filtering and smoothening of acquired data due to noise

I have position data sets acquired from noisy sources and I wanted to filter/smoothen them to rid the effects of noise. The aim is to be able to get their velocities and perform other computations. I would like to ask what a nice filter/smoothening function would be appropriate.

I previously used Butterworth filter, but perhaps that is too 'strong'? Is there any thing more appropriate?

Answer 1

[Star Strider]

It depends on the frequency characteristics of the noise. If your signal is limited to a specific part of the frequency spectrum (usually low frequencies), you can eliminate higher frequency noise with a properly designed lowpass or bandpass filter. Broadband noise can never be completely eliminated by a frequency-selective filter. For that, you must use wavelets.

[Anupam Sharma]

Hello Star, how to estimate Broadband noise using MATLAB function? which fucntion to be used?

[Anupam Sharma]

I have a hi resolution data from an accelerometer as a spectrum and I want to estimate its noise everytime while I am analyzing the spectrum.

I am unable to quanitfy the noise in FFT. Please suggest.