I am working on modal testing of structures using forced hammer excitation. Response time
series has a sampling frequency of 250Hz and the excitation time series has a sampling
frequency of 500Hz. Should the plot of FRF magnitude versus frequency be as a function of the
response or excitation frequency? Or which sampling frequency would govern the FRF plot?

Secondly, for a single-point excitation responses are measured at 5 different points. Is it
possible to calculate the FRF as the ratio of FFT(X)/FFT(P), where X is a (n x 5) matrix
representing the response time histories from the 5 measuring points and P is a vector
representing the excitation time history.

Thank you for the anticipated prompt response.