Add the ability to normalize the reconstructed spectrum to match the data, which allows to work with spectra that are not flux-normalized.
In the first spender papers, we needed to normalize the input spectra in restframe, which requires that we can access observed spectral elements of some reference restframe wavelength (we chose 5300-5800A), but this limits the range of redshifts over which this normalization can be applied.
This PR brings normalization on-the-fly option. The normalization is still computed in the restframe segment shown above, but now from the restframe reconstruction (rather than the de-redshifted observation). We then solve for the maximum-likelihhood estimator of the constant factor that needs to be multiplied with the reconstruction to match the data.
Add the ability to normalize the reconstructed spectrum to match the data, which allows to work with spectra that are not flux-normalized.
In the first spender papers, we needed to normalize the input spectra in restframe, which requires that we can access observed spectral elements of some reference restframe wavelength (we chose 5300-5800A), but this limits the range of redshifts over which this normalization can be applied.
This PR brings normalization on-the-fly option. The normalization is still computed in the restframe segment shown above, but now from the restframe reconstruction (rather than the de-redshifted observation). We then solve for the maximum-likelihhood estimator of the constant factor that needs to be multiplied with the reconstruction to match the data.