资源列表
zero-crossing
- finding energy of speech signal to distinguish between voiced and unvoiced signal
yuyinshibie
- 基于特定人情感的语音识别，语音情感特征提取；语音情感分类器的设计，完成了一个特定人语音情感识别的初步系统。内容很丰富-In this article, we have accomplished a system for emotion recognition of speech by establishing a database of speech with certain emotion by certain people analyzing emotion features and de
attentionModel
- attention model for saliency detection.
voicebox
- matlab voice 工具的子程序软件 主要是进行语音识别-Matlab subroutine of the voice tools software is speech recognition
time_extend
- 语音延长、变速处理程序，实现语音信号时域延长； 主要步骤是先对原始语音信号进行加窗，再对加窗后的语音信号进行时域扩展； 以上处理之后语音信号可能会带进新的噪声，所以再对处理后的语音信号，使用FIR滤波器进行去噪处理。-Voice extending process, extending the time-domain voice signal The first major step is added to the original speech signal window, a
cb_Spectrum
- 用vc编的用于语音的频谱的检测，可以显示语音的频谱图。-It is used to detect spectral of speech and display the spectralgram.
HDecode
- 大词汇量语音识别解码工具，可以在VC2008环境下编译-Large vocabulary speech recognition decoding tool, you can compile under VC2008
Speech
- 基于微软sdk利用Vc++实现语音合成的程序代码-Microsoft SDK-based speech synthesis program by using Vc++ code
1
- QQ机器人源码 QQ机器人源码 -QQ robot source
Smallvocabularyspeechrecognitionofnon-specificpers
- 小词汇量非特定人语音识别在嵌入式系统中的应用。-Small vocabulary speech recognition of non-specific person at embedded systems applications.
11
- 为提高语音端点检测系统在低信噪(0 dB 以下) 下 检测的准确率, 提出了一种基于谱熵的端点检测算法。将每 帧信号分为16 个子带, 选取频谱分布在250～ 3. 5 kHz 并且 能量不超过该帧总能量90 的子带, 计算经过语音增强后的 子带能量以及各子带信噪比, 根据各子带信噪比的不同调整 其在整个谱熵计算过程中的权重, 然后平滑谱熵, 以最终的 谱熵作为端点检测的依据-To improve endpoint detection system in the low
LPc
- 线性预测编码器，基于线性预测模型的语音编码器。-Linear predictive coding, and models based on linear prediction voice coder.