Audio guides and (interactive) sound walks have existed for decades. Even smartphone games taking place in the real world are no longer a novelty. But due to the lack of a sufficient middleware which fulfills the requirements for creating this software genre, artists, game developers and institutions such as museums are forced to implement rather similar functionality over and over again. This paper describes the basic principles of Real World Audio (RWA), an extendable audio game engine for targeting smartphone operating systems, which rolls out all functionality for the generation of the above-mentioned software genres. It combines the ability for building location- based audio walks and -guides with the components necessary for game development. Using either the smartphone sensors or an external sensor board for head tracking and gesture recognition, RWA allows developers to create audio walks, audio adventures and audio role playing games (RPG) outside in the real world.