filtering

02-Analytical capabilities
lanlin
Posts: 1
Joined: 08 Sep 2009, 12:55

filtering

Post by lanlin »

I have a question about filtering.

I want to define butterworth bandpass filter, with lower frequecy is 0.1Hz, high frequecy is 70 Hz, and the order is 4. However, the plot gives me a vertivally straight line connecting a point with gain of 1 at 50HZ, and another point with gain of 0 at 50Hz. I do not think it is a correct plot, since at the higher frequency region, I should get a curve instead of a straight line.
Please point out my mistake.
User avatar
seismosoft
Posts: 1184
Joined: 06 Jul 2007, 04:55

Re: filtering

Post by seismosoft »

You are right. Our filter currently saturates at 50 Hz (something we always assumed to be fine for standard earthquake engineering applications). In any case, we will see if this limitation can be overcome.

Seismosoft Support
chiara
Posts: 2
Joined: 28 Jun 2011, 07:05

Re: filtering

Post by chiara »

I have the same problem.
I would like to cut frequency higher then 25 Hz but in the fourier spectrum the frequency arrives until 100 Hz instead of I have cur tham in advance.
Is that a mistake of the program?
rishi
Posts: 1
Joined: 05 Oct 2012, 23:25

Re: filtering

Post by rishi »


I also have the same problem.
I would like to cut frequency higher then 10 Hz but in the fourier spectrum the frequency arrives until 100 Hz instead of I have cut them in advance and the last value will be so more that straight line appear there.
Is that a mistake of the program?
User avatar
seismosoft
Posts: 1184
Joined: 06 Jul 2007, 04:55

Re: filtering

Post by seismosoft »

Dear rishi,

Unfortunately, we cannot understand the problem you are trying to describe.

Seismosoft Support
sasansas
Posts: 2
Joined: 22 Feb 2014, 12:13

Re: filtering

Post by sasansas »

Hi ,



I have a question about filtering in seismosignal. How it performed. I need to filter frequencies above 40 hz.The following are selected:
a. butterworth
b.bandpass
Questions:
1. Concept of order = ???
2. freq1 always 0.1
3. freq2 = 40

*But the Fourier spectrum does not change#1567;#1567;#1567;


Thanks,



Best regards,
huffte
Posts: 978
Joined: 22 Jul 2011, 10:19
Location: Cookeville, Tennessee, USA
Contact:

Re: filtering

Post by huffte »

Hi sasansas.

It sounds as though you need to use "lowpass" rather than "bandpass" filtering.

If, in fact, your sole criteria is to suppress frequencies higher than 40 Hz, then the lowpass filter with Freq1 = 40 Hz would seem to be the appropriate choice.

The Help System section entitled "Loading & Pre-processing" and, in particular, the sub-section entitled "Baseline Correction and Filtering" provides both a more detailed discussion and several references which should be consulted when deciding upon baseline correction and filtering options.

Best of luck sasansas.
Tim Huff
sasansas
Posts: 2
Joined: 22 Feb 2014, 12:13

Re: filtering

Post by sasansas »

hi,

I thank you for the perfect answer.
Would you explain the items: order
Do I use the default value: 4

Best regards,
huffte
Posts: 978
Joined: 22 Jul 2011, 10:19
Location: Cookeville, Tennessee, USA
Contact:

Re: filtering

Post by huffte »

I am afraid that I cannot claim to be an expert on signal filtering, sasansas. So I would highly recommend that you do one or both of the following:

1. Research the literature on the subject. The references given in the Help System of SeismoSignal are an excellent starting point and may, in fact, be all you need.

2. Read everything in the excellent Help System pertaining to filtering and correction.

You might also wish to do a sensitivity study on the effect or order upon the resulting Fourier Spectrum. But without looking into the literature and investigating the science behind filtering, this would be of little value in itself.

Best of luck sasansas. This field of study is really an interesting one.
Tim Huff
aiswaryamenon
Posts: 3
Joined: 17 Apr 2015, 12:18

Re: filtering

Post by aiswaryamenon »

Hello,
I have the acceleration time history from which I have to get the velocity and displacement time history so as to get PGV and PGD. For some of the earthquake records that Iam having the PGV and PGD values as well as response is very much different from what I should get . So can somebody suggest me what all points should be considered for processing acceleration time history to minimize the noise in the obtained velocity and displacement time histories . How to select the frequency range in which filter has to be applied for noise removal by seeing the acceleration , velocity , displacement time histories obtained by integration and double integration or fourier amplitude spectrum ?

Thanks and regards
Aiswarya
Post Reply

Return to “02-Analytical capabilities”