Smarter handling of inaudible continuous sound effects

[olofson]

Currently, the continuous sound effects are (almost) always started unconditionally, and keep running until killed, whether or not they're actually audible. This is because KOBO_sound::g_move() wouldn't know which sound effect to start after the fact, as the program handle is only passed to g_start()!

(The only exception is when sound is muted, in which case ALL continuous sound effects are restarted as needed from a higher level, where all the program handles are available.)

The easiest way to deal with this is probably to pass the program handle to g_move() as well, so it can just start the intended sound if it isn't running already. We could probably just replace g_start() and g_move() with a single g_run() call, to avoid confusion.

This might not seem like a big deal, but there are a lot of cores on some levels... And there will most likely show up a lot of new continuous sound effects as well. As I've probably mentioned elsewhere, I'm considering using a single A2S program for each object, encapsulating all sound effects for the object, for more flexibility - similar to how the core sounds actually respond to damage, rather than just triggering additional "damage" sound effects.

[olofson]

Note to self: This needs fade-out + kill - not stop + release! We don't want to trigger (inaudible) death sound effects and stuff.