Octave analysis (Algorithm)

[virtman]

Hi Andy,

I agree with you with the points that you talked about,  but if i use the power spectrum  i am calculating the g RMS  values for my power spectrum vector. but for me i am looking only for the peak g level.

Secondly, i am in that example analysing the signal in only one impulse and i think the values are quite near to be real !

on the other hand, that   the FFT generally it split our time signal into a sum of a sinewave so i think that i need to make the sum of my spike's magnitude only of each band and not the sum of all the vector's band ?! it seems to me like i am calculating the energy if i do only a sum square but not the g level!

Thank you again for helping me

Best Regards

[PsyenceFact]

I'm afraid I'm not quite sure what you're asking here.  The peak level in any band is sqrt(2) times the RMS level.

Attached is a modified version of my earlier example showing the calculation of 1/3-octave band levels using filters and using the FFT autopower.  The signal contains two sine waves with amplitiudes 1 and 2.  Both methods give RMS band levels of 0.707 and 1.414 as expected (multiply by sqrt(2) to get original amplitudes).  I hope this makes things more clear.

Andy

[virtman]

Hi Andy,

Could you please make it on an old version because i am using a 2012 LabView version

Thank you

[PsyenceFact · Accepted solution]

I saved this for version 12.0.  See if it works for you.

Andy

[Murph1a37]

Great Job of explaining a complex subject in easy to understand description.