Page 1 of 1

H/V Curve Inversion

Posted: Thu Oct 03, 2019 8:18 pm
by EduardoPitta
Hello everyone,

I'm having some trouble with this new version of Dinver.
Dinver read a .hv file as frequency x ellipticity (Degrees) instead of frequency x H/V (as the old version).
Can someone tell me how to invert a H/V Curve in this new version of dinver?

Re: H/V Curve Inversion

Posted: Tue Oct 08, 2019 2:05 am
by Rexha.vry

I am dealing with the same issue. In the older version, I used to input the .hv file resulted from Geopsy which contains the spectrum of H/V (frequency vs H/V) to dinver. Then, dinver read it as ellipticity (H/V) and I would do the inversion using that. However, in this new version, the .hv file resulted from Geopsy is read as ellipticity (in degree) in dinver. I found some way to consider it as ellipticity (H/V) by changing the header in .hv file, then do the inversion. But, in the inversion, the observed data is automatically converted to ellipticity in degress by calculated them as atan(H/V); and the comparison between observed and calculated curved (ellipticity in degrees) seems off. I prefer to use the older version that consider the data as ellipticity in H/V. Can I change it in the new version?

Please, I hope that Marc can give us an explanation regarding that. I have been stuck on this for 2 months.


Re: H/V Curve Inversion

Posted: Thu Oct 10, 2019 7:13 pm
by EduardoPitta
Yes, the curve produced by the new version is really off, nothing like the old version. I really tried everything but couldn't come up with a solution that didn't include writing my own code.
I gave up on that and I'm now using the older version for Windows.
I'm running late so I don't have time to deal with it.
I thought that it would be possible to change the output in Geopsy, but I couldn't find anything.
I really hope someone could complete the documentation on the Wiki, I will try to contribute when I have the time.

Re: H/V Curve Inversion

Posted: Thu Oct 10, 2019 7:16 pm
by EduardoPitta
By the way, if you have any doubts regarding the theory behind the curve or anything related to that, send me a message. It's really hard to find support for that and I tinkered quite a while with that.
I'm now using HVTFA, I still have to learn how the calculations are made. Now my priority is to get the inversions right.

Re: H/V Curve Inversion

Posted: Thu Dec 05, 2019 12:44 pm
by admin
I fixed reading of H/V curve from geopsy into dinver. They are now properly converted to angles in degrees.
I also added an option to specify if the provided ellipticity curve is an absolute or a signed value (in mode properties).
These changes will be available in 3.3.0

Note that using H/V produced be the classical H/V module is not correct. H/V curves from ambient vibrations does not contain only Rayleigh waves. Body and Love waves are shifting the curve upwards (compared to the theoretical ellipticity curve of Rayleigh waves). HVTFA and Raydec (soon available, Hobiger et al. 2009 in GRL) are better options to obtain ellipticity curves. Effectively, HVTFA is not well documented. Information can be found in NERIES-JRA4 deliverables (european project 2006-2010). Unfortunately the project website stopped running some years ago. I have a copy of all deliverables and I recently made them available at (I fixed the DNS zone this morning, so you may have to wait at least 24h)