Turning spoken voices into musical events
The sound of a person speaking is actually a very complex audio event that contains a lot of interesting evolving waveforms, transients and formants. You can use this in a piece of music, by transforming the voice and breaking it up into elements. This can be done in many ways, and you will get a few tips for this below.
Granular breakdown. If you split the voice event into many small individual snippets and change the order of these snippets you will get a sequence of complete gibberish, that still carries many of the characteristics of a human speaking. Of course this can be done manually, but this is very time consuming. A better option is to use a digital granular effects processor, which can instantly – and sometimes even in real time – chop a recorded or incoming audio into snippets and create a sequence of these.
Extreme pitch correction. Normally the human voice will fluctuate in pitch to create a pleasant melodic flow in the spoken words. Different languages and dialects do this differently, which is why you’ll find that some languages or dialects sound more musical than others. By completely removing the melody from a spoken line you will get an unnatural sounding voice with a robotic and mechanical character. This can be done by extreme pitch correction, where you force the voice to feature only one pitched note.