[Antelope] question regarding dbdetect

[Francesco Civilini]

To whom it may concern,

I have found the solution regarding the problem I previously wrote to  
this community about (using dbdetect for strong motion detection).

The problem had nothing to do with the amplitude of the wave. Because  
we are running dbdetect on segmented data, the processing time window  
(process_twin) in the default parameters needed to be changed to  
something much lesser than the default (30 sec). Now dbdetect is able  
to detect arrivals in all the waveforms. The only downside is that  
running dbdetect takes a far longer time than it did before.





Quoting Francesco Civilini <fcivilini at umail.ucsb.edu>:

> To whom it may concern,
>
> Is it possible to use the program dbdetect in Antelope to  
> automatically mark first motion and second motion arrivals in  
> segmented strong motion event data?
>
> I?m currently working on a project that uses antelope to view and  
> analyze segmented event strong motion data. We can use dbpick  
> function to mark the starting time of the first motion as well as  
> the secondary motions (P and S waves). The goal is to ultimately be  
> able to make this process automatic through use of the dbdetect  
> program.  However, the parameters of dbdetect (sta_win, sta_min,  
> etc), are insufficient to make this process automated.
> 	This is because each waveform has unique amplitudes  that vary by a  
> large margin. The detection uses the parameters STA (short time  
> average) and LTA (long time average) to calculate a ratio to compare  
> with a threshold parameter. If this ratio is greater than the  
> threshold, a detection occurs. This ratio, however,  varies greatly  
> depending on the amplitude of the specific waveform.  The result is  
> that the ratio might be sufficient to trigger detections in some  
> waveforms but not in others. For example, a large ratio will trigger  
> detections in waveforms with large amounts of high peaks in  
> frequency amplitude, but not in a waveform with very low peaks. The  
> similar problem occurs when the ratio is made smaller: there are no  
> detections in the waveforms with the high peaks.
> 	Are there any papers or documents that regard automated, strong  
> motion demarcation using Antelope that could provide a clue to  
> solving this problem?
>
> Thanks,
>
> -- 
> Francesco Civilini
> fcivilini at umail.ucsb.edu



-- 
Francesco Civilini
fcivilini at umail.ucsb.edu