Automatically determine proper pre_filt setting for response removal
Add a tool that uses the noise window prior to the signal to determine the longest period at which a given channel has signal above the noise level, and use that analysis to define prefiltering limits in the station correction step (currently no prefiltering is used by default and that may be degrading the inversions). This could also be used to determine the best period limits for the inversion, by looking at all the stations and optimizing the range so its as long period as possible without starting to include amplified station noise. Consider using wavelet transform methods rather than Fourier spectra for working at these longest periods.
See also #27
Edited by Liam Toney