine structure of strings

We show that the critical behavior of the string tension and most other important qualitative properties of strings are driven by a term in the action containing extrinsic curvatures, which is necessary to add to the Nambu-Goto term. Depending on the renormalization-group structure, strings can be “creased” (with non-zero surface tension at the critical point) or “smooth”. Candidates for both possibilities and their physical relevance are discussed.     

A recursive calculating algorithm for higher-order cumulants over sliding window and its application in speech endpoint detection

Regarding the performance of traditional endpoint detection algorithms degrades as the environment noise level increases,a recursive calculating algorithm for higher-order cumulants over a sliding window is proposed.Then it is applied to the speech endpoint detection.Furthermore,endpoint detection is carried out with the feature of energy.Experimental results show that both the computational efficiency and the robustness against noise of the proposed algorithm are improved remarkably compared with traditional algorithm.The average probability of correct point detection(Pc-point) of the proposed voice activity detection(VAD) is6.07%higher than that of G.729 b VAD in different noisy at different signal-noise ratios(SNRs)environments.     

滑动窗累积量的递推估计算法及其在语音端点检测中的应用

提出了一种滑动窗累积量的递推估计算法并应用于语音端点检测中,用以解决传统端点检测方法在噪声环境下检测性能变差的问题。在对含噪语音信号进行加窗之后,利用滑动窗累积量的递推估计算法估计含噪语音信号的高阶累积量值,并在此基础上结合能量特征进行语音端点检测。实验结果表明,所提滑动窗累积量递推估计算法相比较传统高阶累积量计算方法运算效率明显提高;所提端点检测算法在不同噪声和信噪比环境下相比较G.729b算法点正确率Pc-point值平均提升了6.07%。基于滑动窗高阶累积量的语音端点检测算法具有较高的运算效率及良好的鲁棒性。     

滑动窗累积量的递推估计算法及其在语音端点检测中的应用

提出了一种滑动窗累积量的递推估计算法并应用于语音端点检测中,用以解决传统端点检测方法在噪声环境下检测性能变差的问题。在对含噪语音信号进行加窗之后,利用滑动窗累积量的递推估计算法估计含噪语音信号的高阶累积量值,并在此基础上结合能量特征进行语音端点检测。实验结果表明,所提滑动窗累积量递推估计算法相比较传统高阶累积量计算方法运算效率明显提高;所提端点检测算法在不同噪声和信噪比环境下相比较G.729b算法点正确率Pc-point值平均提升了6.07%。基于滑动窗高阶累积量的语音端点检测算法具有较高的运算效率及良好的鲁棒性。     

Fine structure of analogue states

This article reviews all the high-resolution data on fragmented analogue states taken at the Triangle Universities Nuclear Laboratory, over a period of ten years. There are fifty analogue states observed by proton scattering on targets of masses between 40 and 64, and mass 92. The number of fragments may be only two or three, or as many as fifty. Of the total, only 17 states have a sufficient fine structure pattern where an analysis is attempted. In a few cases, inelastic widths and photon and neutron widths are observed besides proton elastic ones. The article discusses the optimum method of analysis of the data with a view to extracting the physical parameters of the analogue: the energy, proton spectroscopic factor, spreading width, shift and asymetry parameter. The last three quantities are discussed from the viewpoint of the Robson model based on one-channel external mixing of the analogue by Coulomb forces. This model has qualitative success, particularly in describing the large asymmetry sometimes seen, but it is not quantitatively adequate. When the proton spectroscopic factors are compared to the neutron counterparts for the parent (as measured by (d,p) studies), after allowing for Coulomb effects, the proton ones are smaller by an amount which increases to 30% in the heavier nuclei (Ni,Mo). This is in line with the situation on ^{2 0 8}Pb. Thus it seems that the inadequacy of a Coulomb explanation of the analogue-parent energy shift (the Nolen-Schiffer anomaly) has a counterpart in the spectroscopic factors.     

Fine structure of indirect exciton in GaP

Wavelenght modulation spectra of the 1s indirect exciton in GaP are reported. Four lines are found which can be easily explained taking into account the “camel's back” structure of the conduction band. An exciton splitting of 0.9 meV^- ±0.2 has been measured.     

Fine and hyperfine structure of antiprotonic helium

We consider the fine and hyperfine splitting of the energy levels of antiprotonic helium pHe⁺ in highly excited rotational states in connection with the interpretation of the recently discovered delayed annihilation of antiprotons in helium, based on the Condo model.     

Fine structure of two-photon absorption in CdS

We have observed the fine structure of the P-excitons (A, n = 2) in Cds through two-photon absorption. The strengths of absorption peaks of P-excitons and their dependence on the polarization of the light agree with predictions of a four-band model of two-photon absorption. Additional two-photon absorption peaks were identified as phasematched two-step two-photon processes involving second harmonic polarizations.     

Fractal dimensions of speech sounds: computation and application to automatic speech recognition

The dynamics of airflow during speech production may often result in some small or large degree of turbulence. In this paper, the geometry of speech turbulence as reflected in the fragmentation of the time signal is quantified by using fractal models. An efficient algorithm for estimating the short-time fractal dimension of speech signals based on multiscale morphological filtering is described, and its potential for speech segmentation and phonetic classification discussed. Also reported are experimental results on using the short-time fractal dimension of speech signals at multiple scales as additional features in an automatic speech-recognition system using hidden Markov models, which provide a modest improvement in speech-recognition performance.     

Learning the hidden structure of speech

In the work described here, the backpropagation neural network learning procedure is applied to the analysis and recognition of speech. This procedure takes a set of input/output pattern pairs and attempts to learn their functional relationship; it develops the necessary representational features during the course of learning. A series of computer simulation studies was carried out to assess the ability of these networks to accurately label sounds, to learn to recognize sounds without labels, and to learn feature representations of continuous speech. These studies demonstrated that the networks can learn to label presegmented test tokens with accuracies of up to 95%. Networks trained on segmented sounds using a strategy that requires no external labels were able to recognize and delineate sounds in continuous speech. These networks developed rich internal representations that included units which corresponded to such traditional distinctions as vowels and consonants, as well as units that were sensitive to novel and nonstandard features. Networks trained on a large corpus of unsegmented, continuous speech without labels also developed interesting feature representations, which may be useful in both segmentation and label learning. The results of these studies, while preliminary, demonstrate that backpropagation learning can be used with complex, natural data to identify a feature structure that can serve as the basis for both analysis and nontrivial pattern recognition.     

Fine structure and hyperfine structure of some excited states of helium

The fine structure and the hyperfine structure for some singly-excited and doubly-excited states of helium atom are calculated using Rayleigh-Ritz variational method with multi-configuration-interaction wave functions. The calculated results of the fine structure for the Rydberg series are in good agreement with other theoretical and experimental data. The hyperfine parameters and the hyperfine coupling constants of ³He are also obtained for this system.Received: 17 January 2003, Published online: 29 July 2003PACS: 32.10.Fn Fine and hyperfine structure - 31.30.Gs Hyperfine interactions and isotope effects, Jahn-Teller effect     

Guided-mode-leaky-mode-guided-mode fiber structure and its application to high refractive index sensing

We propose a fiber structure based on a single-mode-multimode-single-mode one. A theoretical analysis model based on leaky mode expansion along with a new concept of leaky mode interference is presented to facilitate and enlighten the analysis for the leaky structure. Then we focus on transmission characteristics of the structure in response to surrounding refractive indices (RI) that are higher than that of the silica fiber. In our simulation, output intensity of the structure increases monotonically for RI from 1.46 to 1.55, and changes over 2.0 dB for a change of 0.01 in RI range from 1.46 to 1.48, which is well verified by the following experiments.     

Interferometric grating method and its application in Fe-base shape memory alloy structure design

This paper presents an interferometric grating method used in measuring strain fields on a curved surface. This method can be used to determine the small and large strains with high sensitivity and has been applied successfully in Fe-base shape memory alloy (FSMA) structure design. In this study, five diffracted beam from the specimen surface produce the interferometric gratings through an optical system. Using image processing technique (fast Fourier transform with special interpolation and phase shifting technique), we have obtained the strain fields of outer surface of FSMA joint and contact pressure distribution on its inside surface which has mechanical deformation and transformation.