0

This is largely hypothetical: the answers I receive here will determine whether or not I and the peope involved even move forward with this project.

I am part of a group hoping to build a series of monster suits to be worn at cons or special events, in the vein of kaiju suits and WALKING WITH DINOSAUR single-man suits. It's been our hope from the beginning that the suit actor inside would be able to make their own sounds inside the suit via a voice-modulation system hooked up to short-range speakers inside the suit. We feel that's largely doable, but here's the hangup, and the reason I'm asking:

(1) Is it possible to make modulation system that specifially mixes anything you say or utter with a pre-existing sound? For context, say, an assortment of stock dinosaur noises, or a large sampling of classic Godzilla roars, etc. The hope being that whenever the suit actor makes a noise into a microphone, the system mixes that sound with one already in the system, so that it comes out of the speakers sounding like Godzilla or a T-rex, etc., instead of some joker in a suit making goofy noises.

(2) If it is possible...where do we even begin to look to make that happen?

We have seen that it's somewhat possible. I can't find the video at the moment, doing a cursory scan of YouTube, but we observed a setup where some employees of either Stan Winston or Jim Henson were making a giant mobile suit to take to a con, and they had a real-time voice-modulation set-up that allowed the suit-actor to voice a high-pitched charater riding the monster-suit. We've also seen how voice-modulation is pretty prevalent in the gaming community and how streamers can sound like basically anyone or anything, even familiar and established characters (I just watched a video where someone set up a Pennywise setting for when he played Fortnite, and it was very believable).

We're hobbyists, and this is our first foray into something like this, so we're not sure if this is possible, but we hazard a guess that maybe our set-up might theoretically be easier than all these other examples, since all we're wanting to do is translate grunts into Godzilla roars, etc.

Please, any help would be greatly appreciated. Thank you.

EDIT: Here is a link to the video I mentioned. I finally found it. This might at least give you an idea what I'm talking about, even if our project is decidedly less ambitious.

1

Anything is possible - just depends on how much you want to throw at the problem. The people involved need one or all of the following:

  • Time
  • Expertise in electronics and digital processing software
  • Budget

.... or a willingness to engage someone who does.

From a hardware standpoint you will require a wearable, battery powered computer that can sample and process voice input in real time. Raspberry Pi may be a suitable platform. The processing involved will require:

  • Voice detection and triggering
  • Voice processing
  • Sample playback

There are embedded computers that are fast enough to be able to handle this sort of application. You would probably be fine working at a 16kHz or 24kHz sampling rate which means that the CPU processing and energy consumption could be kept to a minimum.

It is likely that the software would need to be custom written for the application, but there are suitable libraries in the public domain that would be able to be leveraged for this purpose.

For instance:

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.