Closed markbandstra closed 3 years ago
I should also note that this appears to be the only location where this bug crops up. This change may require users to adjust existing code that uses the autocal
tools, although the PeakFilter
kernels are pretty robust (which is why we never noticed that problem) so it is also possible that no changes will be needed.
Ah, autocal tests fail
Dangit, always run the tests. I'll look into it.
A user with a keen eye noticed that the
PeakFilter
class has been using an erroneous formula for the spectral energy resolution:which interpolates between
f(0) = f0
andf(x1) = f1
, but it scales roughly asx
, not the intendedsqrt(x)
.This formula has been changed to
and the notebook has been brought up to date.