Hi,everyone.In my previous questions Dr.Joel explained a lot but these two questions still confuse me.First,how to decide number of repeated FFTs according to RBW/VBW ratio.Dr.Joel mentioned that for a RBW/VBW ratio of 10,there are 5 FFTs averaged before the result is processed.If I change the ratio,how dose the number of FFT average change?A function to calculate that is better.And the second is still about window function.If I use a window before FFT,to maintain the RBW,increasing record size is necessary.Dr.Joel mentioned that the gaussian filter,with 90dB sidelobes is about 1.87 longer than the no window case.I use the SA option in PNA-X and I found things are different.With gaussian filter,record size equals the record size under no window case multiply by about 1.986.And with flattop,kaiser and blckman filter,the number is 3.764,1.964,1.623.I don't know why these numbers,how did they been calculated.Is there some functions can solve this?Another small question:the kaiser window is a adjustable window,the parameter β is the key.I want to know what dose β equal in SA option of PNA-X.Thanks a lot.

Just curious why such detailed questions. What is your application? If you describe it we can tailor a better answer to your questions.

I just did a quick test and of course you are right, 1.988 factor, about. The sidelobes are more like -105 or more so I'm thinking somewhere in the process we tweaked the alpha factor to get better sidelobes (makes sense as we really don't like to see them in the data; you have to struggle now to find a region where you can even see the sidelobes.

I don't know be Kaiser B factor but I would guess 6. That's default for time domain. JP will have to look in the code to find the real value. The extra windows are just there "for fun". for the most part the three windows: none; Gaussian and Flattop will be optimum for the different signaling cases. E.g.: Flattop is optimum for finding close in signals, none is optimum for some classes of coherence measurements, and Gaussian is optimum for noise measurements.