1 method i can recommend is an "attack" or an idle animation , this attack has a very brief animation frame, & a certain range
after the frame, you call a command to the sound you need.
you may also need a script to go with this to make the detection range a big "3d" space (that way your character can "see" the other & do the "attack"
and another script that helps your character detect the other specific character.
another method is to spawn an invisible entity when the 2 characters meet, this invisible entity has a spawn animation with invisible frames , between the frames you put the command for the sound, script may be needed for this detection to happen dynamically (independent of stage coordinates)