Sound designer Meara O’Reilly has posted a fascinating entry on BoingBoing about how the various tones generated in the human voice tract generate recognizable speech — see “Whistling Speech.”

O’Reilly highlights the work of Haskins Laboratories, a research institute in New Haven, Conn., that does work on speech, language, reading, and their biological basis. Here’s a direct link to a great demo by Haskins showing how three different tones that don’t make sense individually can combine to make recognizable speech. Below the tone chart is a series of links allowing you to play the different tones separately and in combination.

AB — 17 March 2010