Time Dependant Smoothing or Filtering

[ckeiderling]

I have a waveform which I am plotting on a log(x) scale and would like to accordingly filter this data such that it smooths the larger x values much more than the smaller ones.

I am strugling to find anything that will do this, are there any VI's that would do this for me?

Thanks 

Chaz

[Marshall_B]

Hi Chaz,

Are you able to send your code over so that we can take a look at what you have implemented so far?

It may be best to separate your waveform and perform filtering independently on each section.

[Bob_Schor]

I think you may have to "roll your own".  When you do a log(x) transform (for example, a Bode plot with log(frequency) on the X axis), the number of points-per-unit-X-axis increases as you go to the right.

Think about smoothing an "ordinary" (not-log) plot.  One way to do this is to divide the X axis into, say, 100 equal intervals (I'm assuming you have 1000 points) and replace all of the points with the average within each interval (like a histogram).  You can also take this 1% window and slide it along the X axis, averaging as you go, the so-called "moving boxcar average".  Note that if the X axis has width W, each bin would have width W/100.

You can do the same thing with your log plot.  The difference is that instead of dividing the X axis into 100 intervals, you divide the log-X axis into 100 intervals.  Now the width of the bins is not W/100, but the 100th root of W.  

This should get you started, I hope.

BS