语音

One of the goals of speech enhancement is to improve the quality and understandability of speech.

语音增强是一种信号的选择性处理技术，主要解决从受到不同方式污染的语音信号中，提取出尽可能纯净的目标语音信号的问题。

In DHMM-based speech recognition, vector quantization permits the measure to phonetic similarity among speech representations to be shared if the acoustic distance among them is below some given threshold.

在基于DHMM语音识别中，若语音帧声学特征间的距离测度在给定的范围内，矢量量化为这些语音帧共享相同的语音学相似测度提供了可能。

The current second-generation and third-generation (3G) mobile communication systems operate with narrow audio bandwidth limited to 200-3400Hz.As wireless systems are evolving from voice-telephony dominated services to multimedia and high-speed data services, the introduction of a wider bandwidth of 50–7000 Hz speech codec which provides substantially improved speech quality and naturalness is quite necessary. The Adaptive Multi-rate Wideband speech codec which was selected by 3GPP/ETSI and ITU-T can provide natural, present and comfortable codec speech and have higher intelligence and robustness in wireless systems. A wide range of applications are envisioned for AMR-WB, including ISDN wideband telephony and audio/video teleconferencing,Voice over IP and Internet applications such as IP video conferencing, voice mail, voice chat, broadcast, and voice streaming.

随着现代无线通信特别是多媒体通信的发展，人们对语音通信的质量要求越来越高。3GPP/ETSI TS026和ITU-T 722.2标准确定的自适应宽带语音编解码器AMR-WB具有的低比特率、互操作性和高语音质量，使之不仅能够应用于无线通信领域，而且可广泛应用于其他各种通信系统和平台，包括：VoIP、因特网、PSTN高质量音频会议、音频流、ISDN宽带电话、ISDN可视电话和电视会议等。

It is based on the world's leading Chinese voice synthesis technology development, apply to the M218A + voice e-book software and allows users to not only read books through the screen, but also voice in the form of "reading" novels and other texts.

它是基于全球领先的中文语音合成技术开发的，适用于M218A+的语音电子书软件，让用户不仅可以通过屏幕阅读书籍，更可以语音的形式"阅读"小说等文本内容。

Unlike speech recognition, voiceprint recognition distinguishes speaker from others by his or her voice, not the meaning by voice.

与常规意义上的语音识别不同，声纹识别不是分辨语音信号所代表的文字含义，而是分辨语音信号所代表的说话人的身份。

Focus on using the spectrogram to recognize. This paper completed the sampling, quantization and preprocessing of the voice signals, and extract characteristic parameters and map spectrogram by MATLAB 6.5. Besides, it also did some further analysis and verification on voice samples recorded by ourselves. Finally, there is an actual cases of voiceprint recognition in forensic expertise.

在MATLAB6.5平台上，完成了从语音样本的采样与量化、语音信号的预处理、特征参数的提取，直到语谱图的绘制，并通过对自行录制的语音样本进行分析，验证了声纹识别的理论以及本文所采用的分析方法，最后，介绍了本识别方法在实际案件中的应用。

Voice of comparison, is the main force of the voice and Mandarin dialects in the initial consonant, vowel, the same point on the tone and different points; Terminology inherited from the era, and image color, geographical aspects of the four-color set their own dialect vocabulary embodied in a character.

语音方面的比较，主要是分析了武定方言语音与普通话语音在声母、韵母、声调上的相同点及不同点；词汇方面则从继承性、时代性、形象色彩、地域色彩四点说明武定方言词汇自身所体现出的特点。

cti-based water-power information system is an interactive voice response system using the voice server model.

基于cti的水电信息系统是一个采用声音服务器模型的交互式语音应答系统；深入的分析了水电信息系统的需求，给出了系统功能和系统结构图，绘制了语音子系统的工作步骤，给出了系统数据库的总体设计，同时给出了语音子系统与数据库交互的函数接口，较好的完成了远程汇报，查询和管理水电信息的任务。

Dialogic voice Cagayan HMP, the Guide to the development of Voice articles, on the east DN series voice cards ...

dialogic语音卡加HMP的语音篇开发指南，对东进DN序列语音卡开发也又作用。

The main work of this thesis includes three aspects:(1) study the existing order adjustment methods and examine the characteristics of speeches, a new order adjustment method based on energy correlation is proposed. It is applied to JADE, KM, c-FastICA, and CMN to separate the convolutive mixtures from simulated and actual speeches, the results of analysis and comparison of separation performance verify the validity of the method.

针对卷积混合语音盲分离的频域算法，本文主要做了以下三方面工作：(1)研究了现有调序方法和语音信号的特性，提出了一种基于能量相关的调序方法，并应用JADE、KM、c-FastICA和CMN等四种复数BSS算法，对仿真和实际语音信号进行了盲分离，分析比较了分离性能，验证了本文调序方法的有效性。

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％。