Re: [MSNoise] QUESTIONS ABOUT ALL PROCESS WITH MSNOISE

Fernando, the parameters look mostly OK, except: if cc_sampling_rate = 20 and your data is 125 Hz, you can't use Decimation by factor=125 ! You have to either cc_sampling_rate = 25 decimation factor = 5 or cc_sampling_rate = 20 method=Resample In a future version, we will use the Trace.decimate() function from obspy which automatically lowpass the signal, but that would not solve your problem there ! Anyway, that should work better already (I'm surprised you even have results ... it should crash at step 3)... Are the CCF looking OK ? when you plot them manually ? THEN: for the MWCS, you need to remember that, for each time lag, the dt is calculated by a weighted linear regression, which needs at least several (>10?) points to work. So: if you have 0.1 to 0.2 Hz frequency band, you need 10 points there, which is equivalent to 100 points for 1 Hz --> you need to have a mwcs_wlen of ~100s !! (except if you hack a little the FFT and add zeros to the time series, but I don't recommend it without being really willing to dive deep in the maths/science of fft). Cheers, Thomas Le 11/02/2015 18:17, Fernando Salazar Monroy a écrit :