Don't forget to up-vote the question or answer if you find it helpful! If it's your own question and you're very happy with an answer you can accept it as best answer.
Hello Sound Geniuses!!
Somewhat of a strange question, but I am trying to figure out a way simulate the way a young man's voice will crack during puberty. Basically to take an existing voice record and process it so it sounds like they are going through puberty. I know it is a little strange, but it is for a project and I thought I would ask around for ideas. It might not be possible, but maybe I am missing something.
I have tried random pitch shifting, EQing in the nasaley/shrieky upper mids, Freq-Mod, etc, but nothing is quite right. Any insight would be greatly appreciated. In addition, this would need to be programmed, so concepts versus plug-in setting would be best. Thanks!!
From where you've said this needs to be programmed, am I right in thinking you're trying to create an effects processor that alters any dialogue track you input to sound like the voice is breaking, rather than just attempting this on specific piece of already-recorded dialogue?
If this is the case then it sounds rather a tricky. You could work out a way of programming in random crossfaded +1 octave shifts, like @Utopia suggested, but I've a feeling this will sound rather sharp and mechanical.
In an ideal situation you'd analyse the characteristics of a pubescent teenage voice and physically model this but that is a huge amount of work requiring a tonne of research etc etc. If that was your real intention then I'd suggest getting in contact with the Antares guys and discuss things with them but I've a feeling it's not!
Please note: When commenting on someone's question or answer, you will not receive notification on follow up comments unless you are mentioned with @Username. So for Andrew Spitz, you only need to use @Andrew
Chances are more than 75% of the result is the actors performance (e.g. them articulating their voice that way the best they can, and/or using ways they can manipulate their voice in real - life like helium, crunching their nose, talking through objects, etc). Check out Antares AVOX Throat (I think that's what it's called?) Even very very CAREFUL uses of Autotune or equivalent tools may work well. Chances are though a lot of it is going to come down to how the actor articulates their voice, either originally or re-recorded. There's only so much that plugins and processing can do, and especially with pitch-related alterations, our ears are very sensitive to the realism versus not real versus uncanny of voice pitch so these alterations are not that that forgiving compared to cadences and tempo, these latter ones which can usually have more wiggle room for alterations but still require a delicate editorial touch and feel for a voice's inherent musicality.
For EQ, the nasal will sit at about 750-900Hz usually, the exact harmonic differing with each voice but this tends to be the usual 'backyard' for nasal qualities.
Sounds something that I personally would not even try. For some reason there sometimes seems to be a tendency to think that everything can and should be attempted to be done with post-processing instead of getting to the source.
I guess Avox Throat (although I have no personal experience with it) or certainly something that does partial and noise based analysis and editing is likely what you would be looking for this type of human voice transformations. It's something that can be used to make a voice sound younger or older (by modifying the ratio of the noisy and tonal parts), to a certain extent, but the effect is still rather, if not clearly synthetic.
I'm sorry to say, but if the actor recorded it without any cracks and you're trying to put the cracks in from scratch, I don't think you'll get very far with it.
If you try to duplicate it right now with your own voice, you'll notice that a voice crack like those of young men tend to go about an octave up, so possibly selective pitch-shifting an octave up and cleverly crossfading between original and pitch could get you somewhere as a rough edit, but ultimately, I suggest re-recording it and have the actor do the voice cracks originally. To put something like this in after it's recorded in my experience has come to no avail and sounds worse and takes more time than it would have taken to re-record it.
I've done this before in a movie using pitchshifting with good results, but there are no way I could ever make an algorithm that even remotely works on all voices. It took a LOT of tweaking before it actually sounded pubertal, and it only works on pretty deep and rich voices, or it will just sound weird, and they must be pretty expressive as the cracking comes from trying hard to control the voice but not knowing how. Myself I actually never went through this tho :-)