The human voice in many ways mimics a synthesizer. The voice box is the oscillator, the throat, tonque and month create the filter from which language and song is created. Modulation is also used to create vibrato effects n the voice. You might just say that from the time we are an infant we learn to play a synthesizer.
The Sonovox is a curious little device. I revised my original post here because I can't really figure out how this device works. It acts much like a vocoder. There are analogue vocoders but they require a series of filter banks which is clearly not what is going on with the Sonovox. The descriptions, which are very terse on the internet. This is my best guess as to how this device works:
A small speaker is placed up against the voice box and the singer articulates the words (accordinig to one web site as a whisper - i..e. with no vibrations coming from the voice box). My suspician is that another mic with a large diaphram picks up the articulated vibrations.
I found this on the talkinig guitar which sounds similar to the Sonovox and the Heil Talk Box:
The principle is the same. The sound of the instrument is amplified and in the case of the talk box and talking guitar sent through a tube which acts as a wave guide. The sound is then shaped by the mouth and picked up my a mic. What I can't figure out on the Sonovox is where the placement of the mic is.