While not specifically a spectral measurement, the cross correlation function can be efficiently calculated in frequency domain. AFAIK, Python tools are missing a good truncated cross-correlation that would work on ultra-large data like we are focusing on. So, e.g., a +/- 1000-sample cross-correlation for signals of order 1e9 samples. I think there is an equivalent overlap based approach for this, but it would take some thinking about.
ckemere
commented
10 months ago
While not specifically a spectral measurement, the cross correlation function can be efficiently calculated in frequency domain. AFAIK, Python tools are missing a good truncated cross-correlation that would work on ultra-large data like we are focusing on. So, e.g., a +/- 1000-sample cross-correlation for signals of order 1e9 samples. I think there is an equivalent overlap based approach for this, but it would take some thinking about.