integral with irregular distances of values

[Schaedel]

I´m trying to integrate some measured data with the standard VI´s. It's a two dimensional array (x,y) and I want to integrate y over x. My problem is, that the values of the x-axis are not equidistant, thus I can not define a fixed dx. So I tryed to interpolate the array. It worked for linear interpolation, but that creates an appreciable error (compared to an Origin calculation). Other interpoaltions didn't work (because of the irregular x-axis?). Does anybody knows a proceeding that would solve this problem?!

Thanks

[becktho]

Hi

At least there is the way of calculating it on your own.

[unclebump]

Search for examples to resample unevenly spaced data.

[jasonhill]

All the examples I found required the Signal Processing Toolkit, which I do not have.  I do have Spline Interpolation 1D and it looks like it should do what you want pretty easily, but that is not in the base package either.  As much as becktho's suggestion hurts, it may be your best option.

[becktho]

Many thanks jasonhill

This motivates me again as I was annoyed after receiving just one star.

Maybe my first post was quiet to short, but I think it's better to calculate the integral on your own, than make lots of workarounds just to get a regular dx. It is not just the x-values that then have to be modified but also the y-values, which is not the best way I'd say.

[unclebump]

Try this link for a 7.1 format base package sample.

http://sine.ni.com/apps/we/niepd_web_display.display_epd4?p_guid=F4CFCADFD0DA2B3AE0340003BA7CCD71

[Schaedel]

I'll try to imply an external matlab procedure. Seems to be the "simplest" and most exactly way...  😉

Thanks all for your help.

[jasonhill]

unclebump,

The example you link to uses a waveform.  Waveforms can only have equadistant points.  As I understand it, Schaedel's data is unequally spaced.

[unclebump]

OOPS!! Try this one.

http://sine.ni.com/apps/we/niepd_web_display.display_epd4?p_guid=B45EACE3D93356A4E034080020E74861