语音

This thesis can be divided into two parts .One is the endpoint detection of noisy speech signals.

本论文根据内容可以分为两个部分：低信噪比下的语音端点检测和语音增强。

In this paper, a new algorithm based on spectrogram for speech endpoint detection is presented.

利用语音在语谱图中表现出的不同特征，提出了一种基于语谱图的语音端点检测算法。

This paper looks into the transliteration of Mandarin loanwords of English person names from both phonological and perceptual perspectives and concludes that such strategies as epenthesis, deletion and feature change are employed to satisfy Mandarin phonology, with epenthesis being the major strategy, and that epenthesis involves consonant assimilation, vowel harmony and default vowel.

文章从音系和语音感知的角度分析汉语中英语人名的音译过程，研究发现：为了符合汉语的音节结构特点，音译过程采用了增音、删音和变音方式，且尤以增音为主；所添元音呈现一定的语音规律，即辅音同化、元音和谐和元音默认。

On the base of the acoustics , a whole linear time - varying discrete speech produce model is established through anatomising the factors including track , driving source and lip eradiation the mechanism of speech in this paper , and we draw a conclusion that the time between glottal closure is pitch period

本文从声学理论出发，剖析了语音产生的机理，综合考虑声道、激励源和嘴唇辐射三方面的因素，建立了一个完整的、线性的和时变的语音产生的离散系统模型，得出两次声门闭合事件之间的时间间隔就是基音周期的结论。

A technology of VAD, in practice, plays an essential role in noise spectrum estimator. The VAD scheme is frequently employed into noise spectrum estimator as an indicator of updating noise spectrum.

事实上，语音侦测技术使也在噪音估测器中扮演非常重要的角色；一般都采用语音侦测系统的技术作为判断何时追踪噪音频谱变动的指示器。

In the meantime, because the material recited from memory is meaningful dialogue or short composition, so, the speech in the social intercourse isn't only euphonic and then cans get a practice, but is the speech of only isolated phrase language is euphonic.

同时， 由于背诵的材料是有意义的对话或短文，这样，交际中的语音语调便可得到练习，而不仅仅是孤立词语的语音语调了。

In this paper, firstly, some basic signal process are discussed, such as signal fil...

论文中首先对语音信号的基本处理问题进行了分析和对比，然后在自己设计的基于TMS320VC5402的DSP实际系统上，进行了语音处理过程的滤波、采样、傅立叶变换和谱包络提取的算法实现研究，讨论了在算法的DSP实现方法，分析了运行实验结果。

The other filiation of morden communication is sound communication, the purpose of sound communication is enable immediate、no error、no interrupt communication.

现代通信的另外一个分支就是语音通信，语音通信要达到的目的是能够实时地，无差错、无干扰地通信。

So the principal component analysis approach is employed to compress the MCF and other fixed dimensional feature vector, an improved algorithm is provided, which has faster training speed than the original one.

为此我们利用神经网络主分量分析的方法，给出了语音特征压缩的具体算法，成功地解决了语音识别特征的降维问题。

The new method employs Viterbi algorithm and HMM parameters to produce the best state sequence and then to warp the time within the same state segment using function approximation technique and intersection technique. So a fixed dimensional feature vector can be got. It is then sent to the RBF neural network.

新方法首先利用隐马尔可夫模型参数和Viterbi算法产生语音信号的最佳时间序列，然后利用函数逼近和插值的方法进行时间规正，组成等维的语音识别特征；最后送径向基函数神经网络进行识别。

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％。