Approximating π is a software implementation of a multi-channel electroacoustic composition by Clarence Barlow based on the sonification of the digits of π. It is based on the following series approximation:

π = 4 (1 – 1/3 + 1/5 – 1/7 + 1/9 …)
We implemented the algorithm for his piece to synthesize sound and produce visuals using his original program notes from 2007, allowing for the first ever performance of this piece with multi-channel visuals.

Each channel approximates π at a different rate, all the channels together form a very specific harmonic series based on a formula involving π (more details in the program notes). The whole piece is conceptualized such that it can be flexibly played on up to 16 A/V channels, and for any duration up to 1.5 years. The longest it has run till date is 2 hours, with 6-channels (6 monitors + 6 speakers).

Implementation Highlights:
⟞ Implemented in C++ using OpenFrameworks
⟞ Synthesizes multichannel audio (2-, 4-, 8- or 16- channels)
⟞ Synthesizes multichannel video over multiple-monitors (one for each audio channel)
⟞ Distributed – can run synchronously on multiple machines (uses OSC)

Approximating π was performed both as an art installation and in a concert setting using our implementation at the following venues:
2016.Apr.24  | Lotte Lehman Concert Hall, UCSB Music, for the Montage Concert
2016.May.27 | Elings Hall, UCSB, for EoYS
2016.May.28 | SBCAST, Santa Barbara, for EoYS @ SBCAST
2016.July.12 | Conservatory for New Music, Köln, Germany, for BAR70W

Installation at SBCAST. 6 Channels of audio-video

Installation at SBCAST. 6 Channels of audio-video

Every channel had a monitor and associated speaker.

Every channel had a monitor and associated speaker.

Panoramic image of setup at SBCAST

Panoramic image of setup at SBCAST

Offshoots: With a little modification, this software was used to perform a spatialized rendition of Hector Berlioz’s Symphony Fantastique for a concert by graduate students of UCSB Music.


Original composition: Clarence Barlow
Software: Akshay Cadambi and Matthias Wagner
Github: ApproximatingPi