On iOS, is it possible to get the user's audio stream in a decompressed format? For example, the MP3 is returned as a WAV that can be used for audio analysis? I'm relatively new to the iOS platform, and I remember seeing that this wasn't possible in older iOS versions. I read that iOS 4 brought in some advanced APIs but I'm not sure where I can find documentations/samples for these.
If you don't mind using API for iOS 4.1 and above, you could try using the AVAssetReader class and friends. In this similar question you have a full example on how to extract video frames. I would expect the same to work for audio, and the nice thing is that the reader deals with all the details of decompression. You can even do composition with AVComposition to merge several streams.
These classes are part of the AVFramework, which allows not only reading but also creating your own content.