I am not sure how it works in most game engines but would the handling of the sound environment be part of the engine or up to the specific developer to implement?
I would assume something as massive as Outerra would need to handle sound reflections and reverb on a grand scale. A gunshot at that airport should have some incredible refractive echo sounds through the mountains. Are you planning to implement anything like that?
I ask this only to get to the topic of voice communication. Most internal game VOIP communications sound terrible (TF2-CSS-ArmA-etc-etc) to save network overhead load on servers moslty. I have a ventrilo server
http://tinyurl.com/28qsary that I use with my friends that sounds fantastic but does not really "help" when in a game and someone screams .. HELP ME and you have no idea where they are.
What would be useful and correct is if Outerra can use a set of assignable key functions. One for Speech from the player body. Then one for a selected Radio Channel. The way I see it. Yelling into your mic should allow anyone within "earshot" to hear you on a decay based on distance and air density etc etc. Speaking on the radio would be audible from the players body of course but transmitted a few miles to anyone on that said channel with a radio on. Some distortion could be added to tell it apart from normal one on one in person voice communications.