The real phenomenon at the heart of this is pareidolia - http://en.wikipedia.org/wiki/Pareidolia. This is the mental bias that makes it easy for us to see a face in places where they really don’t exist. It’s the pattern recognition software that our brains use to help us get through life; language would probably not be possible without it.
The interesting thin about pareidolia is that once a pattern has been pointed out in a non-specific sound our image it is almost impossible not to see it.
There is a brilliant clip where a slurring performance by Joe Cocker at Woodstock is subtitled with inaccurate lyrics, it’s very funny and illustrates the point well.
http://www.youtube.com/watch?v=3xJWxPE8G2c
Don’t believe everything you think.