Audio LOD Questions

I have a couple of questions regarding the loading and "LOD"ing of audio files.

What is the audible range of a sound? Is there a formula? Is it dependent on the size of the object the sound is attached to?

When does llPreloadSound happen? When the agent enters the sim? Or when the agent comes within the audible range of the sound? I suspect the former, which makes me wonder if there is any practical difference between setting up a sensor which starts/stops sounds based on proximity (with llPreloadSound), and just having the sounds attached to objects without llPreloadSound or prox sensors. In fact (if the audible range and the loading time were both acceptable to the author) wouldn't the latter scenario be *more* efficient since it would not be using any processor time with sensors etc?

Looking forward to replies, and apologies if the questions are dumb or unclear.

best as I have been able to tell. Audio emits from the center of the prim, and is based on the volume setting.

The sound emits from the prim that contains the script. Your client "ears" are attached to the camera so zooming in on a sound source will make it louder.

Each player can set preferences for distance and rolloff effects (and their own system gain) so you have no idea how loudly any particular listener hears the sound.

At last check llPreloadSound() was treated as a suggestion to the server to send a sound and this request priority was placed way below textures and geometry so it could be a long time before that sound is preloaded (it may not ever get around to pre-loading, I'm not sure).