Gaussian Filter
29 views (last 30 days)
Show older comments
Hi,
i'm searching for a Gaussian Filter to filter an 1d trace (125 Hz, x-axis:time, y-axis:signal) with a cutoff=4 Hz. Is there a function in matlab?
Thanks for your efforts!
0 Comments
Answers (7)
Rick Rosson
on 29 Jun 2011
In the Signal Processing Toolbox, please try either filterbuilder or guassfir.
HTH.
0 Comments
David Young
on 30 Jun 2011
A Gaussian filter does not have a sharp frequency cutoff - the attenuation changes gradually over the whole range of frequencies - so you can't specify one. This follows from the fact that the Fourier transform of a Gaussian is itself a Gaussian.
What you usually specify is the frequency at which you require a certain attenuation. This is inversely proportional to the "width" of the Gaussian in the temporal domain. (It's not really a width - again it's the time at which the curve drops to a certain fraction of its maximum.)
You can find the detailed formulae and a proper description of all this (better than will fit into a MATLAB Answers answer) in signal processing textbooks. There's a starting point on the web at the Wikipedia article.
0 Comments
Rick Rosson
on 30 Jun 2011
Please review the function reference page in the MATLAB documentation:
doc gaussfir
This page provides information on how to specify the 3 dB points.
HTH.
0 Comments
Michele Bertoni
on 27 Jun 2018
Hi to all, I'm a bit confused.... First of all, I can't understand why should I apply low-pass gaussian filter to digital (1-D such audio or else) signal.... I've always seen IIR Butterworth or similar, maybe is it a zero-phase?
Taking a look to Wikipedia gaussian filter page I can understand gaussian on both time and frequency domain have same shape, ok. St=sigma (in time domain) Sf=sigma (in frequency domain)
St*Sf=1/(2*pi) ok, clear.
It's also said Sf is equivalent to cut-off frequency, in this case: Fs = sampling frequency fc = cut-off frequency St = sigma (in time domain) in number of samples fc=Fs/(2*pi*St)
But in this case fc equals to -6dB, while it's used correction factor c for different cut-off value.
Ok, let's take fcarl example: Fs = 125Hz fc = 4Hz I get St = 5, so I need to take gaussian kernel from -5 to 5 (11 values), is this right? But about coefficient values? Thanks,
0 Comments
See Also
Categories
Find more on Digital Filter Analysis in Help Center and File Exchange
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!