Ian Andrews

Generative


auto-Change (2009)


auto-Change 1 (5:40, 250K)
auto-Change 2 (10:00, 540K)
auto-Change 3 (5:40, 488K)

A number of years ago I came across a box of turntable tone-arm headshells. These are the bare enclosure with a 4 pin connector that work with most professional turntables. I had the idea to construct my own cartridges and styli from found components: speakers from discarded headphones, old microphone elements, needles, springs, small pieces of spring steel, etc..

After constructing the cartridges I went in search of various objects to place on the turntables and play: plates, saucepan lids, metal film canisters, ceramic plates, and even records. The most productive of all was a stainless steel disk with multiple perforations of different sizes to which I attached spring steel clips that would function as obstacles in the path of the stylus.

The objects were played with the cartridges on three turntables at a performance, Turntablism, (Now Now festival) in late 2007. Later I made individual recordings of each cartridge with various objects on one turntable. A Sedco GE12 graphic eq, an Orban 372A parametric eq were used in conjunction with a compressor to isolate the most interesting sounds from the apparatus. Next I edited a short sample from each recording to be used as a loop. Since the sounds produced by the set-up were cyclic, the duration of each sample would correspond to the length of the cycle at either 33RPM of 45RPM. In some cases, where there was significant variety between two or more consecutive cycles, the duration of the samples were extended to 2 or 4 cycles.

About 40 samples were used to create the 3 pieces in auto-Change. Each piece is made up of 13 segments that are played consecutively in random order. Each segment contains 1 sample that is looped a number of times (typically 16). The reason for looping the samples rather than using longer sections of the recordings was to conserve bandwidth. The segments, which have an attack and decay envelope, overlap each other and are thus mixed according to random selection combined with differences in duration.