语音

Abstraction :This paper analyzes the principle of Butterworth digital filter,discuss bilinear transformation method of filter design;record a piece of voice by Microsoft voice recorders, carry on its a time-domain analysis, spectral analysis;then design a specified Butterworth band-stop filter based on the theory of bilinear transform,analyze the filter by the tool named Fvtool;and filter the voice signal acquired by voice recorder.finally compare its characteristics of frequency domain and time domain,study the applications of Butterworth filter in digital signal processing ,in particular in signal denoising.

本文分析了巴特沃思数字滤波器的原理，探讨了双线性变换法设计滤波器的方法；用微软录音机采集一段语音，对其进行了时域分析，频谱分析；然后使用双线性变换法设计了一个的巴特沃思带阻滤波器，并且使用fvtool对设计的滤波器进行分析；然后利用该滤波器对采集的语音信号进行滤波。最后对比滤波前后的语音信号的时域和频域特性，研究了巴特沃思滤波器在信号处理尤其是信号滤波去噪方面的特性和应用。

It programmes in order to extracting the characteristic parameter and carrying out speech recognition.

语音是系统设计的基础和关键所在，首先阐述语音信号产生的数学模型、特征参数和提取方法以及语音识别的算法等。

Using covariation concept in array signal processing, the proposed algorithm can obtain clean speech signal subspace by eigendecomposition of the covariation coefficient matrix of speech signal corrupted by impulsive noise.

该算法把阵列信号处理中的协变异应用到语音增强中，对带噪语音信号的协变异系数矩阵进行特征值分解得到纯净语音信号的子空间。

VQ is easy and tend to realize, HMM can comminute the speech dynamically, it can embody not only the dynamic characteristic but also the stationary characteristic of the speech. GMM is a special HMM, it uses combination of many weighted Gaussian density function to approach the distributing density of speaker characteristic vector in characteristic space, so it is also a good way.

VQ方法是最简单又易于实现，HMM方法能将语音动态分割，既能体现语音的动态特征，又能体现语音的平稳特性，GMM用多个加权高斯密度函数的组合来逼近说话人特征矢量在特征空间的分布密度，收到了较好的结果。

The paper, based on the errors of pronunciation on the level of segmental phonemes by students of Beijing University of Physical Education,isolated and discussed the most difficult phonemes for students to acquire through a contrastive study of the consonantal and vowel phonemes in English and Chinese Putonghua .

在调查北京体育大学学生英语语音发育错误的基础上，分析学生在音素水平上的语音错误以及受汉语影响较大的语音难点，提出了纠正错误发音的方法，并就体育大学学生的基础英语教学提出了个人的建议。

In this method, we use the pooled correlogram based on human auditory model to extract the pitch of speech. An unsupervised lateral inhibitory network is used to get the peak position, which simulates the lateral inhibitory phenomenon in human auditory system.

采用基于人耳听觉模型的相关图来提取语音信号的基频，运用无监督的侧抑制神经网络来模拟人耳侧抑制属性进行基频检测，为了克服在低信噪比情况下侧抑制神经网络的误判问题，引入了相邻语音帧的语音基频的帧间约束。

With the growth of the WAS and the user amount, the presentVSPs could not meet the demands, and began to cumber the further developmentof the WAS.

随着语音增值业务的发展，用户数量的增加，现有的语音增值业务平台已经不能满足需求，变成了阻碍语音增值业务进一步发展的因素。

Experiments on LVCSR test showed that the method can achieve nearly 20% relative error reductions compared to the traditional diagonal gaussian modeling method and cause less computation cost during decoding.

在汉语普通话连续语音测试和广播语音的连续语音测试中，本文提出的相关性特征建模方法在增加较少的内存占用量和解码计算量的情况下，能够比采用传统的具有对角方差结构的高斯混合模型的建模方法获得20％的相对误识率的降低。

By analyzing the current status of speech coding, it proposes that low bit rate speech coding is hot off the press.

首先通过分析当前语音编码的现状，提出了低速率语音混合编码是现代语音编码研究的热点。

So he and his colleagues have devised a system based on what is known as "voice extensible markup language", a cousin of the hypertext markup language used on conventional websites, that allows a website to be built and operated more or less by voice alone.

为此，他和他的同事们一起设计了一种基于"语音可拓展标记语言"的系统。语音拓展标记语言与传统网站中应用的超文本标记语言十分类似，它使得对网站的建设和操作通过语音得以实现。

Breath, muscle contraction of the buttocks; arch body, as far as possible to hold his head, right leg straight towards the ceiling (peg-leg knee in order to avoid muscle tension).

呼气，收缩臀部肌肉；拱起身体，尽量抬起头来，右腿伸直朝向天花板（膝微屈，以避免肌肉紧张）。

The cost of moving grain food products was unchanged from May, but year over year are up 8%.

粮食产品的运输费用与5月份相比没有变化，但却比去年同期高8％。