hi,
let me first explain my idea by using analogy with pictures & text.

1) for pictures , the latest photoshop version offer a feature "content aware scaling",
that allows to keep main features of a picture, and reduce it without affecting size of main features.
Well rather , than a thousand of words, just look at the video at youtube:
http://www.youtube.com/watch?v=vIFCV2spKtg

2) for text: they are many tools to automatically "summarize" the content.
One I remind is for instance , "Copernic Summarize" , but they are others.

So here's my crazy idea. Like for pictures, and text, why it would not possible to automatically reduce
the duration of audio tracks, and get an overall idea of audio content ?
I mean something that involve some "artificial intelligence", that parse audio content.

Is there any tools for this ?
Thanks

Like something that would summarize the minute waltz in only ten seconds?

@pdq,
well I don't know the "minute waltz" (will see what's this later) , but ideally you would decide the final duration (or percentage of original duration) on the summarized track.

Never mind, that was just an attempt at humor.

AFAIK there is research on automatic refrain recognition which does -- as far as popular music is concerned -- usually carry the "catchy" part of a track. But a real content-aware scaling algorithm for music doesn't exist yet to my knowledge. Probably such an algorithm would have to search for
a) monotonous repetitions and
b) very rarely occurring events,
rate them and then drop them out one after another when the target track length is diminished. Open question: how to preserve the rhythm...

edit: grammar

That shouldn't be much of a problem. There's already enough software which can align sequences of music by their beat intervall. To piece it all together you'd just additionally need to pick out the transition between changing rythmns.