Search - LPC CAN

Search - LPC CAN - List

[Audio program] LPC DL : 0: 线性预测LPC的倒谱计算及语音识别的相关程序，可以在VC2005的环境下运行-LPC Linear Prediction Cepstral computing and voice recognition procedures can be run under the environment VC2005
Date : 2026-01-10 Size : 1kb User : wangdi
[Audio program] risidual DL : 0: 预测语音信号的LPC系数，求出残差信号，再由残差信号恢复原始语音信号，由此可以看出残差信号在信号处理中的重要性。-Prediction of LPC coefficients about the speech signal, and a residual signal is obtained, and then restore the original speech signal from the residual signal, the importance of which can be seen in the residual signal in the signal processing.
Date : 2026-01-10 Size : 102kb User : 田静
[Audio program] Linear-Predictive-Coder-master DL : 0: Linear-Predictive-Coder MATLAB Implementation of LPC algorithm for speech signal # Why LPC? In communication systems it is often necessary to transmit audio(speech) signal in compressed or encoded form because of bandwidth limitation of the channel. In this regard, ‘Linear predictive coding(LPC)’ is an effctive method of speech coding at a low bit-rate. # Features ** Analysis/Encoding phase,Synthesis/Decoding phase. **Human voice modelled with all-pole filter. ** LPC parameters(filter coefficients, pitch, gain etc) extraction at the decoding phase. ** Non-overlapping frames of 30 milliseconds in duration # How To Run ** Make sure MATLAB(latest version) is installed ** Put both files(LPC.m with .mp3 file) in the same folder ** Open LPC.m file and run it. ## Comments Different audio (.mp3) files can be coded/decoded by changing the input file name in the code.
Date : 2026-01-10 Size : 2kb User : japaoli
[Audio program] libogg DL : 0: 首先对输入音频PCM信号进行时频分析，决定MDCT的长度，即加窗，然后进行MDCT变换；同时对原始音频信号要进行FFT分析。两种变换的频谱系输入给心理声学模型单元，MDCT系数用于噪声掩蔽计算，H可结果用于音调掩蔽特性计算，共同构造总的掩蔽曲线。然后根据MDCT系数及掩蔽曲线，对频谱系数进行线性预测分析用LPC(Linear Prediction Coefficience，线性预测系数)表示频谱包络，即基底曲线(Floor Curve)；或通过线性分段逼近方式获得基底曲线。从MDCT系数中去掉频谱包络则得到白化的残差频谱(Residue)，由于残差频谱波动范围明显变小，从而降低量化误差。之后可以选择是否采用声道耦合(Channel Coupling)技术进一步降低冗余度，耦合主要是将左右声道数据从直角坐标映射到平方极坐标；最后对白化的残差信号有效地以矢量量化表示。最后将要传输的各种信息数据按Vorbis定义的包格式组装，形成Vorbis压缩码流。-First, the input audio PCM signal time-frequency analysis to determine the length of the MDCT, that windowing, then MDCT transform while the original audio signal to FFT analysis. Two Transformations spectral line input to the psychoacoustic model unit, MDCT coefficients used to calculate the noise masking, H can be used to tone masking characteristic calculation result, the overall structure of the common masking curve. Then according to the MDCT coefficient and the masking curve, spectrum analysis using linear prediction coefficients LPC (Linear Prediction Coefficience, linear prediction coefficients) represent the spectral envelope, ie the base curve (Floor Curve) or approaching the base curve obtained by linear segments. Removed the MDCT coefficients in the spectral envelope of the residual spectrum obtained albino (Residue), since the fluctuation range of the residual spectrum significantly smaller, thereby reducing the quantization error. Then you can choose whether to use cha
Date : 2026-01-10 Size : 494kb User : 张小