A technique is presented to facilitate the creation of constantly changing, randomized audio streams from samples of source material. A core motivation is to make it easier to quickly create soundscapes for virtual environments and other scenarios where long streams of audio are used. While mostly in the background, these streams are vital for the creation of mood and realism in these types of applications. Our approach is to extract the component parts of sampled audio signals, and use them to synthesize a continuous audio stream of indeterminate length. An automatic speech recognition algorithm involving wavelets is used to split up the input signal into syllable-like audio segments. The segments are taken from the original sample and are not transformed in any way. For each segment, a table of similarity between it and all the other segments is constructed. The segments are then output in a continuous stream, with the next segment being chosen from among those other segments which best follow from it. In this way, we can construct an infinite number of variations on the original signal with a minimum amount of interaction. An interface for the manipulation and playback of several of these streams is provided to facilitate building complex audio environments.