Analyzing Audio Data

I'm working on a little project in my spare time. My aim is to analyze tiny chunks of audio data in near-real-time and to use the data in an application (lol game). I have some knowledge of audio but can't think of ways to analyze the data aside from discerning the different wavelengths in a given unit of audio. While it would work, there's no guarantee that there will be enough different wavelengths to provide enough variables.

for tl;dr
How the fuck do I get a few dozen variables from a millsecond or less of audio without implementing an easily abused system?