Resolution Formula


I am confused with the formula provided for Resolution/Bandwidth.

As per ISO 13373-2 and many other books, the formula for Bandwidth is = Fmax/No of Lines.

My Category III Vibration Institute material claims it as =2*WF*Fmax/No of Lines. The second one was even discussed to be correct in our group.

But practically, I am getting a resolution which is consistent with " Fmax/No of Lines" (Refer attached).

If 2*WF*Fmax/No of Lines is the formula for Resolution, why my actual resolution is better than the value calculated by this formula? Then, what exactly is the Bandwidth in the attached spectrum?  Are Bandwidth and Resolution two different terms? 





Files (1)
Formula for Resolution/Bandwidth
Original Post

Additional to the article advised above here is my comment.

Lets assume you have two signals.

Signal1: amplitude 8 and frequency 1 Hz

Signal2: amplitude 3 and frequency 2 Hz

I created a spectrum which has a resolution of 1 Hz, lets say 12 Hz and 12 line for simplicity. As you see from the 1st picture you can resolve these 2 frequencies without windowing, red line in the drawing. But if you use hanning window (blue line) you can not resolve these two frequencies if they have a difference of 1 Hz. Your marker still moves 1 Hz between the points in the spectrum but the result will be unsatisfactory.

So the short answer is yes, Bandwidth and Resolution are two different terms and bandwidth is 2*WF*Fmax/No of Lines

Ideally your spectrum should look like picture 2. I mean there should be another point between those 2 frequencies.  So your maximum resolvable frequency (with a good accuracy) is even less and it is your bandwidth/2

I think i made everything more confusing...

2-Hanning vs Normal 1 Hz ResolutionHanning vs Normal 1 Hz Resolution


Photos (2)

Add Reply