Only create the sound fade modulator once the sound starts fading, saves the allocation for sounds that never fade So we never pay useless modulator allocation