mscohere: reduce high frequencies noise variance.

2 次查看(过去 30 天)
Hi everyone, I am trying to reduce the noise variance at high frequencies obtained as an output of computing the coherence between two tme-series. I played around with the overlap function, chosen different type of window (Hanning and kaiser) and increasing the window length up to 5 times approximately larger than the window length requires to resolve the frequency I am interested to. Below are the parameter:
Fs = 50 Hz; % Sampling frequency N = 3500; % number of sample NFFT = 172.4*5; Noverlap = round(98*NFFT./100); Window = kaiser(NFFT, 28);
[Cxy_B38,fxy_B38]=mscohere(B38_filter(:,5),B38_filter(:,6),wind,noverlap,NFFT,Fs);
Can someone give me a tip on how to further reduce the noise at frequencies above 1.5 Hz?
I am looking forward to hearing from you.
Axel
  1 个评论
Sainath Varikuti
Sainath Varikuti 2015-8-21
Hello Edinburgh,
You may pass the output of 'mscohere' function through a low-pass filter. The documentation page here provides examples of designing low pass filters.
Regards,
Sainath

请先登录,再进行评论。

回答(0 个)

类别

Help CenterFile Exchange 中查找有关 Multirate Signal Processing 的更多信息

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by