Is your feature request related to a problem? Please describe.
The current gmprocess is using ObsPy filter, which is applied in the time domain and yields significantly high frequency content at low frequencies. Hence, ObsPy filter is not preferred for the processing of short time window traces.
Describe the solution you'd like
Could gmprocess be updated so the user can opt for the filter of their preference? Either ObsPy filter or the Butterworth filter?
Describe alternatives you've considered
When we process the records using the same procedure as gmprocess, but applying a Butterworth filter instead, we obtain more accurate frequency content at low frequencies.
Additional context
The figure below shows a trace that has been processed using the Butterworth filter (dark blue) and ObsPy filter (light blue). Notice the amplitudes at low frequencies, it should decrease gradually (like when the Butterworth filter is applied), instead it is constant when ObsPy filters are implemented. We would like to avoid this behavior for the processed records.
The unprocessed and noise traces are depicted in orange and green respectively.
Is your feature request related to a problem? Please describe.
The current gmprocess is using ObsPy filter, which is applied in the time domain and yields significantly high frequency content at low frequencies. Hence, ObsPy filter is not preferred for the processing of short time window traces.
Describe the solution you'd like
Could gmprocess be updated so the user can opt for the filter of their preference? Either ObsPy filter or the Butterworth filter?
Describe alternatives you've considered
When we process the records using the same procedure as gmprocess, but applying a Butterworth filter instead, we obtain more accurate frequency content at low frequencies.
Additional context
The figure below shows a trace that has been processed using the Butterworth filter (dark blue) and ObsPy filter (light blue). Notice the amplitudes at low frequencies, it should decrease gradually (like when the Butterworth filter is applied), instead it is constant when ObsPy filters are implemented. We would like to avoid this behavior for the processed records. The unprocessed and noise traces are depicted in orange and green respectively.