# Filtering pulses by width

Assume we are receiving gaussian pulses and I'd like to do some sort of band-pass filter on the standard deviation $\sigma$ - as best as can be done. Is there a way to do this?

I have looked on google, and there are several mentions of gaussian wavelet filters, but I don't think these are what I'm looking for.

• what $\sigma$ ranges are you looking for, i.e. what's the ratio between smallest and largest std.-dev. that you care about (a single impulse can be interpreted as Gaussian with zero deviation, and a constant as one with infinite deviation, and you definitely only care about things in between). – Marcus Müller Jul 20 '18 at 8:35
• I could start with 3 ranges of std deviation based on some baseline $\sigma$: $1\sigma$, $2\sigma$, $4 \sigma$ – Mike Flynn Jul 20 '18 at 18:50