Issue No.05 - September/October (2005 vol.7)
Bert Rust , Siena College
DOI Bookmark: http://doi.ieeecomputersociety.org/10.1109/MCSE.2005.103
In Part I, we introduced the idea of windowing a time series before estimating its frequency spectrum. In addition to some fundamental elements, we examined zero padding, aliasing, the relationship to a Fourier series, and windowing. One disadvantage of windowing is that it alters or restricts the data, which, of course, has consequences for the spectral estimate. In this installment, we continue our discussion from that first installment with a more general approach to computing spectrum estimates via the FFT.
fast Fourier transform, FFT, spectral analysis, time series data
Bert Rust, "The Fast Fourier Transform for Experimentalists Part III: Classical Spectral Analysis", Computing in Science & Engineering, vol.7, no. 5, pp. 74-78, September/October 2005, doi:10.1109/MCSE.2005.103