Seismosoft | Earthquake Engineering Solutions

Hello everyone,

I have been using SesmoSignal for a long time. Last week I was quite surprised when I compared the Fourier spectrum obtained from Seismo Siganl to the one obtained using Matlab (http://www.mathworks.com/access/helpdes ... /fft.shtml) for the same record. The frequency contents are OK but the amplitudes are quite different. Could anybody tell me the reason? perhaps the definitions of the amplitude are different??

Hi,

In SeismoSignal Fourier Amplitude is computed as follows: FourierAmpl(i)=sqrt[Re(i)^2+Im(i)^2]/df, where df is the frequency step. In Matlab, a different expression is probably being employed.

Seismosoft Support

PS: by mistake, in our first answer to this question, we quoted the Power Amplitude equation, rather than the Fourier Amplitude one; apologies for this blunder.

thank you very much for your prompt response. May I know the reference to the Fourier amplitude definition being used in SeismoSignal?

Not sure we understand your query.

The definition of Fourier Amplitude is a classic expression extensively referred in the literature and the web (just do a Google search or visit Wikipedia).

In SeismoSignal, a small difference with respect to the classic formulation may perhaps be the normalisation by dt, which however may be avoided by users who so wish (by simply multiplying the results by dt).

Seismosoft Support

Hi
I read your answer but I could not undrestand your function?
please help me by write the fourier transform function in seismosignal clear such as MATLAB.
thanks

Hi,
What does it mean?
FourierAmpl(i)=sqrt[Re(i)^2+Im(i)^2]/df
I think it could be correct by this:
fft=sum(x[t]e^-iwt)