Richard Wiseman and Simon Singh do a great bit in some of their presentations, demonstrating how easy it is to make someone hear just what you want them to hear simply through suggestion. (I’ve seen both guys use this, so I’m not positive who developed it but I’m willing to let them mud wrestle over it [naked]. WAIT, EDIT: I’ve just been informed the originator is Chris French! Thanks, Sid.) They play Stairway to Heaven backwards, and it’s all garbled. Then they play it again, this time giving you lyric to read along — the second time, you can actually hear words. It’s interesting to compare this to electronic voice phenomena (EVP), where “ghost hunters” record garbled nonsense and claim to hear the dead telling them to bite the purple banana (or whatever). Then of course it can also be extended to show you how gullible even the most skeptical brain can be.
All that is to say, here’s another fun example I found on a friend’s Facebook, starring Joe Cocker.