VQ_LBG_newdistortion

文中研究了6种常用数字调制信号识别的特征参数集，并采用决策树判别方法进行分类识别。仿真结果表明，在SNR≥5dB时，识别正确率在99 以上，且当SNR≥20dB时，识别正确率达到100 。其特点是，算法简单，识别正确率高，达到了自动分类识别的目的，并有利于实现识别的实时化。(In this paper, we study the set of characteristic parameters of the six kinds of commonly used digital modulation signal recognition, and decision tree method for classification. The simulation results show that SNR 鈮� 5dB, the correct rate more than 99 , and when SNR 鈮� 20dB, the correct rate of 100 . Which is characterized by simple algorithm to identify the correct rate, to achieve the purpose of automatic classification and recognition, and help to identify real-time.)

语音信号的语谱图生成函数，function [ ] = spetro(Winsiz, Shift, Base, Mode, Gray)，5个参数分别为帧长、帧移、取值门限（一般设为0）、彩色模式、灰度模式(This is a matlab function for creating speech spectrogram.)

语音信号的短时分析，主要包括：分帧、短时能量、短时平均幅度、短时过零率、短时自相关函数、短时幅度差、倒谱、复倒谱、lpc系数、lpc谱估计等 绝对保证质量，是保研后导师布置的一些基础程序(Short-time speech signal analysis, mainly including: sub-frame, short-time energy, short-term average, short-time zero-crossing rate, short-time auto-correlation function, short-term rate of poor cepstrum, complex cepstrum, lpc coefficients, lpc spectral estimation, such as an absolute guarantee that the quality of instructors is the security arrangement after the inquest some of the basis of procedures)

语音编码中对LSF系数进行VQ处理，LBG方法生成VQ码书，且采用了新的失真测度(Speech Coding for VQ processing LSF coefficients, LBG method of VQ codebook generation, and adopted a new measure of distortion)

能自动循迹 图像识别 语音识别 避障 红外光感应 过ETC等功能(Can automatic speech recognition obstacle avoidance tracking infrared induction ETC function)

VAD去噪算法，可以有效的过滤静音和无效的语音，同时消除了一些背景噪声。(VAD denoising algorithm, can effectively filter invalid mute and voice, while eliminating some background noise.)

Speaker Identification using MFCC feature and VQ training, very clear MATLAB code

语音合成程序源码！psalo频域基音同步叠加方法。它首先对原始语音信号进行短时频域变换，的到短时谱与短时谱包络。短时谱除以短时谱包络的到声源短时谱，对声源短时谱的实部与与虚部分别进行线性插值，就能达到改变变语音信号基频的目的，然后再进行频域反变换，可的到变换后的短时语音信号。短时谱包络部分也能独立改变，以达到改变音色的目的。 (Voice synthesis program source! psalo frequency domain pitch synchronous superposition method. It was first carried out on the original speech signal a short-time frequency-domain transform, to the short-time spectrum and short-time spectrum envelope. The short-time spectrum divided by the short-time spectrum envelope of the short time spectrum of the sound source, the short time spectrum of the real part of the sound source, and the imaginary parts of the linear interpolation, can achieve the purpose of changing the fundamental frequency of the alternating speech signal, and then then the inverse transform of the frequency domain, can be to the short-time speech signal after conversion. The short-time spectral envelope section can be varied independently, in order to achieve the purpose of changing the tone.)

LD3320 是一颗基于非特定人语音识别（SI-ASR：Speaker-Independent Automatic Speech Recognition）技术的语音识别/声控芯片。提供了真正的单 芯片语音识别解决方案。 LD3320 芯片上集成了高精度的A/D 和D/A 接口，不再需要外接辅助的 Flash 和RAM，即可以实现语音识别/声控/人机对话功能。并且，识别的关键词 语列表是可以动态编辑的。(Voice Recognition ；Speech Recognition ；Speech recognition technology )

谱减去，最常用的程序，实习除噪，语音增强(To achieve denoising)