Determine if group delay is something that we need to consider, and fix it if so. #17
Description
Is your feature request related to a problem? Please describe.
It seems unlikely to me that the vasculature would be completely non-dispersive, but that's the implicit assumption we make when using the same regressor at every time delay. In general, there should be differential propagation speeds of the different frequencies in the sLFO. We should deal with this somehow.
Describe the solution you'd like
Automated group delay compensation.
Describe alternatives you've considered
Ideally, we should make a set of regressor shapes at various time delays and look at their spectra. The magnitude of the spectrum seems to be fairly constant, but we could fit the phase delay in a frequency dependent manner to find the group delay and correct for it.