How do I turn ordinary speech into something inaudible, in Adobe Audition? So it sounds a bit like mumble

Question

Doing sound design for an explainer, with lots of character animation. I have kids and adults that needs to mumble. It's a serious subject, so it can't be too cartoony.

I have some speaks and yells that work well for the characters, but the listener shouldn't be able to hear what they are actually saying, so I need to make it inaudible.

I tried working with the EQ, but it kinda makes it sound like they are in the back of a trunk imo.

Answer 1

I think that comprehensible speech consists of vowels and consonants put in some order. Killing one of the two will make something incomprehensible.

I would try to kill consonants to be more exact, this will bring the sample closer to mumbling.

There are multiple ways you could try doing that like fast compressors targeting specific areas and such but fast compression might sound odd (except if you like the compression effect)

One other technique that you could use is to take the channel and try to target the "easy" consonants (Ks Ss Ps Ts) that luckily also introduce the most articulation. The way you can do that is by using a gate and recording the output or by using a plugin like fabfilter de-esser which provides the user with the "targeted" audio isolated and saves you from some work.

Practically you will need to clone the channel you want to mess with (so A is the original and B is the clone) insert the de-essers to B, flip the phase and start tweaking until you start cancelling one consonant after the other.

Keep in mind that if you use a plugin like fab-filter you will need to clone the channel for as many consonants as you want to cancel cause each de-esser will mute the original sound and only keep the part that de-esses.

To perfectly cancel out a sound the amplitude has to be exactly the same so the tweaking part will have 2 phases, one is to target the specific consonant and the second is to fine-tune the volume so it cancels out the consonant of channel A.

EQing is not easy to avoid completely so after all that some eq-ing will be necessary, Group all these channels and EQ the final output. Do not change anything else to the specific channel because if they are not completely identical you will not be able to cancel them.

So a sentence like "The quick brown fox jumps over the lazy dog" will sound something like "The ui brown fo ump oer the lay do" (ymmv)

Haven't tried the above but I have done similar things over the years, it might work for you :)

P.S. I kinda missed the adobe audition part but I think that it takes plugins like any other program

Answer 2

Conversation can become indistinct due to several reasons:

• It's quiet
• There's much reflected sound (reverb) in the room
• There are many voices talking at once
• It's masked by a louder noise

Experiment with a few of these, or use them a combination:

• Layer up multiple conversation tracks to create a "blur" of voices.
• If recorded tracks are very dry (recorded close, with little ambience), try adding reverb that matches the space they're supposed to be occupying. Err on the side of shorter reverb times (1 second or less), with strong early reflections.
• Reduce high-frequency content and detail with a low-pass filter. Experiment with different slopes (start at 6 dB/octave) and frequencies (start at 8 kHz and go down).
• Add a background noise (air conditioning, traffic, jukebox, live band).

Finally, you could:

• reverse the playback of some (not all) of the conversation tracks
• do the steps above with voices speaking a different language

Good luck!

Answer 3

This answer is for FL Studio but you can also find free VST for bit crushing or ringing for Adobe Audition like https://www.tritik.com/product/krush/ and https://plugins4free.com/plugin/1849/

You can use a bitcrusher (fl studio->effects->effector->low fi) or a ring filter.

Alternatively you can pitch the volume to be pretty low.

here is some demo in FL Studio