My original intent for Fluid Dynamics (2002) was to use two rhythmic sounds I recorded—a raucous faulty faucet in a men’s restroom near the Concert Hall and a gently squeaking gas service regulator outside of Lisser Hall, both on the Mills College campus. As the piece developed, though, the rhythmic elements were set aside as the more subtle sound of gas traveling through the pipes and the soft purring sound that the faucet made on its way to the clacking rhythm became the foci. To these sources I added the sound of a large steel ball and a small brass ball bearing being propelled across a wooden floor, a spare MCI tape machine part rolling on a linotype sheet, and very thin brass sheeting gently swaying. The other main sound is that of a large steel ball rolling down two strings of a miniature koto-like instrument.

The sources are processed using phase vocoding, convolution, granular synthesis, equalization, and extensive layering. Although residual attachments to the original sounds remain, often their origins are rather obscured. The spatialization is natural. At times more static sources are convolved against naturally moving sources so that they take on the spatialization of the moving sources.