The illusion begins with an ordinary spoken utterance:
"The sounds as they appear to you are not only different from those that are really present, but they sometimes behave so strangely as to seem quite impossible."
Next, one part of this utterance – just a few words – is looped several times.
Finally, the original recording is represented in its entirety, as a spoken utterance. When the listener reaches the phrase that was looped, it seems as if the speaker has broken into song, Disney-style.
…the speech-to-song illusion reveals that the exact same sequence of sounds can seem either like speech or like music, depending only on whether it has been repeated. Repetition can actually shift your perceptual circuitry such that the segment of sound is heard as music…