sonofableep

after reminding matt to never again work on one sound for 3 hours straight (buh-bye objectivity!), i cleverly repurposed that dictum to apply to design benders like the one i’m on right now.  “never brainstorm / problem-solve on a single topic for 3 hours straight”.  right.  then i proceeded to do just that.

i’ve been trying to work out the nuances of our server-triggered dialog integration.  a simple enough nut to crack, eh?  but 8 hours later i’m still at it.  and my eyes feel like they have ground glass in them (what’s the max LCD exposure time in a darkened room before one goes legally blind, anyway?).  i’m trying to avoid all the pitfalls i hear in other games, including those i’ve worked on.  dialog cutoffs, mix-wrecking volume levels & fades, unnatural sounding positioning & attenuations, you name it.  and i can’t…quite…seem…to get there.  argh.  my notes here go a little something like this…

define settings based on balancing the following (seemingly at odds) things:

• gotta feel right for “in-person” and “disembodied” speakers.

• intelligibility throughout broadcast area, regardless of distance from speaker.  twiddle max distance and volume/ low-pass / spread curves…probably a lot.

• maintain the sense that the dialog did come from the onscreen source.  this rules out simply making “in-person” dialog 2D.

• can’t be drastically different in attenuation from “standard” settings, or else juxstaposing dialog and combat/emote/misc audio will sound jarring.  this rules out the heavy twiddling required above, making the jarring pretty much unavoidable.

• the player might leave the speaker, crossing the max distance before the dialog has finished playing.  can’t pick a solution for this until the other stuff is solved.

• the player might leave quickly (transportation) or Very Quickly (teleportation).  if your radius is big enough, transporation might be okay.  if there is teleportation, prolly do a big teleport sound to mask the cutoff.  there’s really no way around it.  w00t.

• use these settings as a jump off for 3D-xfading-into-2D-disembodied speakers…then use *those* settings as a jump off for full-time 2D-disembodied speakers.

• can we even arrive at one solution that works for the trivial VO (“Thanks!”) as well as the epic (Uber-Villian monologue)?

• the player might be in a different room than the speaker – yet very close distance-wise – when the dialog is triggered.  will the player get the message?  if not, cool.  if so, do we rely on occlusion / obstruction to do it’s thing, or is more required?

and the non-stop metalcore isn’t helping my focus at this point.  time to heed my own advice, throw in the towel, and go buy Mass Effect 2.

and the cleaning guy just wandered into my office.  what’s he doing here after midnite?  probably wondering what *i’m* doing here after midnite.