Matlab在DFT频谱误差分析中的应用

Application of Matlab in Error Analysis of DFT Frequency Spectrum

研究数字信号最有效的方法通常是离散傅里叶变换（DFT）,时信号进行DFT时会存在各种误差.本文针对DFT 分析的主要误差来源--栅栏效应和截断效应,在理论分析的基础上,采用MATLAB真,更加直观的说明了补零只能减轻栅栏效应,使频谱更加细致,只有增加有效信号的截取长度,才能提高频率分辨率,减轻截断效应.最后给出如何选择适合截断长度的结论.

DFT is always the most efficient method of digital signal processing, but it brings kinds of errors.According to the main source of error fence effect and truncated effect,this article based on Matlab simulation, more intuitive illustrates that filling zero can only ease fence effect, make the spectrum more meticulously, only increaseing effective signal intercept length can improve the frequency resolution, reduce truncation effect. At last,Present the conclusion of how to select suitable truncation length.