Some Practical Considerations of Coherence Estimation.
NAVAL UNDERWATER SYSTEMS CENTER NEWPORT RI
Pagination or Media Count:
Given two processes the complex coherence spectrum is the complex cross spectral density function divided by the square root of the product of the two auto spectral density functions. Estimation of coherence in light of the fast Fourier transform FFT is investigated for synthetic data. The procedure used is to segment the given finite time histories to NSEG segments of size NNN. Each segment is multiplied by a weighting function in this case a cosine bell prior to computation of the FFT. Cross and Auto Spectra are then averaged over the NSEG segments prior to forming the coherence ratio. Practical conclusions drawn are that for non-flat auto spectrum, multiplication by a weighting function is necessary that NSEG must be on the order of 64 that NNN must be large enough to insure sufficient spectral resolution.
- Statistics and Probability