There has been some discussion of either adding exclusion zones to model fitting, and/or fitting different frequency ranges for aperiodic & periodic fitting. This is something that could be broadly interesting, and useful, but I think it needs a bit more looking to figure if and how to best adopt it.
Things to keep in mind:
If there are 'gaps' in the fit spectrum, we have to consider if and how to fit peaks that might be next to, or overlap with, these regions.
Related: we have procedures for dealing with peaks near edges, that might need to be extended
The logic of the model recommends that to fit either periodic or aperiodic requires dealing with / fitting the other component. Though I think there could be reason and utility to fit different ranges / exclude regions, I don't think we want to lean too far into fitting these as separate processes.
Related prior discussions:
Exclude a frequency regions for fitting
This discussion led to adding the interpolate_spectrum function
There has been some discussion of either adding exclusion zones to model fitting, and/or fitting different frequency ranges for aperiodic & periodic fitting. This is something that could be broadly interesting, and useful, but I think it needs a bit more looking to figure if and how to best adopt it.
Things to keep in mind:
Related prior discussions:
interpolate_spectrumfunction