You have a sound, and need a variation of it

  • Yet you need it to still keep its specific sonic characteristics that distinguish the sound from other audio. (i.e. still sound like the same character, gun, vehicle, dinosaur, etc.)

  • What processes (DSP) and techniques do you use to achieve that slight psychoacoustic difference to prevent ear fatigue of hearing the same sound?


I call this part of sound design "building a sound vocabulary" or a "sound language," with which specific instances of sounds can be used that all family well together, as if voiced bt the same object or character. (Seems like this is something that's even more critical in non-linear media like games, perhaps moreso even than film, so I'd love to hear more thoughts from the game sound designers here on SSD...)

If it's synthesized, I use LFO's and other kinds of modulation, usually subtle, recording or printing many takes. [edit: the LFO's or other modulation are set to affect pitch, EQ, filtering, and/or dynamics, based on the context and needs of the sound or action being shown]

If it's based on real-world recordings, I do obsessive amounts of takes with way more variations than might seem advisable. When I get back to the studio and critically listen, the number of usable takes dips dramatically, so over-covering a sound source winds up yielding just the right amount of usable variations. Do this enough, with enough different layers, and these sound languages can be varied but sonically unified. I tend to stick to simple dynamics and EQ tools to even things out, instead of using heavy processing, to keep things sounding organic.

[edit: Since you mentioned dinosaurs and variation...] Randy Thom has said that the thing that impresses him the most on a demo real are emotive creature vocalizations, because that's one of the hardest things to do. Having tried it once or twice, I'd have to agree. Designing a happy sound, a curious sound, and an angry sound for the same creature can be damn tough. But it's like magic when it's pulled off well.

  • In my situation, it is a game. I am working with pre-recorded audio as the budget does not suit field recording of weapons. --- Thanks for the tips NJ! - Ill try some subtle DSP on a layer or two and see what works out.
    – C3Sound
    Dec 15 '10 at 15:57
  • @NoiseJockey great answer. Dec 16 '10 at 19:31

Noise Jockey's answer is a fantastic big picture answer and the ideal approach to sonic variation, IMO. But occasionally one gets called on to create variation in limited source material that can't be resourced again, or at least not in time. My first call of port in that particular storm is formant shifting.

  • @AdamAxbey OOO! Nice, ill try this - thanks Adam!
    – C3Sound
    Dec 16 '10 at 22:47
  • Yeah, hadn't considered that approach...very cool suggestion, Adam! Dec 17 '10 at 1:15

In my opinion (which I got from people smarter than I) you need actual sonic variation, not just processing. When I am designing a sound, there is often a series of elements that make up the final output. When you find the perfect balance of all your elements, make sure to print a series of variations of that mix. I would keep the processing relatively the same between your variations, but adjust the mix of your elements within each variation instead.

The changes to the balance of the mix elements should be minor, but enough to break a noticeable repetition of tone.

Once your variations are in game, you should of course add some random pitch and volume modulation on top of the variations.

Of course, make sure you are optimizing your sample rates to squeeze what you can out of that audio budget. No point bouncing to 44.1khz if the Nyquist theory will allow you to go lower.

  • I like that mix adjustment tip. Thanks Tom! I dont want to add new layers in for variations as that might be a bit much - adjusting volume though might give it just that slight difference in perception.
    – C3Sound
    Dec 16 '10 at 4:16

I am a game guy and avoiding repetition is very important. psycho-acoustically, your brain can recognize sounds within the first few mSecs (10-50ms as i recall). I prefer doing my more extreme varying with the tails so as nut to screw with the basic 'family' of the sound (small changes in mix/pitch over the various elements).

I also think personally that changing the rhythm of a sound sequence is huge in creating variation. As an example, a gunshot may have a slight metal hammer clink before the actual bang. I find that just changing the delay between the 2 elements (and also the relative mix) a small amount adds up to a very different experience. Even with the exact same wave assets.


Slight pitch changes, some eq, stretching, random parts messing (in loops of rain and wind, etc.) Classic techniques in general...


the sonic characteristics might not need to change, maybe only the visuals will be change enough..you could try adjusting the sync as well

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.