The Generative Electronica Research Project

Home

Introduction

GERP is an attempt to generate stylistically valid EDM using human-informed machine-learning. We have employed experts (mainly Chris Anderson) to hand-transcribe 100 tracks in four genres: Breaks, House, Dubstep, and Drum and Bass. Aspects of transcription include musical details (drum beats, percussion parts, bass lines, melodic parts), timbral descriptions (i.e. “low synth kick, mid acoustic snare, tight noise closed hihat”), signal processing (i.e. the use of delay, reverb, compression and its alteration over time), and descriptions of overall musical form. This information is then compiled in a database, and machine analysed to produce data for generative purposes.

Two different systems have been created to interpret this data: GESMI (created by Arne Eigenfeldt/loadbang) and GEDMAS (created by Chris Anderson/Pittr Patter). GEDMAS began producing EDM tracks in June 2012, while GESMI produced her first fully autonomous generation in March 2013. It is interesting to note the similarities of the systems (due to the shared corpus) and the differences (due to the different creative choices made in the implementation).

Both systems are, for the moment, only using the Breaks corpus of 24 tracks.

Both systems will debut at MUME, in Sydney Australia, in June 2013. GEMSI will follow this a few weeks later at the aCoAx Festival in Milan, Italy.

Both systems are coded in MaxMSP and Max for Live, and use Ableton Live 9 for sequencing.

Chris Anderson's work in figuring out how to reference Live instruments through M4L has been integral to the success of automated timbral selections, as well as figuring out how to dump data into Live clips.

The next three steps in the project include: 1) further exploration of autonomous timbral selection, using machine learning techniques on the corpus data, and the available Live instruments (fall 2013); 2) automated signal processing (late summer 2013); 3) complete analysis of the remaining corpus (House, Dubstep, Drum and Bass), and its inclusion in the generation algorithms.

Papers

A New Analytical Method For the Musical Study of Electronica
Proceedings of the Electroacoustic Music Studies Conference, Sforzando! New York, June 2011.
www.ems-network.org

Towards a Generative Electronica: Human Informed Machine Transcription and Analysis in MAXMSP
Proceedings of the Sound and Music Conference 2011, Padova Italy July 2011.
http://smcnetwork.org/smc_papers-2011187

Towards a Generative Electronica: A Progress Report
eContact! 14.4 Toronto Electroacoustic Symposium 2011
http://cec.sonus.ca/econtact/14_4/

The Human Fingerprint in Machine Generated Music
Proceedings of xCoAx2013: Computation, Communication, Aesthetics, and X Bergamo Italy, June 2013.
http://2013.xcoax.org/

Evolving Structures for Electronic Dance Music
Genetic and Evolutionary Computation Conference (GECCO). Amsterdam, July 2013.
http://www.sigevo.org/gecco-2013/

Considering Vertical and Horizontal Context in Corpus-based Generative Electronic Dance Music
The Fourth International Conference on Computational Creativity, Sydney (2013)
http://www.computationalcreativity.net/iccc2013/

Sound

LoadBang

Pittr-Patter