搜索资源列表
G.711
- 语音信号的G.711编码。 The source code has been modified by simplifying the ADPCM algorithm to increase running speed. Here, the quantizer scale factor, y(k), is determined by \"unlocked\" factor yu(k) only. The factor yl(k) and speed-contral fact
G.723
- G.723语音编码 The source code has been modified by simplifying the ADPCM algorithm to increase running speed. Here, the quantizer scale factor, y(k), is determined by \"unlocked\" factor yu(k) only. The factor yl(k) and speed-contral factor a
HDecode-3.4-alpha.tar
- 大规模连续语音识别工具，比较实用。可以做实际建模-large-scale continuous speech recognition tools, more practical. Modeling can do practical
Great_Outdoors_by_sandals82.zi
- 一种简单有效的基于动态时变语音识别源码 对于大多数研究者来说,寻找能够匹配二重时间序列信号的最佳途径是很重要的,因为它有许多重要的应用需求.DTW是实现这项工作的显著技术,尤其在语音识别技术领域,在这里一个测试信号被按照参照模板拉伸或压缩, ,Searching for the best path that matches two time-series signals is the main task for many researchers, because of its importa
svm-scale
- 支持向量机算法，主要是针对大规模对象进行处理-Support vector machine is targeted mainly at dealing with large-scale objects
svm-train
- 支持向量机算法训练部分，主要是针对大规模对象进行训练-Support Vector Machine training component, is mainly directed against the target of large-scale training
svm-predict
- 支持向量机算法预测部分，主要是针对大规模对象进行预测-Support Vector Machine forecast, mainly for large-scale objects to predict
endPointDetect
- 简单的基于门限值的语音端点检测。先将数字化的语音信号封装成帧，利用最大的帧幅值乘以一个比例因子得到门限值-Simple threshold-based voice end point detection. First，framing the digital voice signal , then using the maximum amplitude to multiply a frame scale factor to get threshold
silence_remove
- This code is used to remove silence from a voice signal so that the recognition in voice recognition will give good result. The method is Scale on threshold applied to the envelope for detecting scilence periods. The actual threshold is computed by m
xiaobobianhuan
- 利用小波变换多尺度分析的特性,在小波域的各个尺度上选取不同的阈值对基本的频域谱减法进行改进,并且根据清浊音各自不同的特点,在去噪过程中加以分离,保留了语音中的清音成分,使语音更加饱满。-The use of multi-scale wavelet analysis of the characteristics of the individual in the wavelet domain scale threshold select a different frequency domain of
time_streching_constant_hop_size
- Speech time scale modification using constant hop size
bark2frq
- speech singnal converison from one scale to other
GAUSSIANloglikelihood
- GMM高斯混合模型大规模概率对数计算 需要一个模型地址文件和一个需要识别的声音的mfc文件可以一次执行大批量-GMM Gaussian mixture model probability on the number of large-scale computing need a model of address file, and the voice of the mfc file which need to be identified .can be an implementation
cpcl
- 采用大型通用有限元软件Patran的二次开发语音pcl编写的spar平台参数化建模程序，是学习pcl界面开发、参数化建模的好例子！-Use of large-scale finite element software Patran secondary development of voice pcl prepared spar platform parametric modeling process is to learn from pcl Interface Development, Para
Feature-Points-In-Image
- Scale Invariant Feature Transform was the final project of DIP course. The source code is provided in matlab 20-Scale Invariant Feature Transform was the final project of DIP course. The source code is provided in matlab 2011
Laplace
- 传统的短时谱估计语音增强算法通常假设语音谱分量相互独立，没有考虑语音谱分量间的相关性。针对这 一问题，该文提出一种新的基于多元Laplace分布模型的短时谱估计算法。首先，假设语音的离散余弦变换(DCT) 系数服从多元Laplace分布，以此利用谱分量间的相关性；在此基础上，利用多元随机矢量的高斯尺度混合模型表 示，推导得到语音DCT系数矢量的最小均方误差(MMSE)估计的解析表达式；并进一步推导了基于该分布模型的 语音存在概率，对最小均方误差估计子进行修正。实验结果表明，该算法
Gammashirp-filter
- In this paper, we figure out the use of appended jitter and shimmer speech features for closed set text independent speaker identification system. Jitter and shimmer features are extracted from the fundamental frequency contour and added to basel
Speech-recognition
- MFCC参数是基于人的听觉特性利用人听觉的屏蔽效应，在Mel标度频率域提取出来的倒谱特征参数。-MFCC parameters is based on human auditory characteristics using human auditory masking effect, in Mel scale frequency domain parameters of cepstrum.
JLDATA
- 摘 要：本论文主要研究了语音识别的基本原理,对语音识别系统的构成进行分析处理,其中包括预处理、特征参数提取、建立模块库、识别匹配几大部分。预处理又包括语音采样、预加重、加窗（汉明窗）、端点检测；特征提取的参数是梅尔频率倒谱系数MFCC。 该语音系统采用的是动态时间伸缩算法(DTW)，研究对象是特定人的语音识别，并在MATLAB平台上实现。为了进行后续研究，首先使用电脑中的录音系统录制了阿拉伯数字0—9的语音文件，并转化成 “.wav”格式的文件。-Abstract: This thesis
mfcc
- mfcc used in python mel-scale(mfcc used in python mel-scale)