Dear experts,
I am using a convolution of a Crystal-Ball and a Breit-Wigner using the RooFFTConvPdf to model the signal pdf in the Z’ analysis.
In the past we were using a Voigtian for modelling the signal pdf, in this case the scale variation was modelled by a function that was plugged into the observable parameter of the Voigtian, allowing then the peak to shift to find any potential excess close-by the evaluation point.
However, RooFFTConvPdf constructor seems not to allow a function as an observable as the Voigtian allowed… Is there any work around to this?
Many thanks,
Santiago