Consistency between h/v from geopsy2 and geopsy3

[gaetano.riccio]

Hi,
I am carrying out some h/v on earthquakes with geopsy-hv in geopsy 3.4.1, and comparing the results with those obtained with geopsy 2.10.1. I made several tests, both without smoothing and using the Konno-Ohmachi smoothing (with SMOOTHING CONSTANT = 40 in geopsy 2, SMOOTHING_WIDTH = 0.2 in geopsy 3).
In any case, the h/v that result from the two versions are not consistent with each other; since the results of geopsy 2 have always been reliable, i wonder if I am making something wrong with geopsy 3 parameters.
I attach the image of the comparison, the data files (sac) used, the .hv files with the corresponding parameter files; in the comparison there is also the h/v made with sac, which is in agreement with that of geopsy 2.

PS since the analysis is on earthquakes, the signal window is unique, so in the parameter files for geopsy 3 I set WINDOW_MIN_LENGTH and WINDOW_MAX_LENGTH equal to the duration of the signal, instead for geopsy 2 I indicated the window file from the command line using the -load-win option.

Thank you in advance, best regards.

[admin]

Hi Gaetano,

I ran the HV tool on your SAC signals with geopsypack 2.5.0 and 3.4.2-preview (which is unchanged compared to 3.4.1). When loading your parameters 'param-gp3.smooth', I noticed that the Tukey taper width was 50% while in 'param-gp2.smooth' there was no taper.

Using the same taper width with both releases I could obtain exactly (visual comparison only) the same curve as you can see on the attached plots. Note that geopsypack 2.5.0 does not have the smoothing option 'Y scale type': it is using a linear scale for smoothing in all cases. By default geopsypack 3 is using log scale which introduces a little difference (see plots) but reverting to linear a perfect match is observed.

Best regards,

Marc

[gaetano.riccio]

Ok, I hadn't considered the taper width.

Thank you very much, best regards.