fix bug in velo sensor masking

Together with @ahennequ we discovered that in the masking tool the hit numbering was not correctly incremented such that once there are sensors that are turned off, the hit index does not point anymore to the right position and the masking can create very nasty effects. Pinging also @ldufour and @mwaterla, apart from us the main users (and actually the developers) of this tool.

To show a bit the effects this has, an illustration of the VELO track distributions from 2023 MC (VELO open). When switching off the even modules (C side), we saw before: When switching off the odd modules (A side) however, we saw: The behaviour with the A side is expected, the behaviour with the C side completely buggy. After this fix we see the behaviour as expected when masking the C side: