False alarms

One baseline, then session after session against it, with nothing changed in between. There is nothing to find, so every detection here is the instrument being wrong. Synthetic noise gives 0 out of 15. This is what a real room gives.