Chris_Halkides
Penultimate Amazing
- Joined
- Dec 8, 2009
- Messages
- 12,981
Window functions and Fourier transforms
Maybe some pictures would help. The free induction decay is a set of data in the time domain (like a musical note or chord being struck and dying away), and the Fourier transform converts the data into the frequency domain (like musical tones). Notice how the spectrum changes when one applies different window functions: exponential, Gaussian, shifted Gaussian, sine, and shifted sine. Different kinds processing improves different qualities of the spectrum, such as signal to noise, resolution between peaks, or reduction of artifacts. However, if someone handed you a paper copy of a frequency spectrum, you would have an extremely difficult time to go backwards to the original free induction decay and then process the data set differently. This is only offered as an analogy with respect to EDFs and egrams.I once likened the process of converting an EDF into a egram to doing a Fourier transform of a free induction decay to a frequency spectrum in nuclear magnetic spectroscopy. The data are easier to interpret in the latter form, but certain processing parameters have been irrevocably applied. The choice of those parameters should be in the hands of the experts doing the review, not the prosecution's experts.
Last edited: