I spent a week in October at STEIM working on new techniques for making computer music. Echo Nest Remix is a set of tools for using detailed, automatic audio analysis to remix and compose music. I wrote software in Python and to chop up an existing song into segments and to cluster those segments by timbre. I used Max/MSP to loop the sounds in each cluster and treat these clusters as the voices in a composition, fading them in and out, modifying their speed and repetition, and finding interesting combinations of timbres.
Working with a handful of recordings by harpist Kyle Oppenheimer, I ended up with distinct groups of sounds: sharp, accented notes; soft, sustained notes; treble-heavy chords; the sound of Kyle’s hands running up and down the harp strings. I layered and looped these sounds until new music grew out of them:
The music and software I created during my residency are part of a larger goal of making audio analysis technology useful as a creative tool. The staff at STEIM had many ideas for this project that I hope to explore in the future. Many thanks to Taku, Jon, Nico, and the rest of the STEIM crew for the hospitality and inspiration.