after studying the materials on YouTube about sound occlusion, I found 2 options. 1) this is to calculate collisions from the sound source to the camera (which can be costly in performance on 1 core) 2) this is to use steam audio, which is still not completely clear to me how to work with it at the software level