Running absolute mean normalisation within winzorizing function

[erhmestel]

Some additions to s03compute_no_rotation.py and default.csv, editing the winsorizing function, to add the capability for running absolute mean normalisation. RAMN is selected by setting the windsorizing parameter to "-2". New parameters RAMN_EQfilt and RAMN_window also added.

As suggested in Bensen et al. 2007 section 2.1, there is both the option to weight by unfiltered data, and to weight by data which has been bandpass filtered to emphasise eathquake frequencies.

This function has been tested on data outside of msnoise (Attached an example for 1 minute of data:
ERHM_winzorizing_options.pdf)

This function has also been tested using a version of msnoise 1.6 with this version s03_compute_no_rotation.py (with some minor edits to the rest of the code to ensure compatibility with v1.6).

It seems to work, but extensive testing (including best window length) has not yet been completed.

[ThomasLecocq]

hey @erhmestel , sorry for not inputting earlier here. I'll have a look at the code and cherry-pick the changes to merge them in #227 at some point ! At that moment, it'll be needed to have some bits of documentation/graphical examples etc. But don't worry for now, the doc will undergo a strong re-write so no hurry...

[erhmestel]

Thanks @ThomasLecocq
Let me know when you have chosen what you want to do with it for v2 and I can contribute to documentation writing.

[erhmestel]

Hi Thomas,

I have updated this code to down-weight N and E components by the mean of the two to maintain consistency in processing between them, allowing rotation later. Will upload those changes soon once I have resolved the differences with the updates you've added to this function since my initial commit.

In the meantime, do you have a reccomendation on how I should cite the development version of MSNoise when explaining this updated function within my thesis?

El

[ThomasLecocq]

Hey El, I would love to see the RAMN in the 2.0 release, which I try to puuuuushhh. Would you be ready to help ? I understand it's been super long, so if not, let me know, and I'll try to cherry-pick & make the documentation too

[ThomasLecocq]

@asyates @LaureBrenot if you guys wanna cherrypick some changes made here to add to the workflow :-) welcome :-)

[erhmestel]

Hi @ThomasLecocq @asyates @LaureBrenot,
Very happy to contribute - I think I have an updated version of this code that fixed a couple of bugs - don't know if I pushed that, so will check when I am back in the office.