is there any delay inside the other fsms?
with triggers i usually set a separate fsm with a 2 states turning a bool on/off
with a trigger stay and trigger exit.
Then on the other fsm do a fsm bool test.
but if you have multiple of these sound 'listeners' in a scene you should not set it up in this way (performance wise)
you should set a 'checker' on your player.
and use for for example overlap actions instead of trigger events.
when it found a sound object, send a event to play it, then add to a list.
when its not found the next time, send a event to stop it and remove it from the list.
the overlap you can do for example every 0.2 sec (as every frame, I don't think is required for this)
trigger events are called every frame, then multiplied by your sound object count.
so if you have 10 sound objects you are checking 10 x every frame.
the overlap sample does 1 call every 0.2 sec