Remove seismometer signal from the transfer function of the instrument

Dear All,

i need to remove a seismometer signal, in displacements, from the transfer function of the instrument.

Can you suggest me how to do it?

Thank you in advance.

[Mathieu NOE]

hello

we need a bit more info to help you

a transfer function can be computed with a input and an output signal

here we have only one signal , and it seems you consider it as a "perturbation" , so where are the input / output data for the tf computation ?

[Alessio Lodato]

Good Morning,

thank you for your availability, i have only the output of the seismometer for now, and i have the numerator and denominator of tf.

Do you mean, i need also the input ground displacement to the seismometer?

Thank you,

AL

[Mathieu NOE]

hello

so you have indeed already a transfer function .... but I don't get what it means "removing a signal from a tf" ??

it's like comparing apples and oranges.. you can do math between signals or tfs , but between a signal and a tf the only thing I can do is to filter a given signal by a tf , but that has nothing to do with "removing" ...

is this an homework ? do you have a more refined explanation of what is the intention of this ?

all the best

Answer 1

[Alessio Lodato]

Hello Mathieu,

glad to hear you. Yes i need to filter output signal from the transfer function of the seismometer.

I need it for research activities.

Can you suggest me how to find the exact procedure to do it?

Thank you again.

AL

[Mathieu NOE]

hello Alessio

look at filter and filtfilt to filter a signal by a given transfer function

[Star Strider]

This is a problem generally with instrumentation. The instrument (in this instance a seismometer) has its own characterisrtics, and filters the ground motion presented to it according to those characteristics. It is possible to estimate the transfer function of the seismometer if a record of its responses to a specific input (ideally an impulse of known amplitude) is available.

With respect to ambient recordings however, the ‘inverse problem’ — reconstructing the original signal from the recorded signal and the transfer function of the transducer — is generally not possible, simply because the frequencies not present in the recorded signal (filtered out by the transducer dynamics) are lost forever.

I have not worked with seismometers, however I have worked with physiological instrumentation, and encountered this problem in that context. I am not aware that there is any solution for it.