Because resolution and transmission don't commute, we have to compute the effective transmission of the convolved flux to calculate the calibrated flux and inverse variance per camera. This fixes the flux conservation problem identified for lines in the previous commit.
This flux conservation "bug" was actually quite dramatic for a continuous source flux in regions of sharp atmospheric absorption features.
Ex: before
after
also, for the single line at 5800A, the integrated flux is conserved at better than 1e-6.
Because resolution and transmission don't commute, we have to compute the effective transmission of the convolved flux to calculate the calibrated flux and inverse variance per camera. This fixes the flux conservation problem identified for lines in the previous commit. This flux conservation "bug" was actually quite dramatic for a continuous source flux in regions of sharp atmospheric absorption features. Ex: before
after
also, for the single line at 5800A, the integrated flux is conserved at better than 1e-6.