语音

As an example, using the standard AMR-NB (Adaptive Multi Rate compression, narrow band) vocoder C reference, the CEVA-X1622 DSP core required just 19 MHz when compiled out-of-the-box (for worst-case frames and streams). In comparison, other licensable solutions require more than 45% higher speed for the same out-of-the-box compiled code.

以使用标准窄带自我调整多速率压缩的C语言语音编码器为例，CEVA-X1622 DSP内核在编译现成可用的代码时，仅需19 MHz速率；而其它可授权解决方案却需要高百分之四十五以上的速度，才能编译同样的现成代码。

VeriSilicon specialises in providing expert design services, market leading licensable cores and platforms, industry standard semiconductor IP and scalable ASIC turnkey services across a broad range of application markets, including multimedia, voice and wireless communications.

芯原股份有限公司成立于2001年，是一家发展迅速的芯片产品解决方案公司，公司提供的产品和服务使客户能够达到他们芯片设计的目标，加速开发项目并以较低的成本及时提供市场公认的芯片产品。芯原专注于在多媒体、语音和无线通信等广大的应用市场提供专家设计服务、市场领先的授权内核和平台、业内标准的半导体 IP 以及可升级的 ASIC 全包服务。

It is a gradient-based method, which determines that the final solution is just a local optimum for the objective function, the likelihood that the models produce the training data.

本论文以进化计算为核心优化算法，主要研究了它在语音识别中的声学模型训练以及说话人辨认中的模型训练中的应用。

The highly integrated CX2070X SoCs include an integrated audio/voice digital signal processor, multi-bit codec, digital and analog input/out interfaces, and a power efficient 2.5-watt per channel Class-D amplifier to drive stereo speakers or optional post-width modulation and differential line-out for external speaker systems amplifiers.

高度集成的 CX2070X SoC 内置了集成的音频/语音数字信号处理器、多位编解码器、数字和模拟输入/输出接口，以及每通道功效为 2.5 W 的、驱动立体扬声器的或可选的后宽度调制的 D 类放大器，还有用于外部扬声器系统放大器的差分线路输出。

A neural network architecture, K-subspaces and time-delay auto-associators, is proposed for phoneme recognition. It extends the phoneme filter neural networks approach by adding linear auto-associators to create p-dimension subspace, and an iteration is employed to improve the decision.

提出了用于音素识别的K子空间和时延自相关器神经网络结构，用将时延设计加入线性自相关器，以扩展音素滤波神经网络的方法，产生p维子空间，并采用迭代过程修改划分，以便捕获语音信号中的时间序列信息。

The effect of channel made character parameters offset linearly, so the performance of speech recognition system was influenced.

信道的影响使特征参数产生线性偏移，从而影响了语音识别系统的性能。

And, in the same time, Chinese fonts are the opposite to lingual system.

同时， 中国方块字与语音文字迥然不同，两者的功能与宗旨刚刚相反。

Syntactic function: it shows the relationship between a linguistic form and other parts of the linguistic pattern in which it was used.

句法功能，体现了语言形式和语言模式其他部分之间的关系，语音形式是在语言模式中运用的

The language principles of Chinese names of the most valuable brands in the world were analyzed and summarized in the three aspects as grammar, phonetics and semantics using linguistical and statistical methods.

并用语言学和统计学的方法从语法、语音、语义三个层面分析并总结了世界最有价值品牌中文名称的语言规律。

This article is trying to discuss form a cognitive linguistical angle about the cognitive mechanism of the metaphorical techniques and its source domain, target domain, ground, situation refered.

从认知语言学的角度来看，汉语灯谜文本中蕴涵着特有的隐喻手法，包括语音隐喻、拟形隐喻、语义隐喻和语法隐喻。

However, as the name（read－only memory）implies, CD disks cannot be written onorchanged in any way.

然而，正如其名字所指出的那样，CD盘不能写，也不能用任何方式改变其内容。

Galvanizes steel pallet is mainly export which suits standard packing of European Union, the North America. galvanizes steel pallet is suitable to heavy rack. Pallet surface can design plate type, corrugated and the gap form, satisfies the different requirements.

镀锌钢托盘多用于出口，替代木托盘，免薰蒸，符合欧盟、北美各国对出口货物包装材料的法令要求；喷涂钢托盘适用于重载上货架之用，托盘表面根据需要制作成平板状、波纹状及间隔形式，满足不同的使用要求。