Clarify params

[Adre17]

Hi,

i've extrapolate target.target in xml format

<RealStatisticalPoint>
<x>38.648400000000002308</x>
<mean>34.460796414021615419</mean>
<stddev>1.4712502003305514275</stddev>
<weight>47</weight>
<valid>true</valid>
</RealStatisticalPoint>

<stddev>1.4712502003305514275</stddev> this value give from?
<weight>47</weight> what is this?

thanks a lot

[admin]

The stddev is calculated on a log scale, this is exp(sigma_log). It is hence a multiplicative factor to the mean to get the one-standard deviation range. In this case you have a 47% error range.

The weight is the number of samples used to compute the statistics. In other words, it is the number of time windows for which the process was repeated. It is useful to combine results computed separately (for example, an array recorded at one year delay). The combined statistics are computed exactly as if there is a single process on the complete set.

[Adre17]

Thanks a lot.

this line
<log>Values loaded from Geopsy HV output file C:/Users/umber/Desktop/Molino_Albergati/Molino_albergati.hv.
</log>

in xml file is necessary? if i write manually xml , there is no path source..

[admin]

The log is no mandatory, you can safely skip it.

[Adre17]

Ok it's work.

Just a more question, if i've an HVSR curve freq (Hz) and Amp (H/V) i need to convert H/V value in Ellipticitu (deg) atan(H/V) and calculate stddev in log scale to give dinver correct input or it take also amp (H/V)?

[admin]

You have to specify the atan(H/V) in degrees. The stddev is not on a log scale for ellipticity angle. There is a minor bug in the latest release to display stddev: in the legend, set the line thickness of symbols to something other than 0.

Be careful that a HV calculated in the classical way is not a Rayleigh ellipticity curve. A better approach is Raydec (Hobiger et al 2009) or HVTFA implemented in geopsy. The first option should be soon added.

[Adre17]

ok,

so using file .hv is not the ideal way, is better to use outpout of HVTFA.

I've obtained file test.max from TFA function, then i've modified file adding column of delay value (is missing in output ) and i added for all value of 0.25 (find in in precedent post) is correct?

I load this file to max2curve to obtain ellipticity curve and i get this output

Code: Select all

# Curve name:# 1
# Begin curve log
# 
# End curve log
# | Frequency (Hz) | Ellipticity (deg) | Stddev | Weight | 
0.500878 2.4412997610638825 4.620804002331228 73 1
0.5134 2.4740387542849427 4.715787681419612 80 1
0.526235 1.7037881259669962 4.660730452532595 76 1

is correct header?
Now i can load this in dinvert for inversion , firrst column is Freq, secon column is Ellipticity atan(H/V)? Third column atan(H/V) deg stddev? last remanin weight, there is a strange value for Ellipticity (deg)..



Thanks

[admin]

The fourth column is the weight (the number of samples involved in the mean computation), the fifth column is a valid flag (normally always set to 1).