Ambients in ARE

[divide]

I have played a bit with ambient sound in PST. I've found IESDP description of some fields meaning (of ambient description in AREs) inaccurate. Also I have gained some insight into how ambients are played. Specifically:

• Field 0x0088 labelled 'Number of sounds to use per set' actually has nothing to do with number of sounds. This is maximum deviation from the interval set in 0x0084. So the interval (between consecutive sounds; between the times they are started, that is, no matter the length of the sound) is chosen randomly (from a uniform distribution) from [interval-maxdeviation, interval+maxdeviation] (with single second granularity). The maximum deviation is capped over by floor of interval/2, so anything over it is treated as this value.
  
• The sounds are not chosen randomly; they are played in the order they are given in a file (for interval ambients, that is). Continous ambients always have only one WAV looping (more than one confuses the engine, as it tries to play them all simultanously).
  
• All ambients that doesn't have 'main ambient' flag set in 0x0090 are always at######d with distance and cut off at radius. Also, it seems that radius does not only control the cutoff distance, but also intensity of the attenuation.
  
• Two other flags seem to have no meaning at all. As far as I can tell, in PST the 'point sound' flag is set on all continous ambients, and 'area sound' flag is set on all interval ambients. Setting them to anything else seem to have little effect. The only effect I've noticed is that unsetting 'point sound' flag on continous ambients confuses the hell out of engine and it plays only the first sample of WAV infinitely.
  

[igi]

Added to local copy. Thanks for pointing it out

[Guest Guest]

And where exactly is this local copy? I cannot see any changes on http://iesdp.gibberlings3.net ...

[CamDawg]

Local copy means igi's hard drive, awaiting his next update to the site.

[divide]

Oh, I see... that's good to know

[divide]

BTW, I would be grateful if somebody could confirm my findings... I'm not really sure about anything but the timing thing (I've measured it for an hour with a stopwatch in hand, and then analyzed the statistical distribution to confirm that it's really a white noise, and not Gaussian distribution, for example and the sound cycling (as opposed to random selection).

 

Especially I'm interested in whether there are any ambients in the wild that doesn't have flags set in the manner I have observed ('point sound' for continous ones, 'area ambient' for interval ones). If there aren't, then these flags seem to be simply an additional indication (besides interval being == or != 0) of these types.