Hi all, In a TD game. I have at most 30 monsters in a scene (same time) They have hurt, attack, and die sounds. Which is the best way to implement (performance-wise) Put an audiosource to all monster prefabs. Playoneshot them by a script. Or implement observer mode (learn it first and play only 1 (or 2-3) audiosource in the scene and playoneshot all audios whenever it is needed via that only source. Which is the right way to go?