Making music out of human speech

Originally published at: Making music out of human speech | Boing Boing

Reminds me of this work, which still gives me chills decades after my first listen.


Data sonification has been explored for many, many years, including human speech in just this way, as compositions or compositional elements, stripped of recognizable formants, being converted to MIDI, processed spectrally, sonified directly, etc. long before Melodyne ever existed.

I came here to mention Different Trains and you beat me to it. :grinning:
I was also going to mention Bacharach and David. Burt Bacharach’s melodies are defined by the lyrics. A case in point is “Anyone Who Ever Had a Heart” … the melody of the title lyric is the same note repeated eight times, matched to the rhythm of speech.


That’s pop music for you! And obviously Bacharach and David, with Dionne Warwick’s peerless singing, are an extraordinary example. I’ve often argued with classical musicians that their approach to writing to lyric has been lazy. Look at a page of 18th century music where there’s a stack of notes and one word “magnificat” or “gloria” and big long lines running along the page. Pop music reserves multi-syllabic voicings of single parts of a word for special effects. Typically at chorus time. Much to its benefit I would argue. Western clasical music has tended to be profligate with it which is, for me, where the potency of “cheap music” comes in.

I heard “different trains” one time and it was difficult to enjoy because Steve Reich was sitting just behind me and the kid who had been doing the sound had made a fuck up which led to bombs going off over the PA during the triple quartet before that and Steve was fuming the entire gig. It is however a wonderful piece to introduce contemporary music to noobs as the “sprechtgesang” of the talking samples from the interviews and how it’s taken up in the multiple string quartets is immediately enjoyable to everyone.

It’s also a shockingly dark and powerful emotional piece too. Europe during the war is almost unbearable.

1 Like

I first heard it as a student in a classroom studio setting (excellent acoustics) and my ears wanted to crawl off my head and hide. Once the air raid sirens start up… [shivers] Even without the silent menace of the composer fuming nearby, it carried a lot of emotion. I liked the recording so much (Kronos Quartet) that starving-student me made a copy from the library.

Hearing the piece in a well-lit studio was one thing, but playing it on a solo night drive across the plains of the Midwest… that elevated the weirding out factor in the best way. Decades later, I still associated that movement with the feeling of motion, an unchanging horizon, and a dark starless sky.


Yeah this is reminiscent of Different Trains, but even more the process described is very commonly used in meme videos that harmonize spoken word with music, e.g.:

It seems that the chosen tool for the job is usually Ableton’s “audio to MIDI” feature rather than Melodyne but otherwise the process is identical. Basically you extract the melody and pacing of the spoken passage, write an accompaniment to those notes, and then play it along with the speech.

a variation on that concept:

This topic was automatically closed after 5 days. New replies are no longer allowed.