Top ten things a video game producer should know about creating great audio

I recently requested input from you all on this topic for a paper that I was working on for GDC. I got some great quotes and thought that I would publish them here.

- Mark Miller <mark@groupprocess.com>

• "Just because it sounds like your favorite band doesn't mean it's the right thing for this project."
• "No, you can't play the guitar solo."
• "Listening to a lot of records isn't the same as knowing how to make one."
• "Quality costs money. If you're doing it cheap you're either not getting quality or you're ripping someone off."
• "Tell me the emotion you want to create, not how to write the drum part."

- Spencer Critchley, Director of Production, Silicon Gaming, Inc. <spencer@ubet.com>

• To quote our dear compatriot Brian Coburn: "People don't really like music, they just expect it..." That just might be near number one(!)
  

- David Javelosa, 31st Century Noise Design, <davidja@brandx.net>

• 1) MIDI is a control protocol, not a sound. ("What you hate is the sample set, not the MIDI.")
• 2) You'll get better sound for the same money if you amortize audio department infrastructure across multiple titles rather than expensing against individual projects. ("We know the testers are logging bugs on SGIs, but we just paid for all those massages, so we can't afford that used PowerMac 8500 -and- the AudioMedia III. Can't you just do the game with SoundApp instead?")
• 3) Avoid build processes and APIs that make the sound artist go through engineers to make sound-only changes.
• 4) Just because your cool 3D world geometry engine looks great because it's physically accurate doesn't mean adding a geometrically accurate sound model is gonna sound right. ("But the box says it's -accurate-! It -can't-sound bad!!!")
• 5) Don't put too many eggs in the 3D audio basket, game-design-wise. ("Y'know, with only two speakers, only the .helicopters. can go behind you...")
• 6) Latency .really. matters; a button click sfx heard 200 mS after you click the mouse isn't a button click any more. ("200 mS isn't a delay, it's a postponement!!")
• 7) When all games are using the same sample set you are, there's a limit to how much better you can make your game sound. ("Budget realistically for custom samples!!")
• 8) Dialog recording sessions are like motion capture sessions: turning an actor's output into usable elements can take a -lot- of cleanup work. ("Budget realistically for dialog editing!!")
• 9) When the programmer says that what the sound guy wants is impossible, the programmer is frequently, er, lying. ("How can you tell if the programmer's lying to the sound guy? His lips are moving!!!")
• 10) Allow calendar time in the schedule to review the sounds in context (with real players) for a while, and then get one or two rounds of revisions in. A sound design concept takes calendar time to ripen, and it can take some hours to smooth rough edges in the first implementations.

- Chris Grigg <chrisg@sirius.com>