How program for-loop to calculate RMS with window length of 100 samples and window step of 1 sample

[jcannon]

Hi everyone,

I am trying to make a switch from Matlab to Labview from my signal processing needs....Specifically, I need to smooth and quantify a randon, biphasic signal of ~128000 samples in length... In matlab I would calculate the RMS with a window length of 100 samples and window step of 1 sample to smooth.linear evelope the signal; then do my amplitude calculations...

I am having trouble doing this in Labview.. The matlab code I would normally use is:

___________________________________________________________ 

% Assign RMS variables

x= signal of interest;

window_length= 100;

window_step= 1;

% RMS calculation

EMG= x.^2;

index= 0;

nwindows= floor(nrows - (window_length - window_step)) / window_step;

EMG_rms = zeros(nwindows,ncols); % Preallocation for memory

for j = 1:window_step:length(EMG)-window_length+1;

    index = index+1;

    EMG_rms(index,:) = sqrt(mean(EMG(j:j+window_length-1,:)));

end

________________________________________________________

I have tried  numerous times using the for-loop, but no luck... Ingore the smilies... They should be a colon (:, to indicate all rows) followed by a )....

Any suggestions??

Thanks,

[crossrulz]

Can you show us what you have tried so far?

It seems to me that you need a for loop with an Array Subset inside of it.

[DylanC]

Hi jcannon,

Here is a link with an example on for loops which should get you familiar enough with them to be able to achieve what you need.

https://decibel.ni.com/content/docs/DOC-20828

DylanC

[GerdW]

Hi jcannon,

there's a ready-to-use function that should do what you need...

[crossrulz]

---

@GerdW wrote:

Hi jcannon,

there's a ready-to-use function that should do what you need...

---

To be more exact, you need the RMS PtByPt VI.

[jcannon]

Hi everyone,

Thanks for your feedback with this and the link the the functions...I am ok wth calculating the RMS, my problem is wiring the interation and/or "N" count of the for-loop the get the appropriate sub-array for each signal segment be analysed.

For example, let's say I have a signal that is  5000 samples in length and I want to calculate the RMS for consective windows that are 100 samples long and have the window step of 1 sample each time fr thr total ength of the signal (i.e. smooth it). In this situation the total number of interations for the for-loop runs would be calculated as:

runs= floor(signal_length - (window_length- window_step)) / window_step;

runs= floor(5000-(100-1))/1

runs= 4901.

So, I would the for-loop to run 4901 times (and ingnore the remaining 99 samples at the end of the signal)... In each loop I would calcuate the RMS of the appropriate 100 sample subarray... But how do I wire the interation count and "N" counter to the subarray so that each interations gets the right subarray for each specific interation?...

Thanks,

Jack

[GerdW]

Hi Jack,

use the Pt-by-Pt function and you will not need any subarrays...

[jcannon]

Of course!

Had a late night...

Any ideas about the wiring?

Thanks,
Jack

[GerdW]

Hi Jack,

as PtByPt-RMS only has 2 (useful) inputs for your duty wiring is as easy as this:

But well, one could read the context help to know all this...

[jcannon]

Hi GerdW,

I think you might not be correctly understanding my problem regarding the wiring... I am refering to the wiring of the "N" and/or "I" terminal of the for-loop into the coding inside the loop(?) so that the RMS is calculated for the appropriate segment of the signal in each of the 4901 interations... That is, samples 1-100 for the 1st interation, samples 2-101 for the 2nd interaction, etc, etc, through to samples 4802-4901.

I think with the code you have provided, the RMS will only be calculated on the 1st 100 samples of the 5000 sample array...? How to do get the remaining data? I think I will have to use the subarray function with the "N" and/or "I" terminal of the loop wired so that the subarray specs update on each interation to get the new segment of data needed for processing...?

Regards,

Jack