- From: Daniel Nilsson <daniel_at_oden.homeip.net>
- Date: Thu, 3 Mar 2005 20:38:01 -0500
On Thu, Mar 03, 2005 at 12:04:18AM -0500, Sandy Pond wrote: > On Wed, 2005-03-02 at 22:25 -0500, Daniel Nilsson wrote: > > If not, does anyone have any filter code that you can share the is > > designed to filter out any noise that might be present on the inputs ? > > The issues that I'm having right now is that the inputs don't vary > > much but I'm getting a measurement values that move back and forth > > around the actual value. I assume some of this is digital noise and > > some is analog noise (60Hz noise from the mains for example). I would > > think someone had already written some code to filter out such noise ? > > > First a little DSP info: If there's noise on your input you'll need to > sample at twice the maximum frequency of the noise in order to digitally > filter the noise. Otherwise you'll end up with alias errors in your > recorded input. If this isn't possible then you need to alter your > setup to reduce stray noise or install a hardware filter before the > input to eliminate the higher frequency noise. Agreed, I think in my case it should be fine though. I can't see any high frequency noise in the signal. > Once you collect the samples it is easy to digitally filter the signal > by either using a commercial program like matlab, open source tools like > python, of writing a small C program. If you google around on "digital > filtering" you will find programs. For instance: > > http://ccrma.stanford.edu/software/stk/ Thanks for the pointer. Yes I assume there is a ton of information written on this topic. I even remember some of it from the DSP courses at school ! For my application I assume a low-pass filter should work just fine. For example averaging the last 100 samples thus creating a windowing average function might be enough ? Or is there more to it then that ? Thanks -- Daniel Nilsson
Received on 2005-03-04Z01:38:01