smooth curve smooth curve

Hello, everyone.

How can I best smooth the curve (see attachment) (fitting)? Do you have an idea?

I thank you

[Jan]

How do you define "good"? This must be specified before a "best" can be found. Do you want a low or high pass filtering? Or in other words: Are the low or the high frequencies the noise? "Fitting" is meaningful, if you know the function to fit from a physical model.

[Latifa Bouguessaa]

I just need a function that smoothes the curve

Answer 1

[Star Strider]

That appears to me to be a fft result after using the fftshift function on it. Each of the ‘peaks’ is a different frequency component.

Experiment with zero-padding the fft:

L = numel(s);

NFFT = 2^nextpow2(L);

FTs = fft(s, NFFT)/L;

Here, ‘s’ is the original time-domain signal.

The frequency vector should go from -Fn to + Fn, where ‘Fn’ is the Nyquist frequency (one-half of the sampling frequency).

.