Handling of Birks suppression probably needs improvement
FLSHits report the distance a particle traveled through a cell and the total energy lost during that journey. Dividing these numbers gives a average dE/dx for the particle, but dE/dx shoots up at the very end of the track, so the average dE/dx across the cell is a poor proxy for the net suppression for that cell. In fact, the way the numbers work out, I suspect that we get essentially no Birks suppression at the moment, since the turn-on is fairly non-linear.
Approaches to fixing this:
- have g4nova break up FLSHits when the instantaneous dE/dx changes too much
- have g4nova calculate Birks suppression internally in its stepping, with each FLSHit recording the net integrated suppression over the pathlength represented by that FLSHit (preferably in a separate field rather than via direct modification of Edep.)