An audio pouch is provided that secures an earpiece of a headphone assembly adjacent to the user's ear when worn. The audio pouch includes an inner panel and an outer panel that sandwich an earpiece of the headphone assembly therebetween, such that the earpiece is disposed adjacent to the user's ear when worn. The audio pouch includes attachments that retain helmet straps. The audio pouch is particularly effective with protective helmets having Y-straps extending from the helmet for attaching chinstrap thereto.