SPEECH ENHANCEMENT USING HARMONICS REGENERATION BASED ON MULTIBAND EXCITATION

This paper proposes an algorithm that adopts the harmonic regeneration as post-processing to improve the performance of speech enhancement using traditional Short Time Spectral Amplitude (STSA).The pro...     

一种改进的AMDF求取语音基音的方法

文章在已经被广泛运用的平均幅度差函数(AMDF)的基础上,针对其通过求取"第一谷值点"会产生误差,而求取"最小谷值点"会遇到周期增倍等问题,提出了改进的AMDF方法,它通过简单的预测谷值范围、循环AMDF和线性变换等步骤实现了精确的基音检测.实验结果证明:无论是对不同的语音(男、女音),还是在不同的信噪比情况下.其检测结果都优于传统的AMDF法和LV-AMDF法.     

Perception of Pitch and Roughness in Vocal Signals with Subharmonics

Pitch and roughness were rated according to the extent of amplitude modulation (AM) and frequency modulation (FM) of a subharmonic [fundamental frequency (F0)/2]. The objective was to determine the identification boundaries for pitch and roughness and to discover how both kinds of modulation affect these boundaries. Another objective was to judge the reliability between subjects when identifying subharmonic-related pitch and roughness. Three procedures were used: ABX comparisons, method of adjustment, and rating of roughness. Results indicated that the crossover point to the lower pitch (associated with the subharmonic) occurred between 10% and 30% modulation, depending on modulation type and F0. Subjects demonstrated highly variable perceptions of pitch and roughness, with poor intersubject reliability.     

Wavelet-based cepstrum calculation

In this paper we present a new wavelet-based algorithm for low-cost computation of the cepstrum. It can be used for real time precise pitch determination in automatic speech and speaker recognition systems. Many wavelet families are examined to determine the one that works best. The results confirm the efficacy and accuracy of the proposed technique for pitch extraction.     

一种特征波提取速率自适应的波形内插语音编码方案

该文提出了一种特征波形提取速率自适应于输入语音帧特性的波形内插编码方案。基于双加权长时预测增益最大原则并利用前向基音判决实现了较为可靠的基音周期估计算法,用基音周期、浊音度和波表面平坦度决定波形提取速率以及SEW(Slowly Evolving Waveform)和REW(Rapidly Evolving Waveform)的更新速率。实验证明,该文提出的波形内插(WI)编码算法相比固定波形提取速率的WI算法在平均码率和计算复杂度上均有一定程度的降低,且合成语音质量明显优于4.8kbps的CELP语音编码算法。     

The Effect of Cricothyroid Muscle Action on the Relation Between Subglottal Pressure and Fundamental Frequency in an In Vivo Canine Model

The relation between subglottal pressure (P_s) and fundamental frequency (F₀) in phonation was investigated with an in vivo canine model. Direct muscle stimulation was used in addition to brain stimulation. This allowed the P_s-F₀ slope to be quantified in terms of cricothyroid muscle activity. Results showed that, for ranges of 0–2 mA constant current stimulation of the cricothyroid muscle, the P_s-F₀ slope ranged from 10 Hz/kPa to 60 Hz/kPa. These results were compared to similar slopes obtained in a previous study on excised larynges in which the vocal fold length was varied instead of cricothyroid activation. A physical interpretation of the P_s-F₀ slope is that the amplitude-to-length ratio of the vocal folds decreases with CT activity, resulting in a smaller time-varying stiffness. In other words, there is less dependence of F₀ on amplitude of vibration when the vocal folds are long instead of short.     

Mandarin tone recognition based on wavelet transform and hidden Markov modeling

This paper presents a method of tone recognition for Mandarin speech by using combination of wavelet transform and hidden Markov modeling techniques. A pitch detector based on singularity detection and multi-resolution analysis of wavelet transform is employed for estimation of pitch periods, and hidden Markov modeling with partition Gaussian mixtures probability density function is used for the tone recognition. The algorithm can provide recognition accuracy of 97.22% and 94.47% for speaker-dependent and speaker-independent tone recognition, respectively.     

Electrode Kinetics and Sensitivity of Nanostructured Electrodes from Different Carbon Fiber Precursor Materials

Correlation of properties of common carbon fiber precursor materials, polyacrylonitrile (PAN) and pitch, with kinetics and sensitivity of microdisk electrodes, PAN T650, PAN HCB and Pitch P25, nanostructured by an electrochemical etching method was tested. Sensitivity of PAN electrodes is higher but kinetics are slower than at pitch electrodes due to more defects in PAN. Because of more surface oxides at PAN T650 adsorption of dopamine is greater than at PAN HCB electrodes but for uric acid adsorption is greater at PAN HCB. Uric acid sensitivity is related to high conductivity of PAN fibers, with an inverse relationship observed for dopamine.     

脉冲MPM小型化行波管螺旋线慢波结构的设计和改进

论述了脉冲MPM用小型化行波管慢波结构的计算设计和定型过程。以CAD软件为工具,以实测数据为依据,达到设计预期。